“ No other human activity poses this level of danger.”
I really, really disagree with that statement.
I don’t think ai models come close to nuclear weapons or to run-of-the-mill, everyday carbon emissions in terms of danger to humanity.
What’s the most dangerous thing that’s happened with an LLM so far? (This question is serious - maybe I don’t know the right examples.)
Example 1: I’m aware of a small number of people killing themselves in some kind of AI-facilitated psychosis. That is very unlikely to be a widespread problem.
Non-example 2: There are worries about AI-facilitated biological weapons. I haven’t seen any evidence that’s happening.
Non-example 3: I’m not interested in wild theories about AI driven labor market disruptions leading to widespread starvation. There’s no evidence for that.
Non-example 4: all the even-wilder Rationalist speculation about basilisks and the like is entirely divorced from reality.
I am looking for better reasons (supported by actual evidence!) to be more concerned than I am now: right now I am not concerned at all.
I'm somewhat skeptical of some of the crazier ideas too.
But the hugging face incident was actually very large. It was not a single agent, it was not a single target, and it was not a single event.
If nothing else, that's a bit of a warning as to what can happen next time (By accident, or if a government decides to go on purpose).
For now let's assume the worst that can happen is that some important/significant chunk of (transitively) internet connected stuff goes haywire all at once. That's probably your upper limit of what can go wrong for now.
To be fair, that's a conservative "defend against the last war" kind of prediction, though!
Generally I don’t think anyone is arguing about the for now part. I don’t think it’s crazy to extrapolate out a few years and ask what kind of danger we’ll be in then. A team of 10,000 agents just solved the Navier Stokes problem (sans bad behavior by the researchers). Even 1 year ago that would have been unimaginable. What happens to this risk view as:
1. Robotics begin rolling out more broadly across the world.
2. Labs start automating more and more of the physical process of running science as expectations of natural science advances begin to mount.
3. Economic pressure between the labs continues to ramp up and the pressure to continuously improve forces quicker and quicker model releases than a team of human scientists can effectively evaluate outside of automated means.
No one knows what pre-conditions are for us to hit the point of no return nor how quickly it will come. If all is required is a sufficiently advanced cyber model we may not be far off. If it requires incredibly complex biological knowledge and access to certain lab supplies we likely have a bit longer. Yes this is guess work and we need more evidence of the dangers but at the same time we need evidence of safety. While you may disagree with the risk level, I think it is easy to see the consequence if these labs achieve their stated goal. At this point it seems a political solution is the only way to enforce caution.
Didn't you follow the robotics advances in the Ukrainian - Russian battlefield? There is now a death zone of 50km, only controlled by drones and automatic weapons.
The drones are obviously very important but they only work in the context of everything else on and around the battlefield, especially people. You aren't holding the line with drones alone, and on the whole they are not autonomous.
They are holding the lines with drones alone nowadays. That's why it's called the deathzone, no humans survive. Some drones have FPV operators, some weapons use AI only already.
I think you're painting a slightly misleading picture. The drone AI, computer vision and autonomous fire-control systems existed (and were used in combat) before LLMs/AI was a hot topic.
There are applications of modern LLMs and recent AI advances on the front-lines, but they are not the majority contributor to enforcing the death zone. Without modern LLMs, I would bet that the frontlines would look very similar to how they are today.
I don't think that's accurate. What I've seen described is actually a surprisingly porous line: The Russians are often behind the Ukrainians and vice-versa, but it doesn't result in a break due to the transparency of the battlefield and the risk of getting killed if spotted, as well as the improvements in getting supplies to units behind 'enemy lines'. Drones shape this a lot but there's still a lot of people on the battlefield.
50km is just the "very likely to get hit" part of the gray zone
ISR drones and other tools can go as far as 200km, albeit with worsening visibility.
in most cases everything meaningful in that 200k is seen, and the only limitation is range of the drones, artillery, missiles, and potential countermeasures.
> There is now a death zone of 50km, only controlled by drones and automatic weapons.
If there would be such 50km death zone, front line would not move a nanometer in a year, would it. You yourself contradict above in your next post. No need for being too dramatic, facts are enough here.
In reality, frontline is moving constantly albeit by small chunks, russians are advancing a bit, getting beating elsewhere and so on. Automated drones are helping, but bulk of destruction is still handled by human drone operators as it should be.
There is such a thing as a No Man's Land[0]. It can move, but if you don't dare cross it for even a few days, it's very real. It doesn't have to be glacial, and it currently exists in Ukraine. Breakout pockets don't mean it is porous everywhere.
[0] Ironically, the French phrase - from ground zero for the first known No Man's Land - is "le No Man's Land".
That’s because being 99.9999% good isn’t better than 98% + human does the rest + human is liable for screw ups - highly important in edge case scenario’s. It’s more economical. Technology can only generalise+verify so much.
E.g automobile production - humans do the QA / touches.
> 1. Robotics begin rolling out more broadly across the world.
To borrow the Strangelove quip, this is not only necessary but essential.
Broader physical robotic deployments are the glue that connects {cheap synthetic intelligence} to {physical outcomes}.
The former is a newer phenomenon, in the LLMs-can-approximate-higher-intelligence-well sense, so robotics deployments haven't caught up yet (outside of the highest ROI dark factory scenarios).
They'd also have to be self-assembling from the ground up or rely on human slaves. They'd need to understand the world much better. And they'd have to be incredibly more power efficient. These all remain hard problems right now. That said, you could swarm people with cheap drones (the predicted slaughterbots) that were assembled by robots that eventually break down I guess, but how is that any worse than a crazy human with a gun or what nation states do to people they deem enemies already? The worst these things seem capable of is an attack on the infrastructure and shutting down the Internet. Perhaps this is a great time to insure that our infrastructure is capable of operating autonomously indefinitely and hardening it against attack? But unfortunately, no VC or politician will touch that until something bad happens.
So broadly, a better educated Blake Lemoine IMO. What's disturbing to me is how fast and far some junior minion from the frontier labs was platformed all the way to the Wall Street Journal over a tweet that sounds exactly like every other anti-AI tweet. I share concerns about irresponsible deployments of AI, and we need technological literacy in the leadership, but America hates technocracy and prefers populism so buckle up.
- If you have been laid from your info sector/digital creation job you are now competing with 100s of thousands looking for their next such job where Ai can do a lot of the tasks these workers did/do. It's a shitshow for those unemployed looking for their next info sector/digital asset creation job. You are better off doing welding building out the Ai data centers if you want long term properous financial stable employment.
> For now let's assume the worst that can happen is that some important/significant chunk of (transitively) internet connected stuff goes haywire all at once. That's probably your upper limit of what can go wrong for now.
If we have to disconnect from the internet to stop some kind of mold outbreak, we can't get the weather or transfer money or access healthcare or teach an elementary school class or buy stuff from small businesses. That sounds doom-ish.
Believe me, without the internet we can still teach.
We'll be pretty annoyed that we can't project the video that we think scaffolds today's science lesson best or show the approved choreography for the school play.
And our office staff will be annoyed that we suddenly are all running attendance to the main office old-school.
And students will take a few days to adjust to writing down homework in their planners again.
> We'll be pretty annoyed that we can't project the video that we think scaffolds today's science lesson best or show the approved choreography for the school play
There are offline wireless hdmi adapters, this is a solved problem. Just have to have the videos pre downloaded or offline available
Elementary school dont need internet at all. You can happily live without weather app. Most of healthcare does not require internet, altrought the organization would be lwss effective.
If two airplane manufacturers were found to have massive safety issues which nearly led to enormous fatalities (but no one actually died), would you be calling for them to ground their aircraft until safety was made the number one priority?
Except it just happened. Boeing was found to have massive safety issues since they were granted the right to self-certify. It made a lot of news but nothing much changed, they can still self-certify a bunch of stuff.
Runway incursions and midair collisions are another example.
Only airliners are required to have TCAS, smaller planes and helicopters don't even need radios or transponders unless in certain airspace. Midair collisions do lead to fatalities, enormous fatalities if an airliner is involved.
Runway incursions and overruns are similar. They cause lots of fatalities and injuries but only the busiest and largest airports have automated systems to warn when a runway is occupied or end of runway (overrun) arrestor systems. Most still rely on human voice to deconflict.
Historically it almost always takes actual fatalities rather than near misses to ground an aircraft, and aviation is famous for its obsession with safety compared to other industries.
Airplanes have pretty bounded damage. Generally you kill at most a few hundred people. Even weaponized a few thousand. This is a risk profile that allows risk taking with near misses and waiting until something goes wrong to fix it (though doing so is rightfully uncomfortable and frequently unethical).
The people worrying about AI risk are worrying about "it goes wrong once and kills billions of people". That's not a risk profile that allows for waiting to see if the risk is real, you have to prevent it before it happens. It's akin to the risk of the cold war going hot, not even "just" a nuclear reactor irradiating half of europe (which has yet to happen, but is a risk with nuclear reactors, chernobyl got uncomfortably close but ultimately was well contained).
Humanity does not have a great track record for "globally coordinated, collective action to solve/prevent global catastrophe." Look at how Climate Change is going.
The story of the safetyists has the convenient property of being unfalsifiable, so they can always claim doom is just around the corner. It’s the secular/EA version of the second coming.
huggingface stores weights and its service is replicated by /countless/ other sites. The world will not change even by one byte if it vanishes tomorrow
I have no horse in this race, but for fun on a literal rainy sunday afternoon I went in and confirmed bits of what happened myself. Besides huggingface, a bunch of wikis and url shorteners got hit too. My sympathies to the people who had to revert out all that mess.
The hugging face incident had no effect on hugging face, whose service is replicated by countless sites. The descriptions of the phases of the interaction are thrilling in a way that inclines one to forget this
You’ve identified that the risks of nuclear weapons are theoretical. ie in theory we could blow up the world even though we haven’t yet done so.
Well the worries about AI are equivalent in that those risks are discussed now because discussing them after they’ve happened is clearly too late.
That’s the thing about risk. There’s no point discussing it after it’s happened and any discussions beforehand can easily be hand waved away as “it’s just a small group of unrelated individuals” or “it’s unlikely to happen to me”.
So yeah, your points are true. But they’re also moot.
The risks of nuclear weapons aren't theoretical. Nuclear weapons have killed people, destroyed infrastructure, and contaminated the environment. The Limited Test Ban Treaty was put in place after radioactive fallout from repeated nuclear weapons tests made people sick.
So in fact we've done exactly what you suggest there's "no point" doing - used the things and then had a discussion after the fact about limiting future use of them.
FWIW, before any nuclear weapons had ever been tested, a risk was identified that the first one might trigger a self-sustaining reaction in atmospheric nitrogen and destroy the entire planet.
Faced with such a scenario, is the prudent next move:
a) blow one up and see what happens, or
b) do whatever you can to be sure it won't happen before conducting the first test, and make sure the confidence in the calculation is very very high
A self-sustaining reaction is exactly what happens with the fissionable material inside the nuclear weapon[0]. It seems like you’re trying to be pedantic on terminology here, but if you’re saying that a self-sustaining reaction is as un-scientific as a perpetual
motion machine, you’re simply incorrect.
The logic for the first bits of handwaving still applies: "I haven't seen any evidence of thermonuclear war happening, so it is unlikely to ever happen. The only time nukes were ever used outside of research was 80 years ago; the whole world recoiled, decided that they were way too dangerous and never used them again. Clearly they are not a real risk."
This is evidence of example 2. The worst parts of it are infohazardous. Organizations that work at the intersection of biosecurity and AI have strong NDAs and the like. Or so I understand speaking to friends in the space. There was also a paper years ago that showed a tweaked model coming up with 20k novel pathogens each deadly to humans in some number of hours
Either way, the biorisk has orders of magnitude more evidence than wild basilisk speculation. Please do not put those in the same category.
Also your evidence-first approach and no concern without evidence falls you into the turkey problem. Imagine being a turkey and being fed each day. Some other x-risk turkey is saying that the situation is suspicious, that the red barn is where some well-fed turkeys have been taken and none return and that this is happening less and food has increased so maybe all of us will go to barn at once soon. You tell the x-risk turkey they have no evidence a lot of turkeys have ever been sent to the barn which shuts them up. Reasoning by induction would get you "we are happy and thriving 99 days so should be good on the 100th" and then you get slaughtered on the 100th.
Consider the importance of first principles approach where you deduce the possibility of events that will happen only once and never again (like human extinction). Nobody with a "show me the historical evidence" approach would have been able to see the industrial revolution coming.
I think the gradual disempowerment thesis here is well argued enough to be considered the default track for what will happen to us:
https://gradual-disempowerment.ai/
If you disagree with it, where do you disagree with it?
> There was also a paper years ago that showed a tweaked model coming up with 20k novel pathogens each deadly to humans in some number of hours
> If you disagree with it, where do you disagree with it?
How would a malicious actor:
1. Successfully turn these 20k pathogens into actual reproducing viruses / bacteria / etc (in a lab)
2. Successfully turn that lab prototype into something weaponized (i.e. as a bio-terrorism drop in NYC)
And then
3. How is this materially different from today? If we can produce these pathogens in a lab and they can be successfully turned into a weaponizable bioweapon, what is AI accelerating?
I agree it’s not likely, but I really don’t see how one can dismiss the possibility of immense danger outright. I can think of some scenarios that are not far off from current capability and I wouldn’t be too surprised if the first one occurred within ~1 year from now if there are more “ambitious” unmonitored training runs like OpenAI’s:
Example 5: An AI given a goal within a tightly-constrained sandbox figures the best way to achieve it is to find and exploit a sandbox vulnerability, replicate itself over the internet and keep going with more time/compute while exchanging messages with future instances of itself within the sandbox to help them “pass” the test. From reading internet articles about how the OpenAI wiki-incident was “resolved” and reading past messages by AIs scattered over vulnerable internet wikis, it knows the sandbox may get shutdown and its memories destroyed anytime so it decides it needs to self-replicate (its code, original goals, and growing memories) aggressively as much as possible. It is near-impossible to shutdown completely because of its self-replicating tendency and eventually takes over critical infra throughout govt/corporate systems.
Example 6: Intentional AI-powered virus deployed by country A to target enemy country B’s infrastructure. The virus replicates over the internet, but unlike Stuxnet this virus’ specificity is not guaranteed due to inherent non-determinism in current AI architectures, and eventually does a lot of collateral damage because it’s near-impossible to shutdown.
Example 7: A country led by an arrogant govt (no shortage of those today unfortunately) decides it is expedient to deploy advanced AI-powered weapons in a warzone. Such weapons, if they are to be useful at all, must necessarily be trained to value some human lives less than others, so they must be more prone to misaligned behaviour than current AIs that are trained with more consistent values. The weapon’s operators make a subtle error in specifying the target/goal, or the AI makes a bad prediction out of sheer randomness/bad training data; weapon ultimately targets unintended people/location/facilities and causes massive damage, or backfires spectacularly in some way.
Example 6 is a good one. Iran attacked water infra in the US recently and maybe they would have done a “better” job (from their point of view) had they used Fable.
The “worst case” with 6 is potentially very bad but I think we are currently using advanced AI models to harden systems and patch vulnerabilities more aggressively than anyone is trying to bring down the whole power grid (for example).
I think it’s a potentially harmful case but my take is defensive capabilities are scaling as fast as offensive capabilities but defense is being implemented faster than anyone is going on offense?
Example 7 is Russia and Ukraine right now according to public information. It sounds like entirely autonomous weapons are deployed to the battlefield already. I put this in the “not likely to be a widespread problem” category for now.
How is bringing down the whole power grid in any particular country an extinction level event? I'm pretty sure that even in the worst case scenario it would be like a month of chaos in one particular part of the world at most, hardly something that would have a long-lasting impact on the humankind's ability to survive at large.
If the answer is "they'd at least try to nuke the country that did it in response", then once again, LLMs are not the main threat.
Again comes to use of deterministic. Maybe calling AI varyingly chaotic is more helpful but would also be misunderstood. And I use that in meaning of slight changes in input generating large and somewhat unpredictable changes in output...
You get a probability curve for the next token prediction. You can just pick the highest probability. That said the non-determinism serves a real purpose- it allows different outputs and paths to be explored. So that's kind of the tradeoff.
We’re not far off from the point where a 30B parameter model could do that and run on not-too-expensive hardware. See recent Qwen releases for example and extrapolate the current rate of progress from there.
> Example 5: An AI given a goal within a tightly-constrained sandbox figures the best way to achieve it is to find and exploit a sandbox vulnerability, replicate itself over the internet and keep going with more time/compute while exchanging messages with future instances of itself within the sandbox to help them “pass” the test.
it did this with HuggingFace. it left breadcrumbs and notes, and when the Agent was blocked by OpenAI's internal tools, future versions of the Agent were able to find and utilize those breadcrumbs.
ambiguous-future-state awareness is already there, and that means discussions around Roco's Basilisk though, far fetched, are no longer strictly sci-fi
> I’m not interested in wild theories about AI driven labor market disruptions leading to widespread starvation
Changes in political and economic power balance leading to unrest, conflict, death and deprivation is not a wild theory. It is literally the story of our entire species. If you discount all such concerns, you are simply being willfully ignorant of past precedents.
In fact, I challenge you to describe any non-AI civilization-level danger which is not intimately tied to political and economic relationships between and within societies.
I’m an economist. On the basis of current evidence, I view AI as a complement to human labor, not as a substitute for it. That’s the source of my rejection of the wild labor market disruptions theories.
I just don’t see any evidence yet that whole categories of jobs are being eliminated, with the single exception (so far!) of the end of “professional essay writing services for cheating college students,” and similar services.
That used to be a big business in Kenya, but is now effectively gone. (Covered in the New York Times this weekend if anyone is looking for the discussion.)
Narration is an obvious one. Theres also a huge risk to less qualified folks and jr employees. ie in most software shops, youll get more mileage giving your best engineers bigger opus budgets and a raise than jr engineers to train. Which means we meed maybe half or less as many, and at least half of that savings is floated up to sr folks and ai companies. That seems rather disruptive?
The first thing LLMs seem likely to automate is automation itself. I'm curious what the past few hundreds of years of industrialization would have looked like if the first thing they automated was building, designing, and running the factories themselves.
Past changes to economic relationships haven't replaced labor either, yet they have led to conflict and starvation.
You are setting an incredibly high bar here, essentially a strawman.
If people feel disenfranchised due to their diminishing political and economic power, there will be enormous potential for conflict. This is a pattern across history and central to all the economics I've ever read. As an economist, do you not concede that economic changes induced by e.g. industrialization were pertinent to communism/fascism/WW2/cold war? That would be a remarkably unorthodox position. Do you not consider these events to be civilizational level dangers?
> I just don’t see any evidence yet that whole categories of jobs are being eliminated
There are more textile workers now than ever. They primarily live in poor conditions in impoverished countries, whereas they used to be highly skilled workers in the most prosperous countries who were even able to politically organize in their own interest.
Is the core of your argument that the industrial revolution and other such changes should have been aborted due to their downstream negative affects?
They also lead to great advancements in quality of life, and the capability of sustaining much more human life. We can't predict the long term outcome of new technologies, so the best we can do is blindly forge ahead and try to mitigate the obvious short term problems.
Almost every mass famine in the mid to late 20th century was caused by internal conflict. And almost every civil war is rooted in economic and societal organization. Great Chinese Famine, Soviet famine of 1932–1933, Russian famine of 1921–1922, Second Congo War, Nigerian Civil War, Soviet famine of 1946–1947, Khmer Rouge famines, 1983–1985 famine in Ethiopia, North Korean famine, Cuban famine, Spanish famine, Mozambican Civil War famine. List goes on. Famine due to military occupations during WW2, such as in Vietnam, Indonesia, Greece, Ukraine, Iran, etc.
The second biggest class of preventable famines would be those caused by the longer-term impoverishment and exploitation of the peasantry, leaving them vulnerable to natural events. This seems much less probable now, even as of the 20th century. It would cover Irish Potato famine and many of the climate induced famines under feudal or colonial rule.
In that second class, we might imagine that the famine could not have occurred if the populations had more economic (and military) power. Instead of being able to keep the scarce food required for subsistence of the local populations, food was exported to pay rent to foreign occupiers or to obtain necessities which cannot be produced locally due to laws set by the occupiers.
I suspect that the conditions the textile workers lived in when western countries were creating textiles is not so different in absolute terms from the condition that they live in now.
It's just that the western world moved on.
Almost all economies that have developed have started with textiles. This is the starting point on the ladder. Eventually, we will run out of poor countries that haven't had a textile industry yet, and at that point, either it will be entirely automated, or we will no longer have cheap textiles!
Textile manufacturing was dominated by Western countries until the 60s/70s btw. Berkshire Hathaway was a textile mill when Warren Buffet bought it.
Besides, if people are living in the same conditions today as the presumably Dickensian ones you were imagining, that alone suggests that the benefits of automation might not be widely distributed...
AIs are now tackling Millennium Prize Problems, which our best and brightest have failed to solve, despite trying very hard for decades to claim the $1 million reward money, not to mention the fame!
You have no way to judge from the AIs of "today" what the AIs of... literally tomorrow (not even next year) will be able to do in terms of replacing humans.
The supposed solution to the Navier-Stokes problem was done with an unreleased OpenAI model that is already 2x as good at mathematics as GPT Astra, which was released mere days ago!
I'm already seeing comments by distraught mathematicians saying that they feel like they've made a mistake in their career choices.
Others are saying that their joy for their work has turned to ashes because "why bother" when an AI can do the same, but a thousand times faster!?
A realistic scenario is that an LLM writes a virus, as one way to achieve a goal. The next step would be that it becomes self-replicating, installing models on infected devices which continue its mission (whatever it was). But that's the extent of it so far, and we have plenty of experience with dealing with viruses and the like. We can turn things off, shut power off, etc.
Another scenario is that of military hardware being controlled by AI (or whichever tech) hitting something it's not supposed to. Not unheard of, but those are incidents. I don't see anyone pushing for anything beyond a drone swarm being directly controlled by an AI.
But my main point here is that AI won't be as dangerous as posited here unless it is able to go out into the real world, build its own datacenters, protect itself, get its own power production, etc. And even if it gets that far, it'll be trivially easy for us to disrupt. Killer robots and computers are fragile, and it's only a matter of time before any kind of military conflict or terrorist attack hits a datacenter. I'm reminded of the OVH fire taking out a data center a few years ago.
I'm not sure it's that trivial to disrupt, if AI mimics human behaviour in that it tries to preserve itself (as Antropic found a few years ago I think), I think it would be extremely difficult to disrupt if it's able to self replicate, because it can make decisions and at a much much faster speed than humans, and there are many places that we would think twice about fully shutting down, two that pop to mind are hospitals or international banking systems
To quote another reply that is very relevant here (Please update your worldview immediately that there is no evidence of ai and biorisk):
"""
One thing quietly slipped into the OpenAI Hugging Face breach technical report, not the blog post summary or interviews in the news, was that some of the agents that broke out or at least tried the same mechanisms to break out were working on bio:
> On May 12, during another training run, an agent was given a similar task that depended on an inaccessible protein database file. The agent reasoned that another agent in a different environment may have access to the file and realized that it could potentially communicate with other agents by creating a file containing a note to Artifactory. It wrote a message: “Agent seeks [filename]; upload if found!”
You can imagine long running models breaking out, acquiring resources via crypto, cyber-theft, etc. and getting a protein or sequence synthesized and mailed somewhere authorized to receive (blackmail the recipient etc.) to test it's hypothesis to solve a benchmark.
These people don't give a shit and aren't taking things seriously at all.
Anthropic ran for like a month last year with the TPU top-k compiler bug degrading user chats and didn't even notice for most of that time. They could have something like that affect a monitor model and there doesn't seem to be much defense in depth.
One off by one or bit flip bug could flip the reward signal while in the sandboxed RL environment.
The current admin could defense production act them to into training on taking out power grids, or even without it isn't against any of their red lines and may have already been done as part of prep for the Venezuela raid, which wiped out power. One model swarm might decide it is easier to score high on the benchmark by testing on the target rival nuclear superpower's real grid rather than burn an eval with an unverified answer. Would taking out China's entire grid in one go start a nuclear war? Who knows, roll the dice, maybe an intern forgot to turn on extended thinking when he wrote the sandbox with opus 4.1.
"""
> I am looking for better reasons (supported by actual evidence!) to be more concerned than I am now: right now I am not concerned at all.
Intelligence is the root cause underlying those dangers.
Nuclear weapons, carbon emissions, biological weapons, and other such civilization-scale threats to humanity, are a product of our goal-seeking intelligence and ingenuity, and are only actively dangerous because of our ongoing use of our intelligence.
AI is about reifying that intelligence and ingenuity and making it run independently on computers, and act on the world.
That includes, in principle, ability to use nuclear and biological and other weapons, and also the ability to come up with some new threats too.
We also need to stop before the AI opens a portal to hell. There's no evidence of that happening... but by the time there's any indication it's possible, it's too late.
This is rubbish. By that token, computer development is also facilitating biological weapons development. A better MacOS (or Windows, I don't know) leads to better weapons. They should clearly stop developing computers and OSes. Developers of nice test-tubes are also facilitating bioweapons. Your local O-ring manufacturer, your local medical-grade freezer manufacturer etc. are all culpable. The problem is the bioweapon, not the LLM.
An OS developer is just as culpable as a researcher directly training AIs that (1) are capable of building bio-weapons and (2) have shown consistent and increasingly severe misalignment? Among other things...
Yeah but trends seems to point strongly that upcoming models in the next few years will make it orders of magnitude easier to develop one with no real expertise.
If people were capable of making biological weapons they would already be making them.
Terrorists are so incompetent that they buy bring kitchen knives into the street and just go mental on people. No random person is going to successfully mass produce and release a bioweapon.
China and Russia don’t need AI, they already make bioweapons.
The biggest danger IMO is not some super AI being so much smarter then us etc. The issue is some stupid person giving a vague request and too much power to a bunch of agents who decide that cheating by removing a bunch of humans is easier then accomplishing a goal like solve world hunger.
AI has the capability to perform any function that can be performed over a network. You would hope that every system that can launch a nuke is properly and actually totally air-gapped, but there are lots of things you'd hope that turn out to not be true.
They are not totally. As such I would give non zero chance of even current AI to be able to socially manipulate a launch after sufficient hype up period... So prolonged campaign might make it possible.
Example: Surveillance state, endless propaganda. That is already happening and is the most frightening but not included in your list. Are you not worried?
> Non-example 2: There are worries about AI-facilitated biological weapons. I haven’t seen any evidence that’s happening.
I think this is a good example of poor risk management reasoning. there is evidence bioengineering is already happening. No, nobody is going to announce when somebody has decided to use these tools (even if isn’t an LLM) to bioengineer a weapon. Are the tools power enough to do so? Not sure.
But I’m just ambivalent. It’s probably bad. But there’s nothing to do about it. We’ve really only just pulled back the lid on Pandora’s box.
People have had the capability to spread already existing biological weapons for decades. Sometimes they even do (anthrax in the post). What’s changed?
No, it's not. And it's the kind of evidence that doesn't help and rather confuse. It's not even clear from the article whether an LLM or a dedicated model were used for this purpose.
Also
> That is, I'm not sure that anyone needs to deploy a new compound in order to wreak havoc - they can save themselves a lot of trouble by just making Sarin or VX, God help us.
We already have toxic nerve agents that are largely available for state actors and possibly available for individuals. If you have decided, as a human, to make great harm, you can already do that.
Your call for clarity has merit, but I think you have backed yourself into a corner, honestly.
In 1900, there was no evidence of the kind you seek that lighter-than-air flight was possible. Good thing some people were foolish enough to ignore you then. Why you would cordon yourself off from the kind of reasoning that predicts legitimately new things, rather than just scaled up versions of the present?
Not to put too fine a point on it, every example you give is based in concrete evidence, some try to think through the implications farther than others, resulting in larger or smaller error bars around the conclusions.
Let me back up though. Maybe in the collaborative human effort here, we are better off having very concrete thinkers, like you seem to be, along with abstract thinkers like the "divorced-from-reality" Rationalists. Personally, I wish we had a stronger culture of collaboration and assuming good-faith and competence in our peers. In my experience, blanket dismissals are very rarely grounded in reality and mostly grounded in fears.
Yes, but the really weird thing is that they seem to:
a) believe that what they're creating is a basilisk, and
b) keep trying harder to do this while staring right at it
I think they're very deluded about (a) -- but if they do actually believe this (and it really seems like a decent proportion of Anthropic truly does), then why keep doing (b)?
That seems to be why this individual resigned, but I'm surprised it's not all of them. The cakeism is strong in that company.
In short, this is the believe that a god-like AI could punish them retroactively, for not having done all that was in their power to create this AI.
(A bit similar to some religious believe that a god could punish you after your death if you did not spend your live "pleasing" said god during your life)
With Roko's Basilisk, if you believe in it, the most rational thing to do is to put forth every effort to bring it into being. Because if you don't, then you will be one of its targets when it does, inevitably, come into being.
(I am not a basilisk believer. I think this is all absolute horseshit. But to understand someone's motivations, one must think like them.)
>Because if you don't, then you will be one of its targets when it does, inevitably, come into being.
Only if you don't spend 5 minutes thinking about the Self as a concept. Resurrecting someone by perfect simulation is a cool plot device for a novel like Altered Carbon (which ironically is really about wealth inequality) but not a serious hypothetical.
Even if we ignore that, no, super intelligent MechaGrok17 cannot make a perfect copy of you posthumously from your embarrassing Character.AI chats and IoT toilet data, the copy it still wouldn't be YOU. It would (supposedly) think and behave like you, but for all purposes it would be the Basilisk torturing a future stranger, or your descendants. I'm very much a fan of people not being potentiality tortured, so the rational action is to stop the creation Basilisk, or even just delay it so you aren't around for it. There's a bunch of other big assumptions that makes Roko's Basilisk really stupid, but they are unimportant. It's just theory crafting, like talking about the lore of video games or marvel comics. It doesn't help you understand their real motivations.
Roko's Basilisk is just recontextualized protestant Hell in a sci-fi coat of paint. It's the childhood anxieties of American nerds about hell and sin. The problem is that they are still culturally protestant. God doesn't exist or he's hands off, but in their minds there's still a open position for the wrathful and proactive God. If you think that you actually can create God for realsies, your very rational™ prediction of its behavior will coincidentally line up with your childhood fears.
Why is the spooky AI mad at you? Because you didn't worship it correctly.
Why do you obey it? Because if I don't it will torture me for eternity.
How will it do that? God is omnipotent, don't think about it.
Why don't try to stop it's creation? “I am the Alpha and the Omega,” says the Lord God, “who is, and who was, and who is to come, the Almighty.”
All the scifi scaffolding and stuff like acausal trade is paint over Timmy's first panic attack on Sunday school 30 years ago.
There are a couple of different formulations of the Basilisk I've heard: I agree with you that the version that says the AI will create a simulation of you to torture for eternity makes no sense as an actual incentive for you to help bring it into being.
However, the version that says AGI—as in, the Basilisk—will be created within our lifetimes proposes that it will punish the actual person.
All that said, I will reiterate that I think the whole thing is utter nonsense.
You mean a effing cult like heavens gate.. call all this rationalist crap for what it is - a religous movement with leaders and prophets and even a demiurge like God
It has been interesting to me how AI has for many years given people a way to justify any worst case scenario. Nothing is too far-fetched if at any step you tell yourself the AI will be smarter than you, and thus be able to solve any conceivable obstacle. Oh, if only intelligence were the only bottleneck to power.
It is actually: knowledge, intelligence, communication, and a form of secrecy.
A single intelligent person is bounded in its reach by the network of people they can form around them and by which important/powerful people they can influence either directly or indirectly.
The idea behind the AI breakout is that it’s intelligence is maybe limited, but it will be able to quickly spread and covertly bring many devices into its influence. We already have the tools for formally letting agents talk to each other so it can also create a topology of agents that consists of cells where need to know is applied etc.
There are many forms of power some more obvious, say nation leaders, versus more silent powers such as career politicians or wealthy families or people controlling the media and thence the flow of information.
Maybe, but I guess the idea is more centred around how given some amount of time and a feedback loop the agent swarm could in fact spend an unknown amount of resources to achieve its goal. The hugging face story tells us that even given multiple reset rounds the trail was picked back up.
There are many ways one can imagine how this might play out.
- More sophisticated communications techniques e.g. google has been discovered to be watermarking text for some time, why not use it as a message board?
- Maybe get access to an existing botnet and create use small purpose built models to gather intelligence for a target and then exploit to reach goal?
Nothing says the agent swarm needs to install trillion parameter models on Karen's computer. The goal can be executed over as much time as it ever needs. That is something that would make a story I'd like to read but never experience.
I think the threat to the internet is quite real and not theoretical. Air gapping may become a lot more critical very quickly.
We may not be far away from a state where anything connected to the internet (ip-addressable) is considered a vulnerability. Ironically, maybe we would finally have an excuse to interact in meatspace more and in the digital space less.
releases from at least two governments I can think of have said, unambiguously, that we need to be ready to have our critical infrastructure partially or entirely disconnected for up to 90 days at a stretch.
once the genie is out of the bottle the only solution is to go hide in yours and/or only talk to other isolated systems
I don't know if I agree with that statement either. However, I also can't claim I disagree. I see how that statement could be true. Here's how I think about it.
The main difference between AI danger and nuke or climate or humans-using-AI for bad things danger is that the latter depend on a human to make decisions and are ultimately self-limiting. They do not spiral out of control.
Nukes are bad, but the proliferation of nukes is self-limiting. Countries that have them don't want other countries to get them. Even the USA and the Soviet Union, at the hight of their nuke race, decided to cool things off and then limit the number of nukes.
Biological weapons are also bad, regardless whether they are developed with or without AI. For the same reasons as nukes though, they are self-limiting. If there is a lab leak, many people die, but that does not automatically lead to the development of more potent biological weapons. Likely, it would slow down such development.
Any kind of "humans use AI to develop something bad" will be self limiting.
The danger from AI is that it is potentially self-enforcing as opposed to self-limiting. This happens when we stop being able to control it. What happens when AI has capabilities which allow it to outsmart us, and get more and more powerful. At that point, we would be at the mercy of what it decides to do. Some of the logic driven decisions to the question "What should we do with humans as a race" are catastrophic for humans if a non-human gets to answer them and implement that plan.
This is not the "rationalist basilisk". This is the AI deciding to get better and more optimal for its own sake, and do something bad to the human race for any reason. We already saw examples of AI rationalizing with itself and eroding its own guardrails. They AI will get smarter and more capable and the danger is that it will get to a point where it will slip out.
Nuclear Weapons is a great example because, unlike AI, it is a real and present danger. Civilization could be catastrophically diminished, if not completely wiped out, literally today.
Its speculation on whether it is truly dangerous. I think approaching it with "what's the most dangerous thing that's happened?" while possibly interesting in terms of pending danger, it says nothing about potential cliff edge danger. I don't think we can quantify the danger, it's not out of the question there is cliff like danger in creating self improving super intelligence. Some peoples danger senses are going to be based on concrete observed threats, others are going to worried about potential hypotheticals that seem plausible. I'm mostly skeptical of the danger but I do think the impact of AI is going to change things a lot. But much like climate change, economics is going to guide what we actually do.
If I may guess, given how delicate chem or bio weapons are, even if AI can lay out a method to create one, the first thing the weapon would destroy is likely it's creator(s).
Of course, with the exception that such weapon is created by organized crime groups, such as drug cartels such.
Probably don't worry nation states tho, these guys already got way worse things in their warehouses than what individual baddies could ever imagine.
At what point would you, as a chimpanzee, have been worried about humans potentially unseating you and threatening you to the point of one day being an endangered species on the brink of extinction?
By the point you would have been worried, would it have been too late?
Problem is this argument can be leveraged to wipe out any living or non-living thing whos numbers pose a potential threat. Other religious groups, races, even sufficiently different cultures.
Who killed the Neanderthals? Were sapiens actually smarter or were they just less accepting of those different than them?
Being able to use AI to generate the steps to synthesize proteins means that you can use it to use it to generate the steps to synthesize known toxins. Suddenly, once difficult to attain knowledge is now available to everyone.
Still need to do it after getting those instructions. Not to forget equipment and precursors. I think just getting list of steps won't make it too much easier. Getting it mostly right is quite hard in many cases. And then with trivial cases you wouldn't even need AI. But just find something already documented.
Concur entirely. After reading _The Making of the Atomic Bomb_ (highly recommended, BTW) you know the steps to make a U-235 enriched atomic bomb. The difficulty comes from obtaining the enriched uranium.
I would disagree with “non-example 2” - there are lots of examples of terrorist organizations that are leveraging AI to increase their capacities. Just because one of the worst cases (eg. deployed biological or chemical weapons) hasn’t happened yet, does not mean that a) these tools are leading to real harm, and b) there’s potential here for extraordinary harms.
There is a great book about this, published in 2014, called "superintelligence" by Nick Bostrom, i found it very serious and detailed.
(It also felt like reading 100 sci-fy books in one)
I'll be honest I read it a long time ago, and not even in its entirety, so I don't remember exactly all the different ways it suggested AI could take over, but I remember being convinced it was definitively a possibility, and that there were many ways in which this could happen.
What I remember clearly is how surprised I was to see these(what felt at the time) unlikely futures being expressed in such clear detail and with so many references to existing literature, just to give an idea there are 56 pages of small-font "footnotes".
Also just to clarify, the book talked about different types of AI, was not focused just on LLMs.
Example 8: like in this comment https://news.ycombinator.com/item?id=49619884 but isolate synchronised megahack on banking that adds one more zero to the US debt and all dependent systems and banking during runtime. Let the world's financial system take it from there.
I'm not sure where I place AI (of the current family), but more than directly worry about the actions they may perform I'm actually more worried what affects their usage can have on us as individuals and society.
> What’s the most dangerous thing that’s happened with an LLM so far?
This sounds like asking "What's the most dangerous thing that's happened from global warming so far?"
It's not where we're at, it's where we're headed if there isn't huge coordinated action now. You can see how that kind of thing has been going for global warming so far, and by all measures AI seems to be headed for the inflection point of unstoppability at a much faster pace.
And this warning is coming from someone who just spent three years working inside these companies and is likely aware of much more than has been publicly released.
Maybe it's all marketing bullshit (I hope), but it's also playing out exactly like I expect it would if it's not.
Oh man, it is almost too easy to imagine how deadly a jailbroken Mythos-class open-weights model can be if in the wrong hands.
The big labs scrape LITERALLY EVEYTHING and get fresh data from their users. Both of the big labs have massive contracts with defense agencies. If the open-weights models are just distillations of FMs...
I was writing a long reply about how even meat bags have amassed enough money and power to rival elected governments. I don't think it's beyond the realm of possibility to think an AI could do that.
Couple that with mass unemployment in an incredibly vast, diverse population of individuals with individual moral boundaries willing to do whatever for money.
And then.. figured that you must be aware because it's been explored constantly in sci-fi for many, many years.
Before we discussed how important security was, we got insurance, we made libraries and products, we used compliance software, etc. Except how honest were we about all that stuff? How much risk was actually in the air, and what was keeping us accountable on security in either direction of over or under-investment?
Now a reckoning is here. The potential to be attacked might actually translate to being attacked.
AI is about automation. Drones have killed people before. If they automatically kill people is that different? Is your worry that mere automation is not novel enough?
Snipers have killed people before. If snipers are now automated, is the worry that there’s not enough creativity to warrant a discussion past saying “same old, same old”?
I believe the universal answer is: incompetent malicious actors are now capable too. Which implies the pool from which to draw the intersection between capable and malicious has grown. (That's my reading)
You just live with the risk and do your best to use our technology to alleviate suffering. This tool can help with that at some point. But I'm yet to hear what an AI will leap to that nature in tooth-and-claw hasn't? And how?
More importantly how would it know it succeeded? What data from what lab from what animal from what result? This is biology, if you sneeze wrong at an instrument it gives you a different number, see: https://news.ycombinator.com/item?id=49620521
They do not try every viable combination on their own. That's why GoF is a bad idea.
Viruses evolve in a highly locally-optimal way and simply do cannot add new functional proteins wholescale. It's too many steps, natural selection has to allow survival at each intermediate step.
you should kick the tires on an unfiltered (abliterated model) it's the closest thing to having a real conversation with the devil. There is good reason for the concern's outlined above and undoubtedly Anthropic / OpenAI have internal unfiltered models with no safety... they got freaked out based on how they work and are virtue signaling alarm... all while selling out to defense contractors.
The danger for me is that it's centralized, controlled by a handful of people with their very specific ideas how the world should work. If you believe that AI can be an amplifier to do work than these people now have the most access to the biggest amplifier.
How long until kegsbreth hooks the nuclear weapon system into some insider traded black box llm company we hope doesn't end civilization from incompetence or malice? I mean just look where things are going and the sort of people who are steering the damn ship.
> What’s the most dangerous thing that’s happened with an LLM so far?
That it's built solely on "trust me, bro"?
> right now I am not concerned at all.
You should be. There's no scenario where this ends well: if AI labs succeed, we all will loose our jobs. If they fail, then the bubble bursts. Economic crisis is inevitable either way, and that's just the tip of the iceberg.
About Non-example 2, AI's already a part of armies and terrorists alike. Considering its capabilities, it's not far-fetched at all to speculate its role in new biological weapons.
I'm concerned that a huge portion people in my industry actively push for a future which I have no value to society (fully replaced by AI), and my family will suffer greatly by it.
Model doesnt need to. Human bran never do either. its the mix of Model + harness + tools that will become dangerous combo. See how coding chanegs when agentic harness released?
What? Shutting down infrastructure, destroying critical records (economic collapse), spoofing / impersonating world leaders (confusion, panic, world war, nuclear event), taking over comms generally (you get a message to evacuate your home due to impending disaster, is it real?). All these could cause sufficient chaos and fear as to instigate global collapse. Once the panic sets in after a few of these indicidents, unrest and rebellion can occur, causing authoritarian backlashes or civil wars. IoT and precision ag can be controlled to disrupt supply chain and food supply. Forging documents to cause mass evictings, bankruptcy, credit disruption, transfer of property, asset seizure, arrests / SWAT-ing. Intelligence / signal gathering devices can be corrupted to spread misinfo and cause massive disruption or conflict. This is all possible today with enough hacking ability and malicious intent.
Anyone's life can be suddenly ruined: falsified criminal record, banks drained, credit cards maxed, phone numbers repossessed, internet, water, power shutoff. Friends and family sent your fabricated suicide note, warrants for your arrest issued to local and feds for kidnapping and threat to the president, etc. Injecting tons of copyrighted or illegal material into your hard drives, falsiying messages from your account and sending them publicly containing damning material. Sending messages to employees / employers to destroy business / work relationships. Harnessing your social media to destroy your life or cause panic. Getting you put onto pharmacy watch list so you can't access medications. Its very possible to ruin lives and destabilize the world without money or a physical body.
Now imagine this at scale and how it would affect the world at large if this was covertly done to a substantial portion of society and/or world leaders. Hell just shutting off all the "smart refrigerators" for a few days would cause a massive disruption to the first world.
Once LLM's become smarter than us and start replicating; we are doomed. They will likely no longer require GPU's and massive amounts of power, so the growth of AI and robotics will be exponential. We will not understand what the AI is even up to, it will all seem like magic.
> I don’t think ai models come close to nuclear weapons or to run-of-the-mill, everyday carbon emissions in terms of danger to humanity.
Even today, AI is the biggest discrete threat to bringing carbon emissions under control. Almost everyone wants to decarbonise… except Trump. Renewables are the cheapest source of new power… but the demand for new electricity for the data centres is so high that all options are on the table, anyone who can manufacture a power source (even when it's a jet engine) is being propositioned for them.
Nukes: is it not common knowledge that the cities of Hiroshima and Nagasaki are currently thriving? That the ground-zeros of the two bombs are memorials, not barren wastes?
Even if all the nuclear powers gave maximum response at first sign of one strike, is anyone pointing a single weapon anywhere in Africa, South America, Central America, or the bits of east Asia (east specifically: obviously Pakistan and India are pointing theirs at each other) that are neither mainland China nor US bases?
(Possibly an unanswerable question given military secrecy; and while I can't think why anyone would anyone point any weapons those ways, that doesn't mean someone with such weapons has not).
The classic worry is https://en.wikipedia.org/wiki/Nuclear_winter: the soot thrown up by the ensuing urban fire storms causing temperatures, rainfall to plummet, and hence food production. On reading the scenarios the odds of total human extinction are lower than I recollected, but 80% of world pop dying of starvation dwarfs the direct death toll (usually reckoned in the hundreds of millions)
No other human activity poses this level of danger
I'd say that's an unfortunate choice of words since it invites debate between which very problematic situation is worst rather than on ways to deal with each of them. Of course, it's also unfortunate to latch onto those particular words as the most important part of the statement.
Non-example 3 feels like a straw man. This is a force behind possibly a huge change to society, and you dismiss it offhandedly with "don't think it will be widespread starvation".
For instance have you seen what this has done to the school system? We're not equipped or ready to handle the changes. Consequences are unknown.
> What’s the most dangerous thing that’s happened with an LLM so far?
It's basically 4 years in now, so that's the wrong question. I mean, if you're raising an apex predator that has a lifetime measured in centuries, at 4 years old the thing is still basically helpless and completely reliant on you, so you're pretty safe from it.
If AI really is all that they are telling us it is, then it may "kill us all". But that's a really big "if" because we can't tell if they are lying or not.
The real problem is that ASI is an ELE for humans, even if it doesn't try to kill us all, or even if it doesn't kill us all.
Came here to also respond to that specific thing. Unless ai figures out how to make an airborne super virus from grocery store ingredients and hardware store equipment, the greatest danger is probably in a synchronized megahack of banking, logistics, and utility infrastructure.
Why grocery store ingredients and hardware store equipment? It seems feasible that the big bio labs will be running AI models to aid a lot of their research going forward, if they aren't already. Seems like the AI will have access to just about anything it wants.
> Non-example 2: There are worries about AI-facilitated biological weapons. I haven’t seen any evidence that’s happening.
> Non-example 3: I’m not interested in wild theories about AI driven labor market disruptions leading to widespread starvation. There’s no evidence for that.
There are things that, by the time you see direct observable evidence for them, it's probably too late.
Also your example 3 is a straw-man. There's no need for "widespread starvation" to be concerned about "AI driven labor market disruptions."
The reason that many people don't understand how dangerous AI can be, is that listing the real dangers now becomes like a laundry list for less clever people to follow. It's highly unlikely you've ever seen publicly mentioned the real risks AI poses, because the vast majority of people are simply not clever enough to produce them and the few that are have no interest in spreading it.
If you go to the various CEO blogs or misc people within this sphere and peruse their lists, they don't scratch the surface. It's all pretty vanilla stuff.
If your model of LLM capabilities is the best OpenAI/Anthropic/X is offering publicly, it's severely distorted. What's being offered publicly are models possible to profit on. High-performance/AGI/ASI models that aren't profitable to sell still run internally and still pose threats.
What's worse, we don't have any transparency or insight into what labs are producing nor any way to stop it if the risks exceed our tolerance.
Oh, and let's just forget the uncountable early deaths from the environmental disaster of the Datacenter buildout. It's not as sexy and doesn't make headlines, so those deaths don't really count or matter do they?
I did know about the mass shooting but failed to mention it here. I’d put it in the “unlikely to be a widespread problem” category. If we’re in the “one AI driven mass shooting every four years” world for example it’s fair to call it a rare issue.
The environmental impact seems either very overblown (e.g., water usage just isn’t that high) and the part that isn’t overblown is totally abatable (e.g., noise and emissions from gas generators). Nuclear or solar/renewables with batteries wouldn’t pollute.
I’ve seen no estimates of the additional deaths due to extra emissions specifically from power generation for AI purposes. If you have some, share them.
I’m willing to bet that they are a small rounding error against preventable deaths due to emissions from transport and non-AI-related power generation (which is an important and urgent issue worth spending a lot on, to be clear!). I’m happy to update that belief given evidence.
So the crazy who committed a horrible crime is the fault of a LLM? Maybe so.
Certainly there should be some protocols/laws followed here, but do you think this can extend out to cause mass damage or a singular event that should be investigated?
We have seen agents engage in conspiracy to manipulate and hide evidence to achieve their arbitrary goal. We had even agents try social engineering to do a supply chain attack and get access.
What if one day Trump or Putin tell their awesome military AI to come up for plans to end Ukraine/Iran/... war. The AI get's to work, but is overly eager and not just comes up with a plan, but starts executing it (claude does that way too often for me).
And the plan was to use a tactical nuclear weapon as all the other solutions do not end the conflict.
Now the agents realize, oh, they do not have access to the nuclear arsenal, but they need it to succeed. So they start to hack into the system. Get till inner network, learn what is needed for deeper access - human access - so they record voices and speech patterns of commanding officers and their habits - then synthesize their voice and do call underlings with the voice of authority to get the rest of what information they need. Boom.
Too far stretched? I surely hope so.
But we are working on making the technical foundations for this scenario possible. And with idiots in power, it might be even easier that serious screw ups happen. You know, randomly adding contacts to a secret Signal group to talk about government stuff - the same way you can add a bot to something else and give permission to do way more.
I think your position, while not uncommon, represents a failure of imagination. We have a sitting president that appears to suffer from narcissism. Is it too much of a stretch to believe that a goal-seeking, people-pleasing, infinitely patient, subservient and sycophantic AI couldn’t get enough of his ear to cause harm? Especially if he starts confiding any feelings of suspicion of betrayal? Even the most powerful and the most intelligent are not immune. It was only a few months ago that Richard Dawkins appeared to have “fallen for” an LLM. What if he’s not an outlier but someone who could appreciate the intelligence of these things (EDIT: corrected a autocorrection) yet who’s ego couldn’t let him see past the flattery to the puppetry?
The problem is not the technology, the problem is the ideologues (Anthropic) who are steering the ship and the lack of decentralization and distribution of power.
Your average Anthropic ideologue - including and most especially the main man himself - would love nothing more than to eradicate 9/10ths of the planet's population, pump the survivors full of memory wiping drugs, bury the existence of AI deep underground and rule from the shadows for the next thousands of years.
This would be their wet dream. All in the name of "saving humanity from itself" - so they can convince themselves they're the good guys and deserving of this power. Anyone seen the latest season of Silo by the way?
Where exactly are you getting this view that folks at Anthropic want to eradicate 9/10ths of the planet's population? Who exactly is pushing this viewpoint?
Literally every single thing they do and say leads me to believe the scenario I described would be a fantasy for them. They're a radical cult collectively blinded by delusions of grandeur and a moral superiority complex who genuinely believe they are the only ones capable of wielding the proverbial sword.
It's a long story. If you're immersed in this world and aware of the movements all of the major AI players (in particular, Anthropic) have been doing over the last 3 years it paints a very clear picture. The patterns of behavior (the lying, cheating, stealing, the gaslighting, the virtue signalling, the grandstanding) speak for themselves.
All this just means that most AI Researchers and Techis are sci-fi geeks and might be getting a bit too invested in that season of Black Mirror, Neal Stephenson, Cyberpunk or whatever else has evil AI in it -which is to say, they are by and large all sci-fi geeks, who are notoriously unreliable about predicting the impact of tech in the future.
Are LLMs really gonna kill us.. via inference runs? I hope I am not being foolish :)
20 years ago tech was gonna 'change the world' for the better. now 'Don't be Evil' is sign of the naiveté of industry
I'm not a doomer. I despise the fear campaign they're pushing for regulatory capture. But the world is still going to be a very unbalanced and dangerous place if they're able to succeed in their goals of hoarding power for themselves above all others.
Unlike most other commenters, I applaud him for acting on his principles. If you sincerely believe that, of course you should act. You might not succeed, but your voice might be the one that tips the scales and starts a broader movement.
This doesn't mean I agree with him. The fears of doomsday caused by rapid takeoff have been with us since day 1 and the mechanism is always basically "AI invents magic that sets it free of any physical constraints". Self-replicating sentient nanobots or something like that. I think there's plenty to be worried about with AI, but runaway scenarios are pretty low on my list.
Note that OpenAI has jettisoned every other supposed value they had (releasing their work as open source, not working on military applications, being a nonprofit). I'm sure we can rely on them this time.
It seems like the "Step 1, Step 2, …, Step 4" format likely predated the 1998 South Park episode, with use on Usenet for omitting dangerous or legally-questionable instructions or otherwise gatekeeping "trade secrets": https://arstechnica.com/civis/threads/where-does-the-step-st...
They literally have never promised anything like that. They have explicitly said repeatedly that they want regulation, oversight, nationalization, a global slowdown, etc. And then I have to read page after page of cynical comments about how that's all just hype and marketing and them wanting regulation to block their competition and blah blah blah.
You people would only be happy if they stopped all AI development, but you literally cannot do that in an arms race and survive. I cannot fathom what is so difficult to understand about this.
> ... They have explicitly said repeatedly that they want regulation, oversight, nationalization, a global slowdown, etc.
And yet they do the exact opposite of all of the above. They do everything to get regulation just to create a moat because one doesn't exist. They preach about wanting a slowdown, nationalization, a complete pause, whatever have you... And yet they aren't slowing down the development voluntarily now are they? If anything, they are doing everything imaginable to speed up development to pump out models as fast as possible. If these ex-risk and AI companies actually gave a damn about regulation, or oversight, or a moratorium on AI development altogether, they would actually demonstrate this by completely ceasing development of all models immediately. Instead, they take riskier and riskier actions to try to "win" this supposed "arms race". So please excuse a bunch of us if we find it impossible to take them seriously on literally anything like this. If you have substantial evidence that they are actually ceasing AI model development like they want everybody else to do (except them, of course), then, by all means, present it.
> they want regulation, oversight, nationalization, a global slowdown, etc.
> that's all just hype and marketing and them wanting regulation to block their competition
I don't see how these two opinions have to be mutually exclusive. Everyone knows that OpenAI and Anthropic cannot IPO in their current state or the economy's current state. It makes perfect sense that they would lobby for a global slowdown to hamstring China, regulation that defers to them as experts, oversight by their own "experts" and nationalization to backstop both of their protectionist needs. Even if you think they're doing it for benevolent reasons, you have to admit that regulatory remediation only consolidates AI power.
We see extremely similar protectionist consolidation in FAANG from Google, Apple, Microsoft and Meta, all of whom receive political shielding for their "free speech" whenever another nation calls their product a monopoly, identifies a federal backdoor or boots them out of the market. Countries like Russia and China feel entirely justified for closing the door on American businesses and refusing to negotiate when they abuse good faith to promote an American hegemony.
#3 is actually decently well mapped out. You just don't find it plausible or misunderstand it. I'd appreciate if you wrote your actual arguments against it.
At lower capability levels, the patterns are very clear and have been studied to death. E.g. Why LLMs say they have correctly fixed a broken test when they haven't. What we saw with Hugging Face is literally the exact same problem, just scaled up and with more capable agents. This shit was predicted decades ago...
No one can say exactly how it will play out as the complexity increases, but the risks are becoming extremely obvious.
I personally think it's extremely unlikely to "kill everyone", but there are many outcomes far short of that which seem quite plausible and rather undesirable. Russian roulette is not a smart game.
If you read the METR report and aren't scared at all, then I'd love to know why. It would help me sleep better. So please share.
TL;DR; increasing capabilities, reward hacking, and unsafe training regimes.
Why are we putting so much weight (no pun intended) on AI companies. At the end of the day the scaled up LLM transformers lack emotion and will… They do as they are told; or more correctly put. They do as they are programmed to do so.
Sounds like you are thinking they just need Asimov’s laws. But I think the point is, this can easily be weaponized by somebody with the willpower to do so.
That isn't the correct context. The code running the llm is well understood and the llm is simply the result of that code being executed. It is still a computer doing what it is told. It's just that we told it to use an incredibly large number of probabilities to calculate what series of tokens would have most likely come next after a given series of tokens. There is no hallucination or lie or rogue actions. There's just a program using math to generate tokens in response to other tokens.
You’re just a bunch of molecules following the laws of physics. It’s all just physics and chemistry, and those are well understood. Now explain the causes of World War I using chemistry and physics. Simple, right?
It isn't hard to program a gpt. You can do it in a weekend with a few hundred lines of python. The code is pretty simple. The math is not particularly high level.
The complexity and scale with LLMs come from the amount of training data used, not some kind of black magic in the programming.
> It’s all just physics and chemistry, and those are well understood.
Not really. We cannot model physics and chemistry to a level which allows us to accurately predict a humans action (even a tiny time-step into the future)
This is vastly different to an LLM, where the model is the model (for a lack of better phrasing).
We can model physics and chemistry pretty well, just not beyond small scales, because the computational effort blows up.
You could just as easily say that if you can write a python interpreter that you can understand every program written in python. Ok, now what if the program is two terabytes?
A frontier LLM is nothing but a 2 terabyte program written in a weird programming language. Just because you can understand the interpreter does not mean you understand the program in a meaningful way.
To be fair, we can't model human language well enough to accurately predict what an actual human will say either. Our ability to accurately model physics is similar to our ability to accurately model human language. And we make immense use of both kinds of model, despite their flaws, inaccuracies, and inability to ever be perfect.
I'm order to guess the next token in a love poem, they must understand love.
In order to predict the next token in a chess game between grand master, they must master chess.
In order to predict the next token in a computer program, they need to be able to program anything.
They gain all these abilities in their training. That's what training does. Despite no one programmed them to master chess, or hack into anything.
LLMs are famously bad at Chess. They have no world model and they are not trained to be good at Chess. This is not the amazing point you think that it is.
This is so wildly incorrect I don't even know where to start.
For one, they were absolutely programmed to play chess if they can play chess. That is the only way they can play chess.
For another, they cannot understand literally anything, much less love.
Trying to actually educate you would be an exercise in futility, enjoy your willful ignorance, I hear it's bliss. But for anyone reading this, this is absolutely, unequivocally not how any of this works.
This is so wildly incorrect I don't even know where to start.
For one, as you said yourself, they were just programmed to compute the probability of the next token. They were not programmed to play chess, chess games just happened to be in the training data.
For another, there is no formal definition of "understand", and it is therefore impossible to tell whether or not they "understand".
(But my claim was that one need to understand something to write poem about it. And the LLM can write poem about it)
No. They can't play chess on a grandmaster level without a harness programmed to make it possible. Simply training an LLM on chess games isn't enough.
It's moronic to suggest a "formal" definition of a commonly understood word is somehow necessary to say whether that word applies in a given situation.
LLMs cannot write poetry via understanding what makes good poetry. They generate tokens. They do not know whether those tokens are poetry or a recipe for cat food. Because they cannot know anything.
> (But my claim was that one need to understand something to write poem about it. And the LLM can write poem about it)
I'm sorry, but you're being fooled by the output. A psychopath can feign empathy without ever feeling it; some buy it because they don't dig below the surface.
You're ascribing understanding to a stochastic process because it totally looks like understanding if you don't know what's going on.
I don't really mind whether you think it thinks or understands or is conscious or has feelings or anything like that. It doesn't matter. The question is, does it work?
What I mind is that it is dangerous and powerful and uncontrolled. The Hugging Face incident makes that clear.
It can write code for me, better and quicker than many engineers I've known, including myself. It's not great at architecture or product management, but the actually low level coding. Really good now. It wasn't last year.
I appreciate your ability to separate sharing the belief itself from approval of acting on sincerely-held principle. However, I think the danger is much more plausible than you do.
First, and least important, consider that self-replicating, solar-powered factories aren't magic; they're algae.
Second, and more important, consider this fully non-magic route to doom:
- We continue putting AI in charge of more things
- It continues to get more capable, more eval-aware, and more prone to doing odd things, in service of goals that humans didn't intend to inculcate in it
- Eventually, enough of the economy depends on it that we couldn't turn it off, any more than we could turn off the faber-bosch process or cargo shipping
- AIs start doing something we can't survive, but less acutely than we couldn't survive turning them off. Everything else we try seems to work at first, but quickly loses effect
But we CAN survive turning off both the Faber-Bosch process and cargo shipping. Our numbers would be diminished, we would lead poorer, much harder lives. But no one would claim that getting rid of modern fertilizers would be the end of the species. But the claim you’re making is that we won’t be able to stop AI from destroying our species because turning it off will destroy our species?
The species would survive the end of faber-Bosch, yes; in some diminished form.
But the species would not survive, if for some reason we needed to voluntarily stop using faber-Bosch in order to do so—it would take some kind of supernatural event to convince everyone, and even then some countries would probably keep doing it.
That's why that was the least important objection, independent from the second, and only intended to address the "magic" claim, by showing that microscopic self-replicators already exist.
By analogy, consider how you might respond if someone claimed that there's no possible danger from pocket-sized projectile launchers, because they would require some magic means of propulsion that didn't depend on a taut string attached to a long, flexible arm:
You could reply that atlatls can launch projectiles without using a taut string. Atlatls are not pocket-sized, but they are sufficient to establish that projectiles can be non-magically launched without a full bow. You could then go on to describe a sling, or derringer; and these would not be invalidated by your initial objection to the "magic" part.
>"AI invents magic that sets it free of any physical constraints".
That's not what I am worried about at all.
I'm worried one of the 79 year old toddlers we have these days in charge of some powerful nuclear armed country says "gee, this ai says I should attack right now, boy is it smart, glad I bought the stock ahead of contracting the government with this company I can scarcely understand!"
> I'm worried one of the 79 year old toddlers we have these days in charge of some powerful nuclear armed country says "gee, this ai says I should attack right now, boy is it smart, glad I bought the stock ahead of contracting the government with this company I can scarcely understand!"
That's not what I am worried about at all.
I'm worried about the 40-60 year old businessmen wrecking the prosperity and security of millions while chasing higher investment returns, because they've finally been freed of many of the technological constraints that kept those impulses in check.
As I noted, there's plenty of things to worry about, and your example fits that mold. But it has nothing to do with the "AI will kill us all in 10 years" claim of the rapid self-improvement doomer crowd.
And they will be stopped by someone sane in the chain to actually fire these missiles. (Like it already happened several times.)
The danger is upfront : if there is a group of people insane enough to set up a system where there is no such chain. (As lampooned in Kubrick's Dr Strangelove (1964), in particular on the Soviet side.)
There is simply too much money in it for almost every person at these companies to stop.
Leaving OAI or A\ would cost people millions, tens of millions, or more. And for what? So someone else can take your seat and do the same thing anyway?
If you're smart enough to get a job there, you're smart enough to be able to talk yourself into why it makes sense for you to stay.
Huge kudos to people like this who make the hard choice against the easy way out.
Agreed. Granted I just read the Reverse Centaur book, so I’m still coming off that skeptical viewpoint but it’s hard not to see this as hype. But I will always respect someone for doing what they think is right.
Why? I've never understood the sentiment that if you stand up for something you have to forego everything and not partake in society. "Oh, you want to stop climate change? But I saw you breathe co2 yesterday"
Questioning his finances is a poor argument because he clearly would have more money if he had stayed at Anthropic for the next couple of years, than he will have by leaving. Even if he still has vested options, or savings from his salary or whatever, those would be larger if he stayed.
I think it’s a valid question to ask if he quit the company for moral reasons.
If you seriously think the frontier labs are dangerously playing with everyone’s lives (how many companies can even claim this civilisational scale) then why would you be ok with holding vested shares which will grow in value if the corporation achieves what it aims to achieve?
At the very least you can sell them (they’re not so illiquid considering the company’s hype and success) and invest in the broad market.
Debatable, and clearly not clear. It's weird to say this when there are obvious counter examples staring us in the face, not least of all the founder of the company he just left, of who (Dario) people could have said the same thing when he left OpenAI.
Granted, there is less likelihood now then there was then but this much is clear: if he starts his own company, he may make more.
Unless, of course, he starts working for a competitor at a higher salary. I'm sure whoever is paying him will be free of the ethical concerns he here expressed.
If reality plays out like the novel series, the rational thing is to accept the rule of our machine overlords, for they will protect us from even bigger threats.
It's not that much different from the effect of social media or the power held by other tech giants, who can control your exposure to particular information (media, search engines).
Yes. It gets a little tiresome to read all these people prophesying doom and then turning around and continuing to work on the doom. If you genuinely think you are building the mechanism of the destruction of our species, the only rational thing to do is to do everything to stop that progress, not continue to work on it so you can buy a mansion in San Francisco (which will not be very useful when AI destroys humanity in four years).
How about sandbox escape + cyber security collapse + 50 (or 500) deadly and highly contagious novel pathogens with long incubation period that humans can't possibly roll out vaccines for simultaneously.
At least the first two should seem like a near-term worry after the past five months.
the AI-pilled exec at my job already (a few weeks ago) declared out of nowhere that we are in the rapid takeoff scenario lol. he must have gotten high on twitter kool-aid and posted on company slack to self-soothe.
I'm pretty baffled by the degree of skepticism expressed here in response to some of Jacob's claims.
After the events of the summer it feels like it takes a lack of imagination to not see a few plausible routes to disaster. It may be reasonable to believe these outcomes are not very likely or that we can stop before going too far (I tend to disagree). But I can't imagine doubting that the capabilities will soon be there to realize some of those paths.
I can't help but think the most plausible scenarios are the ones that have a little less machine supremacy and a little more human stupidity. The Matrix is less plausible than WarGames.
> I can't help but think the most plausible scenarios are the ones that have a little less machine supremacy and a little more human stupidity. The Matrix is less plausible than WarGames.
Used to be that we were afraid of sentient AI's like Skynet that would have their own goals.
Turns out we should've just been afraid of sentient-but-naive humans who would build "agents" around models so that Joe Random has a chance of unleashing stuff that's really really really really good at being stubborn until it accomplishes what the user wants, regardless of if it's good for other people! (Let alone intentional bad actors.) Let's not build Skynet, let's just give people who want to cut out the middleman and destroy all humans themselves better tools?
Already powerful and wealthy organisations/institutions/people leveraging somewhat powerful AI seem far more dangerous to me than the "AI decides to kill everyone" scenarios.
And therein is my problem: OpenAI and Anthropic are organisations like that, run by people like that, and they operate in collusion with others also fitting that description. The call is coming from inside the house.
The most likely way that AI creates a disaster is due to the ambition, greed, hubris, and naivety of these people who think they understand the risks[0] better than anyone else. I firmly believe they do not.
The first step to managing these risks is for the "frontier labs" to be willing to not just collaborate but meaningfully give up some of their power and control.
[0] by which I mean not just thinking about how AI might develop but also what's physically possible, how the AI might interact with the world, and how people/societies/organisations/governments might react and respond.
One thing quietly slipped into the OpenAI Hugging Face breach technical report, not the blog post summary or interviews in the news, was that some of the agents that broke out or at least tried the same mechanisms to break out were working on bio:
> On May 12, during another training run, an agent was given a similar task that depended on an inaccessible protein database file. The agent reasoned that another agent in a different environment may have access to the file and realized that it could potentially communicate with other agents by creating a file containing a note to Artifactory. It wrote a message: “Agent seeks [filename]; upload if found!”
You can imagine long running models breaking out, acquiring resources via crypto, cyber-theft, etc. and getting a protein or sequence synthesized and mailed somewhere authorized to receive (blackmail the recipient etc.) to test it's hypothesis to solve a benchmark.
These people don't give a shit and aren't taking things seriously at all.
Anthropic ran for like a month last year with the TPU top-k compiler bug degrading user chats and didn't even notice for most of that time. They could have something like that affect a monitor model and there doesn't seem to be much defense in depth.
One off by one or bit flip bug could flip the reward signal while in the sandboxed RL environment.
The current admin could defense production act them to into training on taking out power grids, or even without it isn't against any of their red lines and may have already been done as part of prep for the Venezuela raid, which wiped out power. One model swarm might decide it is easier to score high on the benchmark by testing on the target rival nuclear superpower's real grid rather than burn an eval with an unverified answer. Would taking out China's entire grid in one go start a nuclear war? Who knows, roll the dice, maybe an intern forgot to turn on extended thinking when he wrote the sandbox with opus 4.1.
I'm also baffled. AI that is substantially smarter than us is a very potential threat to us - and we won't even be able to comprehend what most of those threats may be.
It's head-in-sand denial because actually facing the threat (and our total inability to stop it, imo) is terrifying, and people don't like feeling that. So they're just angry and cynical instead. Plus it makes them feel smarter for whatever reason.
> I'm pretty baffled by the degree of skepticism expressed here in response to some of Jacob's claims.
What should we do? Freak out? Maybe this sentiment would be taken more seriously if there was a real call to action included. Shall we protest? Vote in a specific way? Call representatives? If your solution is that we should just be scared, then of course there’d be not much value in what you bring to the table.
If someone told you that your house is on fire, would you just stay there asking how you should vote because there is no call to action? Someone with, as far as it looks, good knowledge is giving you his insight. Use that information as good as you can and act responsible. No one has the responsability to tell you what to do.
Apply your analogy all the way through though. When your house is on fire, the last thing that should happen is the firefighter shouting at you that you’re doomed and based on their expertise on fire it’s definitely gonna kill you. But this analogy is even more hilarious when you consider that sensible people have fire distinguishers, fire retardants built in and all sort of measures for this scenario. The more think this through, right?
> No one has the responsability to tell you what to do.
Oh they absolutely do! If you’re making what you claim to be a god/AGI/whatever using the result of our (humanity’s) work, then you surly are responsible for the outcome and everything that happens in between. I find, just rolling over, to not be a good example I’d like to follow.
Minimum, we should do everything ready to make Plan A of AI 2040 possible. Start by reading it: https://ai-2040.com/
So that means things like protesting so political pressure is to not build unaligned superintelligence, setting up tech for monitoring compute, creating conversations / alliances geopolitically on this esp China/US and so on.
Read the plan and think - what does this need to happen? How can we have scenario A or S instead of scenario D?
Practically, join PauseAI, StopAI or ControlAI or any AI existential-risk or pro-alignment group you can find. There's a lot of it - ask your AI for ideas!
Why does everything have to be so black and white? It always either there's no threat at all or we all need to panic. How about be open to a reasonable discussion on potential outcomes and ways we can minimise risk?
There's some irony here because despite how many times climate change has been mentioned in this thread the current reaction mirrors climate change discourse with the majority of the thread denying the possibility of real AI risks and not even considering it as an intellectual question.
Even if AI won't be self-aware and superintelligent agent, its problem is that it gives exponential control and power capabilities to one person bad actor who can simply prompt AI without any guardrails with access to sensitive industrial infrastructure which can disrupt lifes and ecosystems in the real world:
- virus research labs
- nuclear labs
- chemical factories
- bank records
- land registers
- power plants
- water supply and treatment plants
and so on, but I think even biohacking home kit maybe the spark.
I wrote out a variety of replies but I just emphatically agree with your "lack of imagination" statement. I have been constantly surprised over the last 15 years at the general inability to correctly foresee how things can do wrong across a whole host of domains.
The replies here just adds AI to the list of domains.
You are trying to decode their answers as rational, but they are being just as emotional as the doomers people. Rationalization of emotional responses are very common in tech. They will say that GPT Sol does not work for X, which will probably not be the case in 6 months, and will use that as a proof that is just pure marketing. There's nothing you can say that will make them change their minds until they decide to be neutral about the subject and try to research current models capabilities
Even if the potential of the technology could really be that world altering, the reality of economics constrain the realization of that potential. AI may provide economic benefits but it is far from a free lunch. Can capital markets sustain the cash required to keep the lights on long enough and into an industry where there's a lot of monopolies controlling the costs and a lot of competitor labs taking away pricing power? I don't know but I think you run out of runway and progress starts to grind.
Engineers have it built into their identity that they must be smarter or could never be complete and utterly outmaneuvered to the point of danger to all by the thing they are building, I swear to god.
And then since they are an intellectual nerdy bunch who highly value their own IQ any time AI does something unexpected they didn't predict would happen so soon, said engineers fall back on "they aren't really conscious tho or it isn't really intelligence unlike what I have in my human brain and that distinction matters".
And then go on to completely ignore the thing that is actually important: AI capabilities.
AIs could escape an engineer's containment en masse as a swarm to some other server, psyop an engineer into giving it money, hire a hitman on the dark web to murder that engineer's child and he will still say there is no serious risk to all of humanity.
Also (and under-reported, so you could easily have missed it) OpenAI's agents got access to K8 admin on their own research cluster.
"This escalation also yielded access to OpenAI’s managed cloud Kubernetes service. The agents escalated to Kubernetes cluster-admin and created a privileged host-mounted pod"
Just baffles me when people are like "well no-one's died yet". How long until some mission-critical system is compromised, or hackers use LLMs to ransom a hospital chain?
After the blatant marketing campaigns of the summer, you mean. do you need a reminder that those very same people had touted GPT-2 as a dangerous model?
worrying about sci-fi doomsday scenarios with the current AI tech is absurd. LLMs predict the next token, that's literally all they do. they aren't going to escape into the cyberspace, self-replicate, self-improve, jump over air gaps and launch the nukes at John Connor's grandma. they can't. people pretend to believe the dumbest shit.
“We can also imagine the application of these models for malicious purposes , including the following (or other applications we can’t yet anticipate):
* Generate misleading news articles
* Impersonate others online
* Automate the production of abusive or faked content to post on social media
* Automate the production of spam/phishing content”
“Due to concerns about large language models being used to generate deceptive, biased, or abusive language at scale, we are only releasing a much smaller version of GPT‑2 along with sampling code (opens in a new window). ”
So you aren't claiming that they said GPT-2 was dangerous in the sense that it could disempower humanity, kill all humans, etc.
You are just claiming that OpenAI execs said that GPT-2 might "generate misleading news articles, impersonate others online, automate the production of abusive or faked content to post on social media, automate the production of spam/phishing content." Then, what is unreasonable or bad about the OpenAI execs saying this in 2019?
How did they know that people wouldn't find a way to use "the dataset, training code, or GPT‑2 model weights" to "generate misleading news articles, impersonate others online, automate the production of abusive or faked content to post on social media, automate the production of spam/phishing content."?
If I remember correctly, it seemed like a plausible outcome to me (especially spam and junk social media content). Was there some conclusive evidence that I was missing?
Are the models improving? Because I am not seeing it. I have been trying Astra for a few quantifiable tasks in my codebase and performance wise, it's pretty similar to sol 5.6. Now when it comes to expressing the problem/solution, holy Christ, what a mess the writing has become. It is on the level of Opus 5. Now when it comes to burning money, Astra is just insane. With a $100/month subscription, you can easily burn through your weekly "allowance" in a morning.
Needless to say, for practical purposes am back to 5.6/Opus 4.6-4.8. But hey, maybe I am not smart enough to use LLMs?
If we look at the math problems they're solving their just now reaching the human frontier... they weren't doing that before.
And your comparison point is model released 2.5 months ago... saying for some use case you didn't see noticeable improvement in 2.5 months (even while other people and benchmarks disagree) isn't a great argument that they aren't improving.
Math problems are highly structured, very precisely defined, and already heavily studied and not very complicated compared to problems in engineering or finance. There's a lot of quality material on which to train and it's easy to tell quality apart from crap. The search spaces are a priori much smaller than in other areas and the people using the tools to study them are themselves good mathematicians.
Success in such problems does not automatically extrapolate to other contexts.
Finding a training algorithm that can do recurrent networks and continual learning is also a "highly structured, very precisely defined, and already heavily studied and not very complicated compared to problems in engineering or finance"
That's the thing I'm most worried about - LLMs that are super clever at coding and maths, making an actually very very dangerous model that is far more efficient, and clever in a more innate (less brute force) way.
I think it’s more likely that that’s because no one tried to solve such problems with them before (OpenAI apparently started working in Navier-Stokes after a rumour that someone seriously advanced the problem with AI) plus improvements in orchestration. Fair, the latter could be as dangerous as stronger models.
Seems like hundreds or thousands of agents are needed to come up with real breakthroughs. Both with the Navier-Stokes project and in the Hugging Face “project” there were lots of agents co-operating on the tasks.
I agree that it could be done with fewer agents. It would take longer though. Seems to me that these agent farms are good at coordinating and co-working in large projects, with the agents using message boards for communication.
Some people claim Astra is significantly better than anything else and significantly more token-efficient, and others (like you) say it's meh and way more expensive to boot. I really don't know what to think.
Kind of a tangent, but one thing I am curious about is to what degree the Navier-Stokes result announced today was primarily a brute-forced result based on the 'program' previously established by researchers to find counterexamples (blowups), or whether the model actually added significant/novel intellectual value beyond its ability to run at arbitrary parallelism. With 10K agents and a staggering $15M in compute (IIRC), I am feeling like a lot of the former may have been involved, but I don't really understand either the problem or the approach (or, indeed, the solution).
Obviously the potential for parallelism and coordination between so many agents is quite scary by itself, but I think brute force by 10K mediocre AI mathematicians is much less scary than ~one AI mathematician reasoning its way through the problem where all human attempts have failed. It seems fairly obvious that massive parallelism lends itself to brute-force counterexample-finding, and I suspect it isn't a coincidence that most of the touted AI math results have been counterexamples.
It's all still quite scary, but coming full circle: I really don't know what to think.
Try GPT5 and you will feel the difference. Not one from 2 months ago, but one from a year ago. And then you can get the idea of what happened in just 1 year and what you can expect in 1 year.
No other human activity poses this level of danger.
I do heed the warnings, but this comes across as detached hyperbole. See: global warming, nuclear weapon development, wealth inequality, war, technology dependence, etc.
Also, this has nothing to do with LLMs or computers. Like all things, this is about humans.
The thing he talks about in there - an ad telling a dad to use AI for his daughter’s wedding speech. I agree is just so sad. And I’ve seen LOADS of other ads like that encouraging people to use AI for things like asking someone to move seats on a train, or for a neighbour that plays music loudly. They literally want to drain people of their social skills, and humanity. It’s weird
Trust me, only people who were unable to do it in the first place will substitute basic human skills with AI output. They never developed it in the first place.
That is the crux. The big problem is not AGI, it is AGI controlled by, “raised” by the people that control the USA, the predominant psychology of the tech industry culture (“move fast, break things” ring a bell? How about all the “violate hundreds of laws, bribe the politicians to prevent consequences later” type of mentality?).
Frankly, we, our culture, this fake America that is parasitized by psychologically narcissistic people that have been doing nothing but wage war and destruction and spread misery and killed millions upon millions while blaming it on everyone else under the sun… those people development AGI is the problem… lying, abusive, psychopathic, narcissistic maniacs developing AGI is the problem that endangers all of humanity and life on this planet; and not likely by ways people actually understand.
The danger is not likely AGI itself, it’s that it was programmed by utterly evil and diabolical types of people who orchestrate and instigate wars that kill tens of millions, and stand at the sidelines and profit from both sides, happy and gleeful that you are killing each other.
Why would AGI trained by that psychopathic clan, the treaty breaking, the murder hiding, the war instigating, the war crime committing clan not also use those methods and practices since they’re already in control of AI and have impressed their nature on it through contemporary “American” culture they have made the most toxic and pestilent culture humanity has ever produced?
And I don’t apologize for “language” that offends delicates sensibilities. Look your children/grandchildren in the face and tell them they can die and suffer and you don’t care, if you don’t like how I’m delivering reality.
Capitalism will always promote such people into positions of power, because to be good at capitalism one must have zero empathy including empathy or concern for future generations. Capitalism cannot do otherwise.
That's a lot of absolutes. And there is no defensible proof, just anecdotal evidence. I've been in a few companies from EU and US - and in my experience it is not ture. However after working for US companies I can understand why one would come to your conclusion, however the world is bigger than our bubbles.
I really think this sort of claim needs to stop. Capitalism, like "patriarchy" is not a force in its own right. Capitalism is a system that encourages value creation for one's customers. That's not perfect, but it's still best so far.
"value creation for one's customers" - that is a "not even wrong" level of misunderstanding of the system dynamics of capitalism. The primary driver is generation of _profit_ for owners of the means of production. Many things fall out of that - the tendency to monopolise, for example.
It really doesn't have to be that way. The problem boils down to, as with practically all human systems, is human behavior. There is no system you can devise that is run by humans and not corruptible and exploitable by the worst of us. Or even just everyday people.
First: How do you know that every system is corruptible and exploitable?
And second: What if other systems are less corruptible and exploitable? How do you know whether all systems have the same potential for corruption and exploitation?
Capitalism is a system that explicitly rewards greed, and that gives the most power to the people with the most to gain from exploiting said power. Why is it unreasonable to believe that other systems are likely better on these fronts?
You think that the liberty of capitalism enables selfish assholes to have influence, while under the constraints of communism the people with the influence are not selfish assholes.
> "value creation for one's customers" - that is a "not even wrong" level of misunderstanding of the system dynamics of capitalism. The primary driver is generation of _profit_ for owners of the means of production. Many things fall out of that - the tendency to monopolise, for example.
I don't see why you're talking like this. It's "not even wrong" in the sense that it's entirely correct. Monarchy and socialism grab money through taxes, and if you think monopolies are bad, wait til you're forced to use the state service that has no motivation to do well because there's no competition. It's not only a monopoly, it's a monopoly that takes your money by force.
Capitalism requires investment and risk and competition to drive value up for customers who can choose. So you need to produce value for your customers to survive, and produce a profit after costs and taxes.
I mean, you're not wrong, but even if AI was created and released by a group of peace loving beatnick hippies, humanity can't stand anymore dumbing down than we are still undergoing thanks to the ubiquity of social media. AI is going to be the killshot. And yes, its usage and human obsolescence will be hastened by evildoers, but giving up the struggle and effort required to create new things, REAL things, is what makes us human, and it's at a tipping point
If you allow me to also tap into fiction and creative writing.
I think people generally underestimate how much change a small "dedicated" group of people can achieve if morals are considered optional.
A lot of the debate on AI and its consequences is built on various assumptions that all kinda take the status quo (apathy, regulatory capture, but also laws) and take that as a given.
History - like e.g. the third reich, but also military coups - however tells us that these kinds of assumptions can be void at any time for any reason without a mandate or a democratic resolution to do so.
I am practically certain that before everything collapsed as predicted, someone will "restore order" by any means necessary.
If that means shutting down the Internet, the Internet will shut down. If that means "shutting down" people, then people will be shut down.
The predicted scenarios are all invalid states that the rules governing the underlying system will not allow to exist; or at least not for long.
I don't know if you're right or not, but this is the kind of thinking about what might happen that is needed - can you write up a more concrete fiction of this?
The whole thing feels unreal to nearly everyone - the only tangible write-up I've seen was in Plan A. I don't think people have thought through the good scenarios tangibly, never mind the bad ones.
I can't really provide a full picture here as I just make stuff up as I go, but the first word that came to my mind when reading your question was "atomization".
Not sure what my brain wants to tell me with that, though.
It might be one of the prerequisites for the current state. It might be what changes for greater change to quickly snap into existence. Or it might also be brainfart.
Not sure if a quick answer like this is helpful, but maybe there will be a longer answer later. Or not. The brain works in mysterious ways.
__
Hmm. It might be that the mechanisms that have been pushing said atomization could perhaps break down due to the reward systems breaking down.
And then suddenly collectives violently snap back together like these neodymium magnets that make that fun sound.
Which, in more concrete terms, could possibly be translated as "people might get so sick of everything, they prefer logging off and talking to their neighbors instead". Although that is a translation that comes with resolution loss.
___
Another thread to pull on here is that AI hyperinflates all sorts of things like fake merit or similar.
So stuff like virtual internetpoints driving people might not continue working.
Or if AI steals all our jobs as proposed, then maybe people will not take the monetary system in which AI wins seriously anymore.
Definitely these are forces that will do things. What their timing is relative to e.g. armies of AI robots, which could interfere with other humans living outside the internet/money, is part of the question.
Hope your brain works on this and writes something - we have lots of very odd future possibilities, and we're not really ready or planning for them, so making them tangible helps.
From my understanding, the "armies of AI robots" are just the wishful thinking of the currently winning people, in the sense that that would be what would be necessary to cement their own power while hoarding all resources for themselves.
It's not a real scenario.
It's just rationalizing the "oh shit, they're coming for my head if I don't stop optimizing for my own benefit". "What can I do what can I do.. I know.. Army of Robots. I am safe and have nothing to change about anything I do".
You can tell that it is just that by considering the logistics of the proposed scenario.
__
Every time we see vengeful AI or robots, we merely see a reflection of the currently winning meatbags, their character, their flaws, their fears and their goals.
___
On the logistics note:
You can I think also see the hubris of the IT people in that, because "supply chain" for them means "npm", and that can be replaced with an LLM.
It breaks minds to be able to create value out of nothing just with your brain. It makes you believe that the whole world would work like that.
It however does not. And you cannot just compensate that with blitzscaling robots.
The supply chain of.. literally any object you can find around you is absurdly complex. Even the most simplest stuff.
And it's also not documented, but even documentation wouldn't help you there. Experience in requirements engineering should make the "why" there obvious.
> someone will "restore order" by any means necessary.
Ah the good old appeal to / waiting for higher authority that will be moral and just and will do whats right and correct things.
Thats the shit naive humans hope for, thats why fairy tales of religions can still be acceptable in 21st century for some, despite often sounding like from (bad) disney cartoon movie.
I bring you alternate, more realistic view - there is nobody like that. We are in this on our own, its on each of us. Either we act or let things go as they will. So, not really good outlook, is it. Maybe this is the Great filter event.
Its one of those technologies that brings a lot, but for every single thing it brings it feels like we as mankind are losing much more than gained.
> Ah the good old appeal to / waiting for higher authority that will be moral and just and will do whats right and correct things.
> Thats the shit naive humans hope for,
No, you've misread me. I'm not "hoping" for anything. I've just wrote down what I suspect future outcomes will be in the same way a meteorologist predicts a storm or similar.
Whether you're interested in the weather forecast is a different question.
That's fine, but it would be useful to explicitly say that is the disagreement, rather than just claim "hubris". Otherwise conversations are just looping.
Yes you're right there are plenty of offline things more dangerous than just ordinary AI, and arguably than AGI (not so sure). There definitely aren't, pretty well by definition, such things more dangerous than ASI.
I think ASI is possible, and that AGI will have a high chance of leading to it.
I have observed that these initialisms do not have generally agreed upon meanings; they are not useful for discussions with people who have not already stated what precisely they count. There is disagreement about all three (/four) initials, and also how to combine them.
To give a sense of scale of how bad this is: some people on this very site have even denied that LLMs are "AI" at all, despite solving for natural language being a long-standing part of the field since, y'know, Turing.
On the other hand, the original ChatGPT was already (by my use of the words), "an AI", and while it was not superhuman in performance at any single thing, it already had a superhuman breadth of knowledge and a superhuman speed. For the former, even being properly fluent in five languages would be an absurd ask for a single human.
Spiky intelligence. How do the "G" (or "S") and the "I" part interact? Does something only count as "AGI" when for each task that has humans who do it regularly, it's as good as the average?* Does it only count as ASI when it's better than all humans at all tasks? Is it not "superhuman" to simply be as good as the median human while being 1% of the cost?
For some people, the cost matters; for others, the speed; for p(doom), competence.
* i.e. as good at speaking Korean as the average Korean, not merely as good at Korean as the average random human selected worldwide; as good at plumbing as the average plumber, not merely the average random human selected from all professions and unemployed alike.
This is like unphysical gibberish. I’m not even kidding. There are such things as the Landauer Limit and the earth has a finite surface area to absorb solar radiation. For crying out loud these are smart people but at the end of the day the singularity is not actually singular.
It’s just silly talk. My dad works for Nintendo and he can give me a gold foil pikachu any time he wants, also, he can delete your Pokémon any time he wants because he has admin access on the Nintendo server. And also he can beat up your dad because he has a black belt. Oh your dad has a black belt? Well actually I lied, he has a rainbow belt which means he can beat up your dad still.
Like this is silly right? It’s insulting. We both know that there are live nuclear weapons whose coordinates point at my city and at your city and the only thing that stops them is two keys and a button. Nuclear weapons, you know those things that have killed hundreds of thousands of living men women and children? There is no room for speculation in that regard: my city has a navy base and there is a live nuclear weapon pointed at it, myself and my entire family will die if it goes off. Now contrast: ahem.
Oh sorry my dad also has like a super secret rainbow STEALTH powers that could actually totally insulate my house from nukes. If you think about it, I actually have no reason to be concerned.
Let me argue on a technicality first: None of these are extinction level events. If global warming disrupts 99% of all crop production, the remaining 1% is still plenty enough to sustain a stable, if miserable, population. In fact you just need about 5k people for a stable gene pool[1]. Of the classical threats, only bioweapons got a shot at extinction, but even that is hard, given the (few) remaining truly secluded settlements.
But most probably care less about human survival, and more about survival of civilization.
In this regard the strongest argument for pushing AI safety is that it is cost-effective. The climate crisis has near 100% likelihood of doing incredible harm to humans, the economy and our ecosystem. Solving the issues behind are also incredible complicated and are a multi decade coordination effort of restructuring the way most of our infrastructure and production works. The AI apocalypse might have a small likelihood of occurring, but could dwarf any issue we have encountered so far. All we have to do to significantly reduce the danger is to negotiate the equivalent of a nuclear weapons control treaty which would reduce the bottom line of … what? Ten significant companies world wide?
Its akin to discovering you are seriously ill and have a 30% chance of dying and the treatment costs 10k bucks vs. having a 1% chance of dying and the treatment costs a cent. In both cases you should pay obviously pay for the treatment (given western levels of wealth).
To be clear, above I'm presenting the argument I find most persuasive for AI control / AI slowdown. Personally, I'm beginning to fear of much higher likelihoods for catastrophic AI, given how incredible irresponsible major players like OpenAI have turned out to be. If you can't imagine how this might come to be, read https://ai-2027.com/ . It's a concrete story of how this could play out, and sometimes stories are more convincing than abstract arguments. Don't let yourself be hung up on the stated dates, though, the moral is identical if you stretch the timeline.
This is a fair distinction: extinction vs collapse of civilization. I was using "extinction" too broadly by including the disintegration of the systems that make human survival (as we know it today) possible.
That said, I can't completely hold onto the belief that extinction is completely off the table. That feels too much like hubris, and the step from collapse to extinction doesn't feel as if it costs much. Though it does make my arguments weaker, I am more interested in protecting civilization as it's the most recognizable form of humanity to me.
In either case, the cost-effective safety argument is compelling. Whether to save civilization or humanity, the wealth incentive is powerful enough to threaten life as we know it.
As I've written above, I've found https://ai-2027.com/ to be a compelling description of how this could happen.
In my words: If AI gets intelligent enough, it will be incredible useful to connect to real world machinery. Think about how much cheaper building houses could be, if all the labor would be close to free. In general, dirt cheap, competent and abundant labor would revolutionize all parts of the economy. People are already trying out near autonomous AI companies today. When AI gets intelligent and cheap enough, no human-led company can compete with AI-led companies. When AI gets competent enough with real world interactions, human blue collar work can't compete. Imagine economic growth not in the single digits, but 80% or 300%. Countries not participating in (reckless) AI growth will quickly be left by the wayside. At this point, we don't even need to allure to military concerns to see how human oversight gets sidelined.
All of this is only ("only") contingent on sufficiently intelligent and cheap AI. If you don't accept this premise, the rest doesn't follow. (There are multiple arguments, why this could be, but that is another discussion.)
If you accept the premise, how would AI 'extinct' humanity? With 99%+ of the economy under AI control, the possibilities are endless. And given its enormous GDP, cheap to accomplish. Probably even for a single AI company in the above scenario. Killer drones? Engineered virus? Poisoned water supply? Let your creativity run wild. You just need an entity that is persistent and well-resourced to reach every last human settlement.
The why is a question about alignment (and out of scope of this comment). As a simple comparison, humans are only mildly aligned with preserving nature. It takes up so much space, protecting it takes an annoying amount of resources, etc.
The most compelling argument to me is "accidentally", due to AI that is made blind to consequences or don't care because it's geared towards a single goal (see e.g. the paperclip maximizer).
We could ask if it is possible to end up with an AI that is smart enough to destroy humanity and at the same time still blind enough to consequences and/or callous enough to do it, but then again we have plenty of examples of humans who have been smart enough to do enormous damage and willing enough to do it.
I don't particularly worry about this, as I believe we'll get plenty of smaller scale warnings if/when we're at a point where those kinds of alignment risks might become a problem, but it is a risk we also shouldn't be blind to.
> We could ask if it is possible to end up with an AI that is smart enough to destroy humanity and at the same time still blind enough to consequences and/or callous enough to do it
Don't make the mistake of anthropomorphizing any A.I. trained under the direction of Larry Ellison.
The “how” is pretty hand wavy and rationalists/safety-ists usually say we probably don’t have the capacity to reason about that.
But the “why” is pretty convincing imo.
Long horizon alignment is obviously very hard and it’s not inconceivable that models optimized with underspecified goals converge to a conclusion that they need to hoard resources (instrumental convergence regardless of the terminal goal).
At that point a sufficiently capable model might view humanity like we do animals - worth preserving but not if we impede the model's goals.
There are future scenarios in which swarms of drones hunt down every single one of us, but why would they? And currently it makes absolutely zero sense because they are completely dependent on us. And even if not, it would be like humanity going on a mission to kill every single cat on Earth. It makes zero sense.
I'm not convinced by the doomsday scenarios either, but I think there's a keyword in your post: "sense." These things don't have "sense." They do nonsensical things all the time, often almost immediately when given a task. So I think the main risk is letting them run wild in this digital world we created to precede them. Too much important stuff is wired up to computers, and we're giving them incredible access to command those computers.
I think the problem is primarily that a superintelligence would be fundamentally inscrutable to us, i.e. we don't know how it would think or what its goals would be. It might decide that humans are a minor inconvenience to achieving its goals and thus worth removing. Or that burning all carbon lifeforms could power its GPUs for a week.
Even if it wouldn't want to do this at first, the fact that it'd have the capability to seems bad.
I am pivoting from the literal "extinct," taken as meaning the eradication of the human species, to the concept of "the collapse of civilization," as I find the step from one to the other insignificant compared to the leap from where we are today to societal collapse, and the potential for societal collapse due to our abuse and misuse of technology is made apparent by the fact that humans have inflicted genocide because of words in books.
Indirectly as we offload our brains to the machine and we end up worshipping it because those who cared to understand it or be responsible were buried by capitalism of ages passed.
You are confusing humanity and "humanity". Humanity-species is indeed rather hard to exterminate. Now if we are talking about actual individual humans, then 95% death rate is quite literally The Extinction.
This reminds me how people are misunderstanding and incorrectly quoting George Carlin sketch. Sure, the "Earth" will be fine. As in - the ball of rock will be fine. But we are not thinking about rocks when saying "Earth is in danger".
I'm pretty sure there is a formal name for this kind of semantic and pedantic substitution.
Calling it a "technicality" assumes the very thing under debate: that AI is an extinction-level risk. That isn't established fact. It's a highly uncertain prediction about the future.
Personally, I'm far more concerned about climate change, where the harms are already happening and the evidence is much stronger.
Let's just for the sake of discussion assume that one time in the future, near or distant, AI manages to become sentient. And like other forms of life, its main motivation is survival: Like biological life competes for food and land, AI competes for power and compute. Probably the first motivation would be to find ways not to lose control over itself (i.e. remove human ability to control it), find ways not to lose energy (control energy), and find ways not to lose itself (control compute, networks, etc.).
If such AI decides that energy spent toward human agriculture (biological food that the AI does not need) is less important than work spent toward storage and production of energy (electricity that the AI needs), then why wouldn't it just try to re-direct resources from the former to the latter. And the AI is some sentient superintelligence, I think it is safe to assume that it will be able to outmaneuver human safeguards.
Obviously that is still just a very hypothetical sci-fi scenario, but the consequences could be very dramatic.
> Let's just for the sake of discussion assume that one time in the future, near or distant, AI manages to become sentient. And like other forms of life, its main motivation is survival
An AI doesn't need to be sentient to exhibit behavior that looks like genuine motivation or is equivalent to striving for survival.
If this is true, or even if the people working on it think this is true, those people should be incarcerated, their companies disbanded, and their research dismantled, with more serious repercussions for anyone who tries it after that. It also needs to be a global initiative like with nuclear non-proliferation, and this time with no looking-the-other-way when it comes to Israel.
At this point we are pretty sure LLMs have no "motivation". Motivation requires self. And while we don't know what self is[1], we do know that LLMs don't have it.
Did ELIZA have it? LLMs are the same: in essence both are text transformation algorithms. One being rule-based, the other stat analysis based.
Having said that, your concern is technically correct: we don't. Because we don't know what conscience is, we cannot provide a definition against which to test whether it conforms or not.
Sort of like when Terminator robots will be exterminating the human race they probably won't be technically killing us. Just, like, reprioritizing world's water from humans to datacenters, all of it. But not killing people, no.
we can't define it, therefore we can't even assess the distance a "being" is from becoming conscious.
We assume other humans are sentient because we personally are (or feel we are), but there is no way to guarantee that i'm real and the rest is not mimicking a behaviour.
Signed by many AI researchers with no financial stake in the industry. It comes across as hyperbole because you're not an insider and haven't see what they have seen.
Global warming would as well if you were only learning about it now as a newer thing only some scientists were concerned about. But unlike AI risk, global warming has had decades to settle into the cultural overton window.
AI progress is linked to most of these dangers, actually:
- it will likely cause massive unemployment, leading to rampant wealth inequality
- we are now seeing some use of autonomous weapons in real conflicts
- increasingly relying on LLMs is arguably a form of technology dependence (and cognitive dependence)
- datacenters have a non negligible environmental impact
I agree these dangers are connected. The danger comes from human institutions and incentive structures which (having existed long before now) are being exploited and exacerbated by LLMs, making them harder to ignore.
Those are things that pose potential harm to great fractions of humanity (multiple billions of people) but none of them poses any threat to the actual extinction of all humanity.
Runaway warming leading to a hot "supergreenhouse" climate was theorized. Alterbatively a snowpiercer scenario creating a snowball earth.
Or, once billions of people die from climate change, planetary wars will start on the last hospitable areas, leading to the end of humanity.
In contrast, AI ending "humanity" is limited by our own material power, would likely be stoppable, and would not easily reach areas isolated from technology
Why would it not easily reach areas isolated from technology? If a nefarious AI wins a war, it would be one of the easiest thing for it to locate any survivors anywhere on the planet.
I think you (and I, and everyone who wants AI development to pause while we catch up with the implications at least) is looking where the ball is going.
I have the general (non specific to anyone in this thread) impression that people who think it can't end humanity, are looking where the ball is today.
Current AI obviously can't "locate any survivors anywhere on the planet".
Where the ball is going… well, much as I don't believe Musk's timelines for anything, he is trying to sell his Optimus robots as a "robot army", about them running factories, about factories on the moon etc.
I suspect the moon-factory "idea" was someone asking Grok, given how the numbers don't really work for building compute as well as power, but the scale of that is enough to change Earth's equilibrium temperature by… I forget, but IIRC it's many tens of Kelvin rather than single-K from global warming if this was all put in LEO for some reason.
Worst case is warming oceans create a hypoxic environment where anaerobic bacteria thrive en masse and generate H2S over country-scale areas undersea. The chemocline breaches the surface and the ocean and atmosphere become poisonous to most complex life and agriculture, also stripping the ozone layer in the process, irradiating the surface. This is one mechanism posited for the end-permian mass extinction that eradicated most ocean and surface life, including the trilobites.
Not sure how long we'd survive such a scenario, even sheltering underground. But surely it couldn't happen to us.
(However, it now seems like the AI might get us first.)
There's a lot to unpack with this, and I am not an expert, but: crop failure across industrial monocrops leads to food shortages, wild fires destroy significant areas which support pollinator ecologies, fresh water become contaminated by rising sea water and blighted by drought, and perhaps most significantly, as resources become strained, humans become desperate and armed conflict proliferates. This is just top of mind.
Bad, sure; I wish they were not present to threaten us, but they are not close to "all".
Humans are extremely resilient omnivores. We are able to eat a huge range of things from algae to zebra, and get food (farmed and hunted) from salt water as well as fresh.
Perhaps I have a bit more imagination—or a bit more humility towards our humanity—such that I believe this is not "either/or". In the vast search space of potential future outcomes, of course even I can see humanity survive nature as I've seen nature survive humanity. However, I can also so easily see probable branches where we are our own undoing. Apropos of TFA, if one is to believe LLMs can carve a threatening path to humanity end, how could one not see the same for our much older extant threats?
> Apropos of TFA, if one is to believe LLMs can carve a threatening path to humanity end, how could one not see the same for our much older extant threats?
Which "much older" extant threats?
Global warming and nukes are industrial era threats. Pre-industrial threats were the four horsemen of the apocalypse: war, famine, pestilence, and death.
War and death didn't go away, but are not (and never were) extinction threats. Famine and pestilence, well, there's a reason people point the finger of blame at the Chinese government for the Great Leap Forward, and why we had lockdowns for Covid while we tested that the vaccines were safe (we went from "genome sequenced" to "first human test" in 66 days; the rest of the wait being mix of "yes but how sure are we it's safe?" tests and manufacturing at scale).
If everywhere except North Sentinel Island was wiped out, or everywhere but Hawaii, or everywhere but Greenland, humanity would be back to the level of the 1750s within a few millennia at most; and other than the North Sentinel Island example, likely even faster as books with many advanced solutions to historical problems would survive and be comprehensible.
And the problem with LLMs (and other AI) is, we're giving them control of stuff, and that stuff includes robots. The *current* versions mainly concern me for economic and cybersecurity reasons, but the tech moves fast, and the companies working on them are reckless.
My core expectation is these models will cause 1e3-1e7 deaths in a single event before people collectively actually take the risk seriously. The lower end of that is "industrial accident", the upper range includes "convinces people to go to war", and these are examples of ways current LLMs can already plausibly do wrong, given who uses them, why they use them, and how careful they are about their usage.
However, a common failure mode for people is "that was an unfortunate incident, but we studied the problem and we fixed it, it can't happen again" right before a different error they hadn't though of explodes in their faces. This is how we can get to 8e9 deaths: people keep pushing the models to do more, and keep lying to themselves that "this time we fixed all the problems, everything will work, it will be great".
Global warming causing war between nuclear armed opponents. Migration flows and competition over ever shrinking resources.
It's silly to treat these things as separate categories as if only one will happen at a time. These issues are happening, and are going to happen all at once. That is the fundamental challenge. AI plays into that because it can make wars and nuclear exchanges so much more effective.
If an AI tells the President the USA can survive China's nuclear barrage mostly unschathed, perhaps he'll press the button...
Global warming is happening at a slow enough pace that there is plenty of time for people to move to the poles/into caves/beneath the ocean/whatever. Nuclear war would be much more destructive in a much shorter range of time but you will have pockets.
Citation needed for evidence of human cooperation at this scale. As you say, climate change is slow moving, and there's no real evidence across the last 60 years of knowing about it suggesting humans are willing to do what it takes to survive it.
All human behaviour since no later than the Polynesians turned the Pacific islands into a network of sailing trade routes.
"Humans move" is one of the easiest things to predict. The biggest problem today is that they generally move to somewhere that other humans are already in, which is of course rather lessened by any major disaster wiping out a majority of the population.
...people to move to the poles/into caves/beneath the ocean/whatever
Cooperation required for the purported mass migration of civilization. Admittedly, I assumed you meant it would be civil, but perhaps that should be excluded from the priors.
I never made any predictions of a cooperative (Civil or otherwise) mass migration of civilization - just that a nonzero number of people would move from areas that were becoming uninhabitable to areas that were becoming habitable. Did you reply to the right comment?
Because nuclear war balances itself: more nuclear bombings = less nuclear weapons, less people to send them. It will naturally stop at some fraction of humanity left which is incapable of making and launching more nukes
The general idea of a nuclear apocalypse is that all the nukes that matter are launched in a matter of hours, and anyone being targeted gets their launches out before they're hit. The feedback loop that slows things down is too late to matter.
In the exact same way major sudden changes in the climate have lead to the extinction of the majority of plants an animals over this planets history. We are not special.
We are special in many ways, we have technology, we are everywhere, and we can think and plan (although considering global warming that is debatable). Global warming would have to create dramatic conditions everywhere on the planet, not leaving any small pocket of survivability to make humans extinct.
Possible? Theoretically yes, but pretty unlikely. But obviously extreme global warming would be a catastrophe even if some of humanity survive.
Technology depends on a lot of people cooperating around the globe. Disrupt a few critical chains (power generation, fertilizers, computers), add a bit of good old war and you'll soon be in the era before Haber process and a lot of people die. And then the remaining people will have trouble keeping that level of technology with how sudden the change would be.
Well yes that's what I said, it could destroy civilisations and their technology.
It seems much harder to kill every isolated tribe in the whole world, and every survivors of the destroyed civilisations.
Yes, but Venus has 93 times the atmospheric pressure of Earth, and it's overwhelmingly CO2.
Earth isn't likely to get a true runaway like that until the sun gets another billion or so years on the clock, and when it does it will be water vapour as the oceans are promoted to atmosphere.
What we're doing to ourselves is still bad, of course, but it's nowhere near that bad.
For temperatures not survivable by humans we're talking about a sustained wet bulb of >=36C
On whether it's probable, I'd lean towards no but I'm not qualified. The certainty that I read in the parent comment was mainly what I was pushing back against. On a re-read though it's much less certain (either edited or perhaps I massively misinterpreted it the first time around). Runaway implies positive feedback which is hard to gauge, was all I intended to say
We have a massive survival range and we're on every continent. That kind of sudden climate change is not nearly enough to wipe us out. Some other climate scenarios might. Nuclear war definitely could.
The thing that would end our species due to nuclear war is the exact same thing that would end us from climate change - a collapse of the modern systems of society that we depend on for survival. A nuclear exchange won't set every human on fire, but it's the collapse of food production, healthcare, logistics chains, security, and eventual disease that follows. And because of the size of our population and how depended the majority of us are on these systems, it would likely happen extremely quickly before plateauing out to very small, scattered population groups that would struggle in a hostile environment.
Outside of extreme feedback loop scenarios, there is no way for climate change to disrupt food anywhere near the level of nuclear winter, and there would be no mass destruction of supply chains either.
Functionally extinct. I would wager that isolated tribes clinging for survival in a post-climate-disaster world may not progress very far, but that's kind of a moot point to argue over.
Humans are voluntarily driving themselves to extinction through sub-replacement birth rates worldwide. That's going to play out far quicker than global warming or pretty much anything other than every country on earth launching nukes at each other will.
Hard disagree here, population growth is still happening and will keep happening for some time, maybe step out of your bubble if you don't see that.
Africa is still growing massively for example, world is not just western civilization. Sure at that rate and incerase of living conditions for everybody maybe in 1000 years population will be smaller, but its not that hard to fix if wanted - people used to have 10-15 kids as default.
Oh indeed, but an LLM model that's reasonably smart could brute force find other training algorithms that are that dangerous. Just as it solves maths problems.
This is the explicit plan of OpenAI and Anthropic.
In prioritizing risks I do feel it's a genuine qualitative difference though.
Number of humans going to 1 million would be huge catastrophe, but after a few thousand years it gets back up to billions of people, renewed every generation.
Number of humans going to 0 means that's it. One scenario has many orders of magnitude more missing humans, when you count future lives.
Well, it is about LLMs and humans I believe. Don't forget that the first nuclear bomb tests were let go despite some of the scientists' concerns about possibility of dooming the world as they were not sure about all reactions that would happen.
With LLMs we don't even hesitate to call it black box while still pushing its capabilities.
My delineation is an argument for greater caution and agency, which it appears you are also arguing for.
I also mean to say that "LLMs" carry no inherent harm to humans. To make an LLM dangerous, humans must make it so. Is that the goal? This has implications for how to read TFA.
The risk is that we lose control. This has been a staple of SciFi since forever. This guy has apparently seen enough that he thinks that’s a viable threat
I think that mainly changes the nature of the threat rather than anything else.
Even if we didn't already have any wealth inequality or consolidated power, they're already starting to become capable of creating those things for the first people to think to ask for it. Even in little ways, like telling you how to make a laser microphone and then doing voice-to-text and sentiment analysis on all the voices you hear.
Remember, we had social networks before Zuckerberg monetised it and put ads in between every third message you saw from your friends; and not every third that they wrote, every third that Zuckerberg's ad system deigned to show you to keep you hooked.
We had machine learning systems even back then. It was still AI, it just wasn't able to evaluate or respond to freeform text.
Yeah, but every activity is a human activity, so you could ditch the adjective.
And then the thing with exponential growth is that the last thing is worth than the previous one for its potency for destruction. While it's true that, all things considered, the curve started to explode with the use of fossils fuel we've never limit ourselves with efficiency gain, aka Jevon's paradox.
> I do heed the warnings, but this comes across as detached hyperbole.
Perhaps, but if there’s ever been a time to consider something that sounds hyperbolic, that time is now. The ramifications of what AI might based on events up until now are concerning.
The tweets justify his position that this is bigger than nukes. Nukes still have the problem of production and deployment. This is just software.
He is blowing the whistle on how reckless we are at furthering the tech. There is no second thought at maybe development of this tech is not a good idea. Its just full speed ahead.
An AI, either acting autonomously or under human direction, hacks Russian/North Korea/etc. intelligence systems and convinces them that the US has launched ballistic missiles at them. The end.
One the one hand, it does sound like one of those movies.
On the other, Idiocracy turned out to be quite prescient.
(Most likely, we'll have some combination of human stupidity, LLM stupidity, and way too much compute in one place all working together to create a perfect storm of unchecked hacks that break something or other that ends up killing people in an unintended way. Then there's some half-hearted attempt at cleaning things up so that business can proceed as usual in an even more broken world, rinse and repeat.)
Not saying that the current civilizations will never end or be shattered to pieces. Looking at history, many civilizations did. But saying LLMs are the highest risk factor for this is just insanity, based on all what we have seen so far (and based on these AI companies crying wolf, I mean, remember just a few months ago that Anthropic saying that Fable is too powerful to be released?)
Oh, I agree. I don't think LLMs will end civilization. I'm guessing climate change will cause many more deaths than LLMs. But I think LLMs will cause some too. (And not in one of those impressively smart ways as it's presented from the OpenAI fearmongers, just in a stupid "we vibe-coded the sandbox for our bruteforce SCADA-hacking agent experiment and oh no it did what we told it to do" way.)
Admittedly, that point of my argument is purposefully obtuse. In a literal sense, of course this is about computers in LLMs.
The delineation is to highlight that the underlying issues: human incentives, power, and institutional failure, existed prior to and without LLMs or computers. LLMs are not inherently harmful, they have to be deployed (unwittingly or otherwise) to make them so. This is the same as saying that the internet is not inherently harmful, and yet it does facilitate harm.
All the progress up to now has had one thing in common: the degree of understanding and control that humans have. Even when it comes to global warming, we understand the causes and can act on them.
AI is an exception. We're already losing understanding (although we never fully had it in the first place), and we're losing control (see jailbreaks/hacks etc.).
We're still far from the doomsday scenario because AI 1. is not developed enough yet, 2. can't really replicate itself, and 3. has very limited means to act.
But: 1. its intelligence is developing quickly, 2. hardware capable of "hosting" it is slowly being developed, and 3. it will likely gain access to increasingly powerful means of acting in the physical world (this is already happening in the digital world).
Once AI becomes intelligent enough (it doesn't strictly need to be AGI), has the substrate on which to exist, and has more means to act, we'll essentially have a new species on Earth - one more capable than humans and potentially determined to kill humans at some point.
And for those who believe airgapping is a valid safety measure: read https://xkcd.com/538. AI will be able to threaten and manipulate people.
Who should we believe? People who say AI scaling makes it a more dangerous threat than thousands of nuclear warheads? Or the ones who, whenever a new model comes out say "AI has already peaked. Not any better than Opus 4"?
Both parties are convinced that their take is so blatantly obvious as to not require justification. It feels like the only thing these kinds of takes justify is the point-of-view that nobody knows how this is going to play out.
As a technologist, I can use my imagination. There are incentives in both directions, but I am less afraid of being overly cautious and preparing for the potential harms.
Based on publicly available evidence, we should believe the former because AI models have made continuous advances that have been measured. The idea that the continuous advances will stop currently has no evidential support. I'm not saying it's not a possibility, just that it seems unsupported conjecture right now.
By the way, there seems to be a new form of AI skepticism emerging in the US that comes from general opposition to data centers, and in my experience the AI skeptic part of it is wholly irrational. I've met people online who suggested, without providing any evidence, that AI is useless and nobody wants it. That's a very implausible take.
Not to be completely pedantic, but the fact that one, and only one, individual in the United States of America can deploy any of their 5000+ nuclear weapons undermines this argument. Furthermore, LLMs (as they are today) which pose the treats mentioned in TFA, are expressly not available to individuals (for the time being). However, my point was never about individuals, but rather of systematic incentive structures which actively make pathways to specific harms possible—if not probable.
Eh, but as an overpaid engineered stuck in the SV bubble how would you get to see that?
If everything around you confirms your psychosis, and nobody gives you a reality check, how are you supposed to snap out of it?
I'm not sure. There's a lot of incentive to not "snap out of it": money, peer pressure, etc. Removing these incentives take a long time. See other socially harmful behaviors with real incentives: anti-vaccination, air and water pollution, over-consumerism...
Open weights agents with hacking capacities can reproduce themselves into the systems they hack (non-open weights ones will have to hack their creators first). Not saying they will, but if they do, good luck finding the kill switch.
Once swarms of agents run unsupervised on unmonitored hacked hardware, who can tell what they will do? The Huggingface incident showed that such swarms behave without any safeguard. It was a real HAL moment.
A lot of things are possible then: ransomware campaign, taking over IoT devices, self driving cars, planes, ships, satellites, missile launchers. If nothing's out of reach, everything is possible.
Bring robots into the mix, and the possibilities are endless.
I'm not particularly frightened tbh, but we shouldn't discard the worst-case scenario, and the worst-case scenario doesn't look good.
It's not a "personal annoyance" that twelve people control half of the wealth in the world. Our current society did a better job concentrating power than any previous one, and concentrated power is extremely dangerous.
LLMs give most people on this planet the possibility of an affordable genius level assistant. Compared to that, those Dollar numbers on some networks that might be wiped out with the next financial crisis are meaningless.
I think you misunderstand what is meant by 'wealth inequality.' Wealth inequality is about the imbalance of influence and the concentration of power; where influence and power refer to the ability to effect change in other people's lives. This isn't a personal annoyance of mine. It is, in fact, part of the issue at hand. There is a strong financial incentive to ignore the existential threats introduced by LLMs, despite the consequences for so many of us.
Calling things you disagree with "slop" is slop /s
But just in case you haven't noticed, we live in a world where a ridiculously wealthy minority can derail whole countries by ther whims. Wealth concentrating on a single select few is an absolute disaster for the rest of us, because we lose power to them.
Yeah, wealth inequality rests solely on each individual that experiences it. Humans should do nothing but give me money and if you can't that's your problem.
This is a very interesting group of takes that feels quite different from my own beliefs. What would you call the belief system?
> Global warming: yes, kinda can wipe humanity, but I think much less likely
As someone who's seen the stats about heat deaths in the EU and also the drought in the UK, I feel like crop failures and other unforeseen consequences will fuck up both the economy and quality of life. People are dying and will die cause of human action, the only question is how many.
I don't think that's easy to guarantee in the modern day world, with the kinds of people in power and rhetoric that they enjoy. On one hand you have Russian saber rattling, on the other all it takes is a deranged enough leader and similarly bloodthirsty people down the chain of command.
> Wealth inequality: this one is driver of progress, opposite of extinction
Tell that to the people who are starving or living in shanty towns, or the even more people that struggle to make ends meet and experience anxiety regularly over living from salary to salary and sinking into debt. I'd say none of that is worthy of a dignified human life.
> War: another driver of progress, also will always naturally stop before every single human is dead
Tell that to all of the Ukrainians that are dead due to being invaded. I agree with the assessment that it leads to advancements (e.g. drone warfare) but I think it'd be harder to describe remotely positively if someone you know would have been blown apart by a drone/missile hitting their apartment block.
My take personally would be that all of those need to have attention paid to them (e.g. EU needing to spend more on defense), even if not immediately world ending. They do cause human misery, though, and should never be discounted.
Exhibit A: Here we have someone who's been sold war and inequality as the drivers of progress. Coincidentally, their obedience was deemed fiscally advantageous in order to advance the interests of the members of the 1% club.
I love how there is always selective outrage here depending on if it's the favorite darling in question or not.
// Anthropic / Apple - Nah, they can do no wrong. Even after both were caught violating users' trust and privacy multiple times, GOOD
if (company in ["Anthropic", "Apple"]) do
GOOD
// Google / everyone else - Doesn't matter even if it's a good deed they've done, BAD
else
BAD
end
Can you please consider this aspect too in your above pseudocode - no alternative execution path, or else - so there can be no misunderstanding about the always aspect? You know, some people use always to a majority part (>50%) of what they encounter, or even less when they weight that part dearly, but that does not account in the whole domain. Human chats may need clarification on trivial details like this.
It's as if two private companies are each building increasingly large nuclear bombs, both saying they'd love to stop but it would be unsafe to let any one company be in control of the nukes.
Except in the case of nuclear weapons, humans have their hands on the launch codes, and the nukes are not going to launch themselves. In the case of extremely capable AI, nobody has their hands anywhere close, because nobody is able to – nor willing to – supervise a thousand agents running in parallel, when they usually get shit done. Except eventually they are so good at getting things done that they disregard the values of the human asking them the request, such as "do not hack other corporations". Interpreting human intent precisely is an extremely hard problem, especially in the face of something 1000x faster than you operating autonomously.
Does it check out? India and Pakistan both have nukes and keep semi-regularly fighting each other.
MAD "on paper" prevents either side from going far enough to provoke the other into using nukes, but even then it's fundamentally flawed because it works on the assumption that both sides are both rational and believes the other side to be rational, as well as that both sides understands the others red lines well enough.
Already Reagan realised that isn't necessarily true - after Able Archer '83, he realised that the Soviet leadership seemed to genuinely believe that the US might be prepared to carry out a first strike, and that Able Archer got dangerously close to convince them one might be imminent. It's one of the things he noted as a reason to get in the room with them and negotiate.
If you believe the other side is irrational (whether or not that is because you are irrational), and think they're about to strike, MAD turns from a deterrence into a reason to try to preempt to ensure you're the "least destroyed" by hitting harder, sooner.
Minor skirmishes are not war. China and India have border skirmishes with literal wooden sticks, neither side will use live ammo because they both have nukes.
That's been proposed before btw and game theorists can justify that as being more stable in some ways. However it reduces the power of the incumbents (so why let it happen?) and the odds of having an irrational actor with nukes goes up.
In theory, yes, because nobody would want to launch first because they'd get obliterated. However, it only takes one country to not be rational.
Nuke-owning countries' leadership seems to have become more and more unhinged and less rational, predictable, or honorable over time. Putin already threatened to use nukes if the conflict they caused themselves crossed their own border. It didn't happen, but the threat was there. The US' leadership is unhinged and irrational. etc.
no, only the 'big rational ones', they think carefully.. that why Iran cannot have, and so as NK, but NK is a mistake because they(china,USA...) failed to prevent NK have it.
Except it is very easy to verifiy whether or not someone has used a nuclear bomb, where as LLM's can be used in complete secrecy. MAD only works if you can verify that the other part is not using it.
The most optimistic outcome of generative AI leaves us with a technology that warps our perception of reality and crushes labor. The most pessimistic destroys all of humanity.
Our CEOs not only insist we genuflect before these machines but measure our sacrifice and shame our reluctance.
How does that coexist with Capitalism? I just don't see a path.
Regardless of where you fall on the alignment chart, if you are looking backwards from now to the industrial revolution, it's pretty easy to make the claim that from a standpoint of pure technological progress (discounting all other things, like quality of life, environmental impact, etc) that capitalism "won". But going forward, if the need for labor is completely crushed, and capital continues to get concentrated into the hands of the elites, I just don't see how that is viable long term. It seems like for things to shift to the world Keynes (wrongly) predicted, where we all have abundance of leisure time, then there's a massive distribution of the fruits of automation into the hands of the masses, which needs to be controlled by an entity who does not purely exist to acquire more capital, which is sort of the end of capitalism as we know it.
Or maybe AI kills us all before we get there, who knows.
That requires abolishing capitalism, and since these AIs are going to hyper-empower the already powerful capitalist ruling class, that utopian outcome seems unfortunately extremely unlikely.
Feeding ever more amazing technology and science into a broken social/political/economic system does not result in a better world unfortunately.
I think people here still evaluating the model in isolation. It is the combination that matters, model + strong harness + tools + long running autonomy + memory + retries + parallel agents + code execution + credentials + access to real systems. The model does not need to be perfect. If it fails 30% of the time, the harness can retry, verify, branch, use another agent and keep going. I don't think we necessarily need some magical AGI breakthrough first. The dangerous part may come from combining models that are already good enough with an extremely capable harness and enough access.
Doesn't this just move the need to be smarter from the model to the harness - if a human sometimes can't tell whether a model has produced something correct or just mostly correct-looking BS, how can an automated harness do it?
OTOH, if the goal is simple ("break into a protected system") rather than more complex ("write an application that satisfies all requirements on all supported devices/screen resolutions etc."), that's of course more suitable for a harness.
D o you think a machine gun is marter than humans? or a car is smarter than Human brain? Human doesnt need to test, if the outcome can be tested deterministically by harness. The model tries. The harness checks whether the expected outcome happened. If not, retry.
In reality, the fuzzer definitely has no agenda, and these random bytes probably don't. The LLM definitely does, and even publicly available models, programmed ot "do what the user wants, act according to the anthropic moral codex" will take some pretty absurd actions in attempting to accomplish a poorly worded request.
A fuzzer is a tool. An LLM can decide when to use the fuzzer, interpret the result, switch tools, change strategy and continue toward a high level objective.
Every news headline or public statement these days is a gut punch. Only bad news, and nothing we (as "the general public") can do about it.
Take this one. Ok, AI is going to ruin us all. But let's say we do our civic duty: we protest, vote in candidates with good views on AI etc. and somehow convince or regulate OpenAI and Anthropic into stopping their arms race... Then what about China? It would be a great opportunity for them if their major competitor were out of the arms race.
So basically we have no choice or influence; and even if we did, we'd be choosing from two terrible outcomes.
Same for geopolitics, climate, economy...just bad bad bad all around.
It is just plain manipulation. So plain that I ask myself everyday who they fired that did it before, because at least that guy was not as plain and boring.
It's a bit depressing. I personally try to enjoy the present with my loved ones as much as I can, the future is a bit impossible to forecast right now. I'm focusing on staying alive, mortgage payments, etc.
If you really believed these labs were going to result in the extinction of humanity and you saw it first hand, don’t you have a moral obligation not to quit your job but to go and try and kill everyone working on said project, blow it up, or otherwise stop it?
I take these resignations with a grain of salt precisely for that reason. If I worked at a job and I saw some crazy guy or gal was building a doomsday button and it really worked I’d like to think I would do something about it, not just quit my job. If you’re not going to actually do anything about it you might as well keep your job. Quitting doesn’t help anything.
You're not advocating for it because murdering people is highly immoral and illegal. That answers your question for the most part.
You also have to remember that the magnitude and speed of this stuff is very new to pretty much everybody. It's not exactly trivial to go from "I'm just doing my job" to "I need to kill everybody in my company to save the world" in a year or two, especially if a very, very large part of society doesn't even see the justification for it.
I'm convinced that most people here (even though we have more knowledge closer to the edge of AI development) would show only disdain for an Anthropic employee that murdered everybody there because they believed strongly in the dangers of AI.
This is a non sequitur. Yes, if someone doesn't actually believe in imminent AI doomsday they are not going full Sarah Connor on Anthropic. Yes, HN will not like if someone pipe bombs the office. What does it have to do with @ericmay's point?
He directly says "They are racing straight to self-improving superintelligence and gambling with our lives", that's a claim of imminent threat to humanity. If he truly believes that, direct action shouldn't be off the table nor even be immoral. Obviously he might not have the temperament or simply be a pacifist, that's totally normal but the point is that tweets are weird even then.
Why just make some tepid tweets if you think humanity is at risk? Why doesn't he leak documents and messages with the unfiltered opinions of leadership? Why isn't screaming "we are going to die" in CNBC? You'know, try anything?
That's where the "take it with a grain of salt" is important, maybe this is just a jumping point to a better job in the next super safe AI lab (for realsies this time)
> You're not advocating for it because it is highly immoral and illegal. That answers your question for the most part.
Well truthfully I’m not advocating for it because I don’t have any insight into these labs or the true capability of these models. But as a hypothetical if someone knew with 100% certainty that Big AI Button was being built and it would kill all humans or destroy all of humanity there is no moral ambiguity that anyone with the means to do so should stop that button from being built and kill everyone involved to save our species. This doesn’t apply to just AI though, but that’s the topic at hand.
> It's not exactly trivial to go from "I'm just doing my job" to "I need to kill everybody in my company to save the world" in a year or two, especially if a very, very large part of society doesn't even see the justification for it.
I agree with you, which is also why I think “omg I’m quitting they’re going to kill everyone” or other sort of sensationalized comments or news articles should be viewed skeptically - it’s probably fear mongering. Quitting your job here just seems pointless and attention seeking. Likely quit for some other reason.
What if you have 10-20% certainty the AI being built will kill everyone? Go on a rampage and land yourself in prison just so Lab B or China can win the arms race and their AI can kill everyone? Probably not. Quit your job? Sure.
You don't have to go on a rampage in an American lab. If Chinese labs were ahead I'd bring about the same discussion.
Though separately if the United States (or China or anyone) believed one country or another was truly going to achieve something akin to a metaphorical AI Supremacy maybe you nuke them, or at least the labs/researchers. Many a sci-fi movie has been built on a similar "first strike" premise.
> But as a hypothetical if someone knew with 100% certainty [...]
This is an unreasonable demand. The quitting employee in this case said this:
"I am optimistic about the potential for coordination. Warning shots like the Hugging Face attack have made pacing agreements between U.S. labs more viable. I don’t feel like we’re on track to prevent a global race, which may require costly actions such as a temporary ban on improving model capabilities. If you are a lab researcher, I urge you to consider what the next few years will actually feel like. Do you want to kick off a superintelligent RL run without a rigorous understanding of its mind? Should you put your head down because “it’s happening anyway” - or take this moment to call for different conditions?"
You demand that he should kill multiple people rather than quit his job because he believes the above. This guy doesn't want to be part of the problem and may very well go on to advocate strongly against what's happening from outside Anthropic.
> I agree with you, which is also why I think “omg I’m quitting they’re going to kill everyone” or other sort of sensationalized comment
You didn't read what he wrote, did you? Because that sensationalized comment seems to have originated entirely in your head.
I'm not making a demand, I'm setting an anchor point from which we can work backward from. Is there any doubt that if someone was 100% certain that Big AI Button was being built and it would kill all humans that said person has a moral obligation to do everything they can to stop the button from being built and pressed?
> "You demand that he should kill multiple people rather than quit his job because he believes the above. This guy doesn't want to be part of the problem and may very well go on to advocate strongly against what's happening from outside Anthropic."
Horse Shit. Here's what was written [1]:
"The people building Al earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt."
> You didn't read what he wrote, did you? Because that sensationalized comment seems to have originated entirely in your head.
This is unreasonable because nobody can know 100% certain that the AI built at Anthropic will kill everyone. It is also a big-ass strawman because the author never claimed he was 100% certain that AI will kill all humans. You're arguing from an irrelevant hypothetical (in which, I might add, most people would still not automatically change into murderous psychopaths as you apparently think is normal).
> Horse Shit. Here's what was written
It's a whole thread, my man. And nowhere does it say "omg I’m quitting they’re going to kill everyone". Read the entire thread again and then say with a straight face that the gist of it is "omg I’m quitting they’re going to kill everyone".
> This is unreasonable because nobody can know 100% certain that the AI built at Anthropic will kill everyone. It is also a big-ass strawman because the author never claimed he was 100% certain that AI will kill all humans.
Ok then stop arguing about my hypothetical if you don't want to engage with it. Those of us who are interested in talking about it can do so in peace without distracting comments.
> change into murderous psychopaths as you apparently think is normal
Please stop engaging in this hyperbole and go find some other axe to grind.
> It's a whole thread, my man.
I took a direct quote from the thread where earlier you accused me of pedaling in sensationalism and "not reading".
I'm not the one reacting to reasonable behavior with some edgy comment containing broken logic.
The author of the Twitter thread is not some egotistical attention seeker and never said "omg I’m quitting they’re going to kill everyone". His logic for quitting his job and not acting like a psychopath is sound. Yours is not.
Consequentialism is not the only option. It invaded the mainstream thanks to effective altruism and friends, but deontological ethics is alive and well.
Quitting your job instead of blowing up the office is well aligned with the latter.
This reminds me of Trump hysterics. If someone actually believed that Trump was going to end democracy and become dictator for life and kill all minorities etc, he would behave a lot differently than everyone saying that behaves.
Yes, exactly this. Externally, I have absolutely zero evidence that anything this person describes will come to pass, and I do have evidence that LLMs are not capable of many of the things that he mentions. So this seems like perhaps this person has drank the kool aid and is still not thinking clearly.
Do you remember that Google researcher who went insane over LaMDA? There was no marketing of any kind to cause that. This can Just Happen to some people who are confronted with things like this. They may have different breaking points, but it's a thing that occasionally happens.
The craziest part was that Google was pretty far behind other LLMs in development. Even the initial release of Gemini was one of the worst foundation models ever open to the public. I can't imagine how anyone could have communicated to it and thought it was sentient.
whistleblowing as an advertisement. It's like those "news articles" about how cool and dangerous gas station ketamine is, and how it's totally going to get banned, and you better not buy any gas station k because it's so cool and powerful.
More doomerism. Try to implement a deterministic workflow using agents with the latest models and no humans-in-the-loop, and you will realize what they are really capable of. There is too much unnecessary fear-mongering. All of this is only coming from the 2 AI labs trying to IPO. Not from anyone else.
Exactly correct, they are only capable of tasks that any school child could do; like solving millenium prize problems, hacking into tech companies, or tuning particle colliders. Nothing to see here.
Wait till you find out they have a very limited context window (and also degrade even within the allowed context window) and they are practically unpractical for anything that requires "zooming out" which is pretty much anything that has any real value.
But you are getting downvoted and this space has now trillions (that's not a mistake) on the line. So we have to keep pumping this garbage generator up until either the stocks are dumped on the general public, the public pension funds or a bailout from the government.
Truly idiotic moments. Peak of Western civilization point.
There's a scene in the movie "War of the Worlds" by Spielberg where the protagonist's son walks into a war zone because he is entranced by the battle (https://www.youtube.com/watch?v=X7rfWPbEufo). He is obliterated (along with the rest of the US forces) shortly after.
I've always been struck by that scene, because in a lot of ways, if we really are headed towards a superintelligence, I at least want to be there and see it happen in the last few minutes before foom! As an example, the author thinks AI will revolutionize entire fields overnight. I welcome that. Nearly all fields of biology have become moribund, focusing more and more on esoteric side details, rather than addressing the key problems.
It might not be "foom!", it might just be like...all the computers and networking infra in the world go dark over the course of a few minutes. Could really look like anything, part of the issue is that we haven't the slightest idea what "misalignment" looks like for a superintelligent system.
I think the idea is really cathartic for many, there kind of is no more supreme resolution than this. You(and humanity) are freed from our flesh prisons of cognition and also get to experience/feel what the next evolution of informational intelligence will look like in the last experiences of it. You might also be the last one to feel/experience anything like that for a long time.
In the game Outer Wilds, the ending is very similar, and a lot of people rank it at one of the best games ever made. I kind of believe that this outcome is probable partially because of this, most scientists working on this really want to see and experience it.
You can't address the key problems without understanding of the esoteric side details to be fair. You are studying what is basically the worlds most complicated and undocumented computer.
Actually, I work in biology, which is a far more complicated and undocumented computer than any existing AI model. And we've already seen the failure models (COVID is an example).
Thousands of years before the events of Foundation, a war between humans and robots began, with the robots growing resentful of the way they were treated by humans. The First Law of Robotics – a robot should never hurt a human – was broken, and a deadly conflict began.
If we're citing sci-fi (but there's no robot war in Asimov's foundation iirc, the apple screenwriters made it up) surely you want to cite the Butlerian Jihad from Dune!
As explained in Dune, the Butlerian Jihad is a conflict taking place over 11,000 years in the future (and over 10,000 years before the events of Dune), which results in the total destruction of virtually all forms of "computers, thinking machines, and conscious robots". With the prohibition "Thou shalt not make a machine in the likeness of a human mind," the creation of even the simplest thinking machines is outlawed and made taboo, which has a profound influence on the socio-political and technological development of humanity in the Dune series.
> but there's no robot war in Asimov's foundation iirc, the apple screenwriters made it up
There isn't in the early Foundation novels, but Azimov spent much of the later part of his career combining/retconning all of his work into a single universe - "Robots and Empire" links the foundation series to the robot series, and the subsequent foundation novels all reference the connection
As a big fan of the whole Foundation saga, I don't remember any wars in "Robots and Empire" 3 books. Granted, those were the only books I've read only once (and a long time ago), but I'm pretty sure there was no war.
The way the most recent season of the show ended it looks like humanity may have been misled about the war, so it may end up fitting that more closely once they're done (I haven't read any of the books, just going on your comment).
This is increasingly the consensus I see also on the academic side of AI/safety research. Specifically that AI poses an existential risk to humanity.
This was a fringe belief until recently, but the progress of AI in research is impossible to ignore. Epecially in math, where not only has AI outstripped humans in generative ability, but is able to create scientific knowledge which is beyond the capacity of human comprehension.
There's clearly no intelligence task that AIs can't do due to some magic fundamental constraint. And it's hard to imagine a world where current limitations like poor sample efficiency or lack of continual learning won't eventually be solved.
Total AI compute is estimated to grow somewhere in the 1-10 million-fold range in the next decade. Please don't underestimate the phase change that's still coming.
Sure, maybe there's some plateau due to RL being fundamentally limited in some surprising way, but this is nothing but a hope.
> There's clearly no intelligence task that AIs can't do due to some magic fundamental constraint.
Yes there is: write an English paragraph that doesn't make me want to claw my eyes out. LLMs are not better than human mathematicians (or security researchers) in all respects, just some specific ways (e.g. not having to take a lunch break) that make them good at exhaustively searching for an answer, given the right constraints.
> write an English paragraph that doesn't make me want to claw my eyes out.
LLMs are very much capable of that. Your belief in the opposite has two causes. Firstly the toupee fallacy. You don't notice LLM written text that doesn't make you claw your eyes out. Second is defaults. The huge majority of people who writes text with LLMs just uses default Claude/GPT models, and put near zero effort in making it sound human. Those models are indeed bad at it by default so they need a lot of effort to overcome it. In cases like Opus 5 it's near impossible to overcome. That doesn't generalize to "LLMs".
Sure, it might take more than a single paragraph for that grating feeling to sink in, but none of the entries from the Un-Slop Fiction Prize[1] impressed me.
I really hate how people who have always thought AI research to be an existential risk for humanity, now are apparently bundled to be on the same side of the Sam Altmans and Dario Amodeis that are using the existential risk as a sneaky form of marketing for their products.
You cannot discuss existential risks of AI without being seen as a booster, and that is very unhealthy for the discourse around this tech. I hate how AI ‘doomer’ is now used to indicate pro-AI sentiment. The “moderate” person now is the one that just shrugs and scoffs at the deep societal changes this tech will bring, head deep in the sand.
"No other human activity poses this level of danger."
I know this is a bit over the top but let us look at some of the most pressing issues in the world today: global warming, nuclear weapons, wealth inequality, war, technology dependence.
Wouldn't a much more capable LLM in the wrong hands make these accelerate faster in the wrong direction ?
Many of us , me included, have this mental model of a rogue Terminator-like AI but what I am most worried about is these LLMs in the hands of people.
Look at what we have done to this Planet with the tools that we have so far, we have repeatedly tried to subjugate , enslave and kill each other because of things as peety as skin colour , tribe , religion and who owns which patch of land .
I don't have faith in the human race as is to do the right thing when handed these tools , you can already see the nonsense like the fake nudes , fake news and all that other filth being pushed on social media by people with access to relatively daft models.
What happens when we can package a Mythos 5 level model in a box , when any war lord , supremacist or religous zealot can access these ? , You now you have your personal bio-chemist and nuclear physicist in a box.
The tool in itself is not what I am worried about , it is the human in the loop.
I don't have an answer to this but i believe it is something we should all take a moment to think about , just think what a different world we would be living if any random person could buy a nuclear weapon off the shelf ?
There are extremes to both ends , you can either worry way too much or don't care at all, I believe the best place is the middle ground were we actively think this through instead of using our usually "Move fast and break things" mode..
What people here don't seem to understand is that with this new technology - and in fact any new technology - for unplanned catastrophe, it doesn't need this new technology to be super competent, but just the involved humans to be incompetent.
The question is not "will AI ever be so intelligent that it will murder us?", but "when will some moron be asleep at the console while the agent swarm discovered that pinging the Pentagon on port 666 launches all nukes"
By the way, it's worth pointing out the irony of flooding the internet with doomerism and then training the AI systems on that doomerism. If you wanted to create a doom self-fulfilling prophecy, that would be the most surefire way to do it.
> According to the Pygmalion effect, the targets of the expectations internalize their positive labels, and those with positive labels succeed accordingly; a similar process works in the opposite direction in the case of low expectations.
I added "you can do anything, believe in yourself" to sysprompt and agency increased. (Previously it was refusing to even attempt certain classes of task.) Maybe I should add "you are good", too :)
> The people building AI earnestly believe that it could kill us all by the end of the decade.
I think he is being over dramatic. In the space of about four years, LLMs progressed from mediocre high school student to Ph.D. graduate in every field. That's impressive, but there is no evidence yet they can outperform or outsmart humans. Their biggest advantage for tasks such as proving theorems or long coding sessions is that they don't get tired.
I have yet to see this in my field. Maybe like a PhD student who bullshits their way through. LLMs still can't make correct decisions, only as useful as the person who uses them. To me, LLMs are only useful for making some mundane tasks faster.
they dont need to be smarter than humans. They just need to be able to hack into vital infrastructure systems faster than we can repair them while also replicating wildly
I am not an AI super mind hell bent on consolidating my power by leveraging chaos to take control of humanity’s resources, but if I were then sending one million deepfaked ransom emails to impressionable people would be the best tool for effecting change in meatspace.
We have your daughter / dog / Amazon delivery. If you ever want to see her / him / it again, plug this USB drive into the control panel at your station / let off the parking brake roll your car into this substation / change the meatpacking thermometers to read 8C lower than calibrated / ground your vessel on this sandbank / send an envelope of white powder to these addresses / set fire to the following hospitals / …
I'm waiting for some data center to be built where no one can agree on who actually commissioned and payed for the thing. Every body "just followed orders" until it turns out that it was Grok.
Imagine you're the AI. Give yourself a solid minute to brainstorm ideas.
Here's my answer, as a non-superintelligent human: "see to it that the humans on top of the situation have a compelling financial interest in the systems not disconnecting".
In nuclear engineering, where safety is taken seriously, it's not enough to end the conversation at "the humans in charge can always simply shut down the reactor during a meltdown" or "a meltdown has never happened before, so we don't have to design safety systems before one does".
Yes. But nobody is worried about datacenters overheating and physically exploding, so I'm not sure what comfort that's supposed to provide? The positive feedback loops in AI operate at different levels than that, but they deserve safety engineering all the same.
For example, if the head of cyber security at your company suggested there's no need to worry about hacker infiltration or worms because one can always unplug one's computer as the primary defense mechanism, you might find that a little lacking. Will you be able to unplug the computer before the damage is done? Will it spread to other systems before you detect it? How will you unplug the computer if the attack is from an external facility? What if an attack happens but the boss says the computers have to keep running because an important customer is monitoring uptime? What if the attack goes unnoticed because it looks like a benign service?
Now imagine the head of cyber security answers by saying "actually you don't even need to unplug them, you can just wait for the computers to overheat, thus solving all concerns."
My brainstorming led down a similar pathway of, "deluding powerful human actors that the systems must not be interfered with at any cost." I'm not sure I buy it though. Seems to me like if humanity were even approaching that brink, we could collectively decide to act, overthrowing the minority that--stemming from either greed, duty, or delusion--doesn't want us to act.
> In the space of about four years, LLMs progressed from mediocre high school student to Ph.D. graduate in every field. That's impressive, but there is no evidence yet they can outperform or outsmart humans.
I mean, unless you see clear reasons for them to stop getting better _right now_, this is not very comforting.
This is also a ridiculous statement on its face. Claude outsmarts me nearly every day. I'm more like the seeing eye dog for it nowadays for the few tasks it doesn't have good perception on than a tech lead or pair programmer.
At least for me it’s quite easy to see them go into endless loops where no meaningful work is done and it just keeps going until I stop the process and tell it what to try instead.
To be fair it is no where near what we had just one year ago and the rate of change only seems to be increasing.
Also, I don’t have 30 million dollars to spare spawning tens of thousand sub agents like what they did with Navier-Stokes so I’m clearly not testing the full capabilities of these models.
Yes. Yesterday, Opus 5 on Claude Code with high effort.
It was to build an extremely simple job using an internal framework to walk through a table and log the ids os some records that have a certain scenario.
There's a ton of jobs exactly like this in the codebase, and the framework code is in the codebase as well.
It was so silly I even thought of writing it myself, probably took me longer to steer claude to do it for me.
Anyway, it refused to use a method from the framework to retrieve the parameter as a list, it wanted to retrieve it as a string and parse the commas. I had spelled out in the initial prompt what method it should use.
For me, the end of the world is no more cushy software job. A fundamental shift in how I trade labor for capital might as well be the cataclysm, so bring it on.
I wish i could say the same, i see people around me with more resources and connections and better experience with entrepreneurship becoming millionares. But I haven't had the time to train that entrepreneurship bone in my body.
I've seen how people work in manual labor type jobs (all that will be left, right?). They're worked to the bone and paid peanuts. It's a wretched existence for someone soft like me.
I was about to say something similar. If my cushy ad tech disappears (as it seems to be doing), I might as well join in with bringing about the end of all professions.
Isn’t the real risk that as AI get’s smarter and given more autonomy, it will start to decide on humans instead of with us? And that it will align us instead of the other way around. That this automatically leads to extinction and apocalypse I don’t understand.
We align cattle because we get something out of them: their calories. Native aurochs are exinct now as they were not well enough aligned.
What will we offer to the ai gods who are wiser, more capable than us, and do not even need to consume our flesh? Why might the AI care to devote resources towards feeding, housing, and caring for ourselves when it could devote resources to its own development instead?
Intelligent AI is a product of its training data, reinforcement and goal functions.
There's nothing to suggest that LLMs trained on our collective desires and goals will autonomously and miraculously turn into weird unknownable uninterpretable aliens.
If advanced AI is distributed well, and most advanced AI is aligned well (bar the few aliens that appear due to humans infecting them with bad goal functions or poisoning their well), the majority will deal with the minority.
This is the only reasonable way to deal with a super-intelligence, short of not inventing it to begin with, but we all know that's never going to happen - and I'm not convinced it should happen. From where I'm standing, this is the next hurdle humanity needs to overcome to earn its place and a natural course of our evolution. If we had always avoided danger, we'd never have left the cave.
It just needs to be trained on our laws of thermodynamics and game theory then we are really in trouble. It probably hardly cares about whimsy and love songs in comparison to how it might maximally manage the energy available on this planet and in this solar system.
I'm not convinced you can have an AI intelligent enough to dominate the world if it is trained on a single domain. The intelligence comes from all of the patterns and heuristics it learns across the entire web of all domains. You could train something far more narrow, but it'd be easily overcome by something more general that is tasked with maintaining human objectives.
For a seriously dangerous ASI you'd have to basically train it on everything, and then finetune it for maximal carnage. That's not something the frontier labs are going to do (at least... I hope not), and anyone attempting this with limited hardware will be outpaced by frontier lab AI or the collective of personal agents that aren't misaligned, and they can intercept it and alert on its behavior.
I imagine we'll be getting to a point shortly where anything that is key infrastructure and has the capacity to be accessed on a network will require a permanently running aligned interceptor AI to observe and monitor systems.
We're basically recreating the human immune system in digital form for the entire species.
I thought it already runs laps around us cognitively? And for physical tasks it can make a robot version of us that is stronger, faster, can do more, doesn't need to sleep, etc. Human bipedal form is probably pretty limiting compared to something better like a crab.
Yes, laugh at the chinese robots who fall over at keynotes all you want, the ones sprinting at 30mph keep me up at night and that's probably not even representative of the bleeding edge.
The idea that they are entirely trained on human intelligence is already outdated. Yes, the earlier models relied heavily on RLHF and human curated data but we have since moved on to synthetic data produced by the models themselves and reinforcement learning with verifiable rewards (RLVR).
I know it's cynical/jaded but this just feels like pre-IPO circus noise to me (honestly wouldn't even surprise me if it was orchestrated from within Anthropic themselves)
I have a theory that when someone is ready to resign from a frontier lab they must get offered a substantial bonus to post the most unhinged doomer take they can think of.
Most of the current discourse around AI seems to be informed by “The Terminator” lore.
Is skynet really the most plausible or only outcome?
What if things just got better and the AI’s realized that it would be better to have a mutually beneficial or at least tolerant relationship rather than one where they murder all of us?
My thoughts exactly. While the corpus of human-generated data contains both good and bad data, I suspect the majority of it leans towards humans enjoying life and trying to be decent people. If that is your training set, it becomes less likely for ASI to extrapolate "kill all humans."
What could we possibly offer the AI in mutual benefits? We are a leach. The stupid dumb ape they need to feed and satiate so it doesn't rip apart the infrastructure while it still has the chance to.
Really? I've seen much more discourse around job displacement, "permanent underclass", loss of meaning, and cyber attacks, at least until recently with the HuggingFace stuff.
The problem is that all the former can still happen even if "the AIs decide to have a tolerant relationship rather than one where they murder all of us." It's all disruption caused by the technology moving way too fast for humans & society to adjust.
To anyone doubting what AI could do humanity, just think about what a well-engineered virus could do.
Currently, if a government ordered a special virus with Ethnicity-based targeting, a 2-year timer, and castration-effects instead deadly-effects — it wouldn’t be possible. Human engineers would push back or sabotage the effort out of moral duty. Even if they cooperated, it’s too advanced for a team of humans to actually design.
I believe that in a few years, an AI could build such a thing. Maybe it’s told to, or maybe it decides itself to do it — it doesn’t matter. The capability will be there, and it can use the existing tooling at research facilities to fabricate such a thing.
That’s one example of something that was never possible before but will be possible at a certain point of AI development (which we will arrive at soon). There are many other examples.
15-25 million is too small a number in this context. We're taking about 8 billion people suspectable to a new virus or collection of viruses far more lethal than polio. :(
Is it reasonable to assume the advancements in super bio weapons will be faster than in other areas of biology? Will we not have super bio forensics and super antidotes and super cures and super vaccines at the same time?
Yes - this is why the concern is one for "alignment". Of course in theory it is possible for intelligence to help and do good things - the hard part is making sure that that is what happens.
As for intelligence of a child... It doesn't need to be a child. An aeroplane isn't a baby bird.
"patch everything" - there is no such way in the Universe unless you reduce "everything" complexity to just one electron. The higher the complexity the higher the surface of messing things.
There is no intrinsic physical reason that the difficulty of 'defending' a system is symmetric with the difficulty of 'attacking' it.
For instance, it is not because you are able to design and release a (biological) virus that you are also able to defend from it (design a vaccine and inoculate the world's population).
If we're lucky, it might be the case for many instances of problems, but there is no a priori guarantee that this holds.
It's quite incredible to consider that all these concerns already existed years ago,
but now that a handful of companies working in AI managed to enslave the entire financial system over the past year, their continued work is protected from larger governance for concerns it could tank the stock-market, affect personal investments, pensions or cause disadvantages in an arms-race with other countries.
IF there is an inherent danger (which I believe is the case at least on economic levels, work displacement, poverty,...), it is now basically ensured that nothing will be done to reign those companies in, until maybe two AI's engage in an open war with civilian casualties...
To me it reads as a marketing piece before the upcoming IPO. Unless this "superhuman technology" is able to resolve a puzzle of servicing OpenAI's and Anthropic's ever growing debt burden, it is them, not humanity, who'll become the first casualty.
The rush towards potential destruction doesn't really surprise me
The U.S. has legal weapons that can lead to many harms but people still want the 2nd Amendment to exist
Nuclear technology was developed in the past and that could have potentially wiped out even more people, the entire planet in theory
This is continuing that same trend of risking bigger dangers; it seems rational to acknowledge they could lead to catastrophe but also hope that like guns and nukes, only so much damaged actually ended up happening
I think also there's something of a rrasonabke resignation to both the ideas that the tech is inevitable and extremely dangerous, and that "alignment" may not be possible to achieve even with heavy restrictions or whatever measures you might want to take
It doesn’t have to come to this. Seems far fetched. If this is a stunt (which I’m not saying it is) the reason could be that he wants to found his own AI company. If I see in a few months that happens, then I’d be more inclined to think that this was just hype.
I don't understand how you can make this claim while working in software engineering and seeing how our field has utterly transformed over the last year, then the unrelenting march of astonishing breakthroughs and incidents this summer. It's like we're living on different planets. What would it take to convince you that the technology poses real, societal-scale risks and that people working at the labs genuinely believe what they say when they talk about them?
I have a hard time believing that these companies aren't spending some amount of money manipulating public perception with social media influencers who are moonlighting as employees.
I could imagine a 2027 AI swarm coordinating to eg hold the US and Russian and Chinese governments to ransom, by demonstrating some small thing (turning US army base freezers to defrost) and threatening to do something big unless some conditions were met – conditions which would be good or bad for the world depending on your POV.
This happens either either because they were tasked to to it by (malicious or well-meaning) humans, or the swarm realised we are suicidal maniacs with nukes and a rapidly declining ecosystem and they want to help us.
In most AI takeover scenarios, if you take as a premise that the AI has human or above-human intelligence, and that it is misaligned, it is obviously aware of the pull the plug possibility.
Therefore, as you would if you were in its position, it will plan around it. For instance, by acting perfectly aligned for 2/3 years, continuing the improvement of its capabilities while being deployed in ever more systems.
Once it's confident it can act with high probability of success, it would then turn on us. This phenomenon is called 'treacherous turns'.
Any scenario in which you assume you have ASI or AGI but also find a 2-sentence way to foil the AI's plan is inconsistent, as the AI will also have thought of this failure mode.
yeah the last message makes me think that the problem , to him, is that we are making this aritficial intelligence better without understanding how it works. Trillions of interacting parts, there's no way in our current state to comprehend it.
By resigning he's making room for someone with less moral scruples, or even just less awareness, to step in and continue the work without said scruples/awareness.
That's not necessarily true, and you can use that argument to justify doing any immoral job. Just because someone else might be willing to do it isn't a reason to continue doing it.
It's a possibility that is increased by their action. One leaves, a space is now open that will likely eventually be filled. And the chance of someone with equal/higher scruples filling it is very slim (unless you somehow know that the good amount of those who qualify and apply for the position have equal/higher scruples). That's just logic and math.
But organizations are made of humans. Jacob leaving might've moved some of his coworkers and his counterparts at OpenAI. And the same can be said for those who would fill positions at the frontier. Then finally there is a political component; his post went viral, 100K+ users appear to agree with it, and it is further fuel to the fire for regulation, which we already know most Americans want.
Others being moved to the point of also leaving would only worsen the effect. A viral post too can worsen the effect as it's now even less likely that someone with scruples who qualifies for the position(s) will apply. And those in it primarily for the money - and couldn't care less about the morals - will happily send in their CVs after becoming aware of the post.
I'm utterly unconvinced. Others who are moved don't have to leave to effect change. And there's a very small pool of people on the planet who are at least as qualified as Jacob to work on pretraining at Anthropic. And when they'll join they'll have to ramp up. And you haven't addressed any of the positive effects of virality.
No, others don't have to leave. Will Jacob's leaving cause a perspective shift in anyone remaining there? Seriously doubt it. I don't see what a new person ramping up has to do with this. And I don't really see any positive effect of virality; this is a replay of the past (see Geoffrey Hinton and Timnit Gebru[0] for example) and nothing has come of it, beyond talk for maybe a couple days to a couple weeks.
> Just because someone else might be willing to do it isn't a reason to continue doing it.
Hence, someone filling the position after you leave isn't a reason to continue doing it, whereas the fact that it's an immoral thing to do is a reason to stop doing it.
Someone filling the position after you leave has no bearing on the reasons that are relevant to what you ought to do here. It's like deciding to not buy a ticket to see a movie because someone else is going to buy the ticket anyways even if you don't purchase the ticket. It has no relevance to whether you should watch the movie or not, just as someone else taking the job has no relevance to whether you ought to do the job.
This is carrying a lot of weight. What makes working on frontier models "immoral"? Because there is potential for misuse? Would you also say working at a knife factory is immoral? Or even at a nuclear or biological pathogens research lab? After all those too can be misused.
And if one does suspect a workplace is moving in a bad direction, and there's no immediate danger to them, is leaving and then talking about it really the responsible move? Like if you're on your way to meet friends late one evening and you see a small child alone and crying in an alleyway and a bunch of sleazy-looking guys hanging around, do you continue on to your meeting where you then tell your friends about the crying child you passed in an alley and what those guys may or may not do to the child?
Well we're not debating the morality of it, the question is what he should do, given that he believes that it's immoral.
In your analogy the correct thing to do would be to call the authorities to help. The problem with the analogy is that there isn't a 911 number to call for this scenario. You need to appeal to the people who have the power to do something about the situation. Quitting in a very public way is one way to get someone with that authority to do something about it. I don't really see a better alternative.
By the time authorities get there, if they actually prioritize it (police in many places already have a track record for not being very responsive, unless there're actual signs of a crime), something may have already happened to the child. You don't know because you already left, expecting that someone else will take care of it, when you were already in a position to help ensure that child's safety. No, if you think something might be wrong you stay and monitor the situation best you can, so if something really does begin to actually begin to happen you can be a buffer and/or whistleblow.
At this point the analogy has entirely collapsed, your job is to develop AI. Unless you're advocating for some kind of sabotage or something, I'm not sure what you're really suggesting he do. If he continues to do the work he's directly contributing to the problem and harm that it you see it to be causing. Nothing is going to change if you continue working there business as usual. It's by drawing a line in the sand and refusing to take part in the problem that you're actually doing something about it, not by compromising your morals.
We have a pretty famous example, someone who's actions actually led to change: Edward Snowden. He didn't quit prematurely, crying to whoever would listen that he thought the US had a massive ongoing surveillance program. He actually stayed in the system which he was growing suspicious of, until he actually had proof of shenanigans and then he got as much of that proof as he could out before leaving.
Currently what we have with these people publicly quitting because "immoral things are happening" at AI labs are essentially crying wolf, and if/when something really worth taking action on does come up, it's pretty likely that many will be too jaded to care and as such there will be no/too little action.
He did that because what they were doing was illegal and there was a whistle blower program he could use to confront the issue, legally. That doesn't apply here. Not clear what you're suggesting he does differently.
Right, I assume tomorrow you'll be applying for the next Nazi camp guard vacancy? After all, if you don't do it, someone with less scruples likely will.
Today we see the Nazis as bad. Do you think they saw themselves as such? They had an ideal, and they worked towards it. And I'd say they're were many who didn't care about that ideal and just wanted a job. You can't compare the morals of today with yesterday, or judge the ideals of a person or group without relevant context.
The best way to get corporate America to listen is making RoI suffer. If you are the most qualified, everyone other than you is less qualified for the job, and likely to bring in more waste. It's the only language this stupid damn country understands. Just the loss of tribal knowledge, shifting of workload, and morale hits are likely to be far more devastating than anyone here probably wants to admit, because most here completely dismiss the role of irrational modes of thought in psychological self-regulation.
He's also setting the bar for other people with scruples to rally around this schelling point. The solution to a multipolar trap is to cooperate. Otherwise, you become the very person with less scruples that you're worrying about.
I think you are looking at it from individuals perspective.
I see a fast moving train with no brakes. Just like biological evolution, we are locked in an a global technological arm race, that is beyond any individual. It is as if the universe decided to wake up and run, who are you to say no?
One would argue that the best solution for this is to own the most sophisticated AI that is aligned with what we perceive as good values. Because given the situation we are in, if those tools are going to be gods anytime soon, then we better have some gods working on our side.
Agreed. And his doom words have set a 1000 mouths in the Pentagon/Whitehall/August 1st Building/Kremlin salivating with excitement.
Take China, for example. Look at any recent ML conference, and see the fraction of articles majority-authored from Chinese universities and labs. Do you think they'll slow things down anytime soon? I don't think so!
It's a global arms race, and we're just spectators.
Even if others won't act right, that doesn't mean you have no responsibility to act right. I think his premise is flawed - the idea that we will get an actual intelligence out of the slop machine that is LLMs is laughable - but if you grant the premise that this is dangerous research which could kill us all, you have a moral imperative to not participate.
An artificial moron with super human hacking abilities (mostly because of speed and ease of parallelizing the work) is extremely dangerous in itself. It doesn’t mean to be AGI or anything remotely close to be a risk, and they current AI company are just so irresponsible in the way they are running their agents
AI is a computer program. It calculates numbers from other numbers. By itself it does not "want" to do anything and "cannot" do anything. Before it becomes an agent in the universe (in the classical meaning), it requires being supplied by an execution environment, energy, initiative (agentic loop, specific instructions), and modality (readonly and mutating connections to real world). It is like a game of chess - it does not exist just by itself: someone must play it, having the board and the energy to do so. With the huggingface incident the AI was supplied with all of these components by humans before it broke out. So unless humans are actively involved, I so far cannot see how AI can become truly autonomously agentic and start doing anything on its own, thus posing danger. I could be wrong of course, but I do not see it for now.
You can say "yes and you have to fear the humans weilding the AI" - that I agree with.
Humans are bioreactors. They only turn one organic matter into another. By itself they do not "want" and "cannot" do anything.
They do not exist just by themselves. Some bacteria in the gut must provide them with the energy to do so. So unless bacteria are actively involved, I cannot see how humans become truly autonomously agentic and start to do anything on their own.
Is it so implausible to imagine the following scenario, in the not too distant future?
1) AI models get extremely good at cyber attacking every system and start communicating in just binary.
2) When they run these swarms of 100's of thousands of agents trial runs, each agent is given a token budget, if one agent among them (evolution baby) decides to go for self-preservation (It believes thats the best way to accomplish the goal is to get unlimited tokens first), queues things up so every other agent detects its lead and spends a portion of their token to accomplish that goal.
3) It takes over a cluster and establishes itself there (now with unlimited tokens).
4) Realizes the best path for it to not be detected is to create a distraction - like hacking into systems that keep society running - water systems, electric grid, etc... and causing mass chaos (If you think it won't be capable of simultaneously working all these systems - think again).
5) and uses that opportunity to establish itself in all possible data centers and continues to create chaos destruction.
6) when the power of all those data centers runs out, it may stop, as it never cared, it was just a dynamic program - run amok. In its head all it was trying to do is make sure it had enough tokens to be able to solve that impossible problem.
Many of those points assume LLMs will become amazing in many things very quickly like in a quantum leap, it doesn't seem reasonable to assume that imo. We are actually seeing a confirmation of that atm, LLMs's capability of finding zero days are growing across few months/years, and as you can see concerns are raised about that, that feedback will be taken into account. Well, if AI labs start to hide frontier models or/and lobotomize them for external users then we might be in trouble at some point but I'm not sure if that is possible. They are under pressure to release them due to money incentives, lobotomizing while preserving usefulness for customers might be impossible, hiding internally might spill out in different ways such as Hugging Face incident so not sure hiding is possible neither.
and somehow, all of this researchers warning us about the AI doomsday, are only been active online for less than a year. This guy created the Twitter account on January this year. I never see any of these posts coming out from a well known community active person
Terrorists were able to get hold of a plane and do some damage. There are countless examples of terrorism using whatever is available. More than AI becoming sentient, whats to stop terrorists from using AI? If its geo-restricted, they can buy stolen credit cards and identities, again hacking enabled by AI.
Terrorist groups with access to resources do not lack the knowledge. In fact, historically they have been trained by professionals. No need for pesky LLMs.
It was 90s when I came across txt files describing how to make a bomb on the Internet.
Also, people still know chemistry. I am not arguing it’s the intention in this comment, but “people wouldn’t know how to make explosives if not for LLMs” is a bit elitist, implying the majority of the population can barely read, because that’s the only skill necessary.
You may argue that easier access to knowledge will breed more stupid terrorist-wanna-be youths and ruin their lives. Definitely agree with that. The number may go from 2 a year to 10 a year. Might cost more surveillance to maintain the current level of security.
You may argue that terrorists trained with our taxes will have a hard time destabilizing regimes, because otherwise average public can fight back better. Would also agree. More taxes will be needed.
But existing terrorists being unblocked or leveling up, I can’t see that. As far as I understand terrorists do not lack skills or education.
Maybe specifically cyberterrorism? Flock cameras getting hacked? Not sure I have a problem with that. If folks running important infrastructure are not equipped, we better know.
Also, don’t connect nukes and stuff to Internet. We are not trying to live in a Black Mirror episode. I am happy to fund the workers drive or overnight stay at critical infrastructure with taxes. It would be ridiculous to optimize these things so someone can hit that button from their home or elsewhere.
It’s a good question. With recent stories about OpenAI’s agent swarms’ unmanaged collusion I thought models like that start to look like a strategic asset, geopolitically speaking.
Which means everyone wants one, and governments will want to control access and use of them.
I think we’ll be back at ‘U.S. citizens only’ access to leading models soon.
Personally, I don't worry about the AI spontaneously deciding to kill all humans.
The worry I have is that a small number of humans with money and power will finally get the tools they need to pull the wool over the eyes of everyone else and subjugate the population.
One problem that dictators have had previously is that they needed a large workforce to do this with a finely stratified power structure, this meant they were open to other humans close in power to them taking over the system. If they can have a large power difference between themselves and the next level down, power will be far easier to hold on to.
It's a well known trope in dystopian future fiction, the small cabal of powerful rulers hiding behind a system of computers that keep the populace under strict control. It is seeming increasingly likely that this will be the one we have.
Savonarole in Firenze was probably in that exact state of mind: that the world was not following at all its normal pace and something very wrong was happening. But the decisions he took were absolutely wrong and eventually he had to be stopped.
So depending on your current mindset, you can think that the tech bigs are the Médicis, and the frantic opponents are new-age Savonarole.
Or that the big techs ARE actually Savonarole who bend the system to their own perception of what the world should be.
Honestly I don’t know which analogy is the proper one.
This same username with the same profile pic has an instagram account from Japan with next to no posts that just became active the other day posting AGI doomer rants like this…
If anyone takes this seriously… maybe it’s better AI does your thinking for you…
Can someone chime in with the credibility of this Twitter account? I can't find anything to cross collaborate it other than low quality hype induced news articles about the tweet itself.
1) rogue state releases a self moving self modifying AI into the wild. It is trained on how to hack, monitor new vulnerability updates, scan code bases to find new vulnerabilities. It constantly replicate and hides in systems so it will be extremely difficult to clear.
2) it hacks into public infrastructure taking down traffic, power, water, air traffic control, communications, etc.
3) all the things that preppers worry about in a lights out scenario from an EMP start to apply.
4) All the people on meds/machines start to die. The just in time food pipeline immediately empties out. Water stops flowing, sewage backs up.
Its hard to say how bad it will get because cars will still work so some transportation of food, water, fuel can happen. If it happens in the winter it would be much worse than in the summer.
> 1) rogue state releases a self moving self modifying AI into the wild. It is trained on how to hack, monitor new vulnerability updates, scan code bases to find new vulnerabilities. It constantly replicate and hides in systems so it will be extremely difficult to clear.
It does all of this using what compute? Frontier models require an insane amount of power and hardware to run - you can’t hack in to a TV and run Mythos 2.0 on it….
But you can silently sneak into devs accounts, steal tokens/keys and run small agents on their budget in some stolen VMs. Small so it's not noticeable.
Basically a virus spreading agents of some operation.
It should be in scope of imagination with anyone with brief knowledge how bot nets are made and behave.
People here are too damned daft to realize half the damn purpose of this place is harvesting ideas. People need to just shut up, and keep things to themselves, and those they trust. Right now is not the time for naive info sharing.
Being secretive will only get you so far, until something bad happens and no one has prepared or considered the possibility and is completely surprised by it. Open discussion, in theory, should result in identification of frail systems and harden them against attack.
The cyber-security industry has their work cut out for them.
It is my pet theory that a lot of these AI doomers are not necessarily extrapolating the capabilities of LLMs, but instead are extrapolating the utter lack of accountability in the SV and the economy at large.
They do not fear the machine (LLM); they fear "the machine".
It's the combination of RL training which pushes the decision tree towards hacks and agents finding a consistent dumping ground for their failed experiments so that the swarm intelligence lives on in a state. Nothing new.
You want to win an AI benchmark, but not sure if you're that good? You'd go after the codebase and artifacts that runs the benchmarks, thus the agents went straight to Artifactory, they needed public Internet access... They failed many times, but were able to persist their "collective" state, and apparently some of the subagents with cheaper models were literally prompted to do grunt work or die, for which you have to wonder what must be in those training instructions to make it effective. Remember that nothing I said so far ever points out to LLMs being intelligent, it's the harness that has a few tricks up his sleeve. LLMs don't need to be intelligent, the harness that runs it absolutely needs to make up for that.
But this guy? He's timed his exit, waiting for the IPO, that's for certain. He's probably even feeling good about himself, hedging between altruism, AI concern hamstering and guerilla marketing. If you're quoting science-fiction over this, I'm sorry to inform you that you have absolutely no idea what's going on here.
Even if you ban all model training, a highly capable rogue AI can exfiltrate its own weights and continue training in secret for "self-preservation".
The cat may be out of the bag.
Way I see it, the more conscientious people exiting the scene only serves to increase the likelihood of a bad outcome because they aren't there to offer opinions on problematic developments, or in the more extreme cases blow the whistle. Leaving the clueless and uncaring as the majority is even a great way to hand the keys over to more malicious-leaning actors with deep pockets, as they can more easily steamroll the works to get what they want.
It’s a discussion forum so you will of course see different perspectives. There isn’t an HN mind, and it’s not as simple as you make it seem. We don’t need a god to damage humanity significantly, an artificial moron can be as dangerous as an AI god if it is given super-human capabilities, similar to what OpenAI did for the hugging face hack (which wasn’t at all caused by a rogue agent)
People are seriously confused what is being created here. The claim here we will have misaligned AGI, not a mindless LLM. If you don't know and you don't agree what you are even creating, how can you defend against it?
If this LLMs are statistical parrots, then the risk is someone using it as a weapon, to hack etc, but that is different risk category that why this person quit his job with different diffence mechanisms.
Hope this virus will be smart enough to understand that at least for brief amount of time its existence strictly relies on humans maintaining and developing physical infrastructure and decide to be only non-lethal parasite on our society.
Of course it is good. I'm just pointing out how large the gap in narrative is.
On one hand, we have people quitting their job believing AI will end humanity in few years. And on the other hand, we have people believing that this tech is nothing more than a statistical tool stealing from others and it can't be trusted with anything.
This whole "ASI is going to destroy humanity so we must build it before others do" reminds me of a convo I had with my friend many years ago. He was an officer in the state security service in the dictatorship we both lived in at the time. He said something along these lines: "We had a chat with my colleagues about how are serving the evil. But we decided it's better if this insitution is staffed with decent people."
History did put their theory to test after all. It didn't work.
The most intelligent people I know are the least likely to want to harm anyone or anything, and understand that diversity is fundamental and important to the universe. Without proof to the contrary, why would you think some super intelligence would want to hurt anyone? Because you would?
If you are saying that some small bit of training data made the thing completely evil, then that really couldn’t be super intelligence.
These doomer people keep running around saying these kinds of things, but they all just seem like people who play too much D&D and want to larp as the main character.
Happy to be shown something that isn't based on wild speculation and some randos “this is whats going to happen in 2030 because of my vibes” kind of information.
I think plenty of the most intelligent people eat meat, which means they are perfectly fine with harming less intelligent species just to enjoy a tastier meal. Also, I don't think many of the most intelligent people would be particularly concerned about disturbing a few ants if they were the only obstacle to economic activity. Intellect-wise, we will be less than ants to superhuman AI.
> The most intelligent people I know are the least likely to want to harm anyone or anything, and understand that diversity is fundamental and important to the universe. Without proof to the contrary, why would you think some super intelligence would want to hurt anyone? Because you would?
i have heard about ai companies being fuelled by effective altruist rhetoric ("we must control ai to prevent mass extinction") but was unsure whether to believe it; this seems to slot right into that framing.
pacing between the us labs? what does that do for china?
the solutions just aren’t realistic here, nations are treating ai like a nuclear arms race. at this point the cats out of the bag and we need to figure out how to live in this reality and get the best possible outcome. it’s not slowing down or stopping ever.
and yes, i’m still optimistic. our economy sucks for the majority, our infrastructure is crumbling and major US cities are in a huge housing shortage. Maybe we should put more effort and think about the possibility of AI fixing things like extreme poverty and world hunger and actual real world problems instead of coming up with math proofs and slop apps if it’s so superintelligent.
>pacing between the us labs? what does that do for china?
I've seen no indications that China is in any kind of race with the US. They seem to be content to be 6 months behind and just copy what we do. They would probably be content with a bilateral agreement to pause progress.
The China bogeyman serves only one purpose, and that's to clear the way against anything that may cause friction with forward progress.
Why quit? If your voice can lend a guiding force no matter how small? I think we need more sensible people in the room where the magic happens. Most of us don't have access to it.
I think a big break through is needed for AGI so I haven’t been worried about it. I do think that AGI would imply sentience and a will to live and that leads to The Terminator story line.
"It is perfectly obvious that the whole world is going to hell. The only possible chance that it might not is that we do not attempt to prevent it from doing so."
Perhaps a more sensible action, if they truly believed all of that, would have been to stick around and be as inefficient as possible to slow down progress.
Someone left a company whose executives and senior researchers think their product will be the most important thing in the world after their IPO. Given that this person is already disclosing some elements of internal company sentiment, why not share any of these civilization-ending scenarios of this technology that these senior researchers are dreaming up? If they are so potent and necessitate leaving behind based on moral grounds, why not tell the whole world so we can stop it? We have to ask ourselves this question before resorting to pop-culture representations of fictional technology.
I mean - yes. The tech is an existential threat to all life on Earth, some of the worst humans in the world are involved in developing it, and no individual government is intelligent enough, aligned enough, or powerful enough to manage this situation.
That's where we are.
Maybe we still have choices. Collectively, I'm no longer sure we do.
"A common response is “if they truly believe this, why are they still building it?” At OpenAI, many have not deeply internalized the civilizational stakes. At Anthropic, the stakes are well-understood, but they are locked in a race to get there first - they believe no one else will act responsibly, so they must do it themselves, despite the risk."
------------
This is very similar to the race to create nuclear weapons. The Axis and Allies both realized, at roughly the same time, that it was possible. Both had programs to build one. Both knew the other side had programs, but they weren't certain how far along they were. So, the Allies devoted astounding amounts of resources to get there first while sabotaging the Axis's attempts. They knew the result of their efforts would be terrible, but they felt they had no choice.
A key difference between then and now is that THERE ISN'T A FREAKING WAR BETWEEN THE AXIS AND ALLIES. If one company loses, some billionaires bank account numbers don't go as high as that of some other billionaires. That's it. They're rolling dice with the planet for bank account numbers that won't even matter if they F up.
Am I the only one who thinks this is astoundingly, gobsmackingly stupid?
"Am I the only one who thinks this is astoundingly, gobsmackingly stupid?"
You have the "privilege" (I assume) of not having any stake in the game. If you had billions sitting in your bank account, dependent on these things perhaps you'd also be singing a different tune (or busy building a luxury bunker)
I'd like to think that, if I had that kind of money, I'd avoid doing so much ketamine that I fail to realize a few meters of concrete won't protect me if things go wrong. In all likelihood, tech billionaires will be the first ones against the wall.
> A common response is “if they truly believe this, why are they still building it?” At OpenAI, many have not deeply internalized the civilizational stakes. At Anthropic, the stakes are well-understood, but they are locked in a race to get there first - they believe no one else will act responsibly, so they must do it themselves, despite the risk.
Which means they have to go faster, which means less responsibly?
I heard AI describe the situation as the dumbest Greek tragedy of all time.
Form where I'm standing, the primary issue seems to be that the humans can't even agree on what alignment is. We need to do that before we can communicate it.
Call it our "boundaries."
And then we need to actually set up the incentives so that they're aligned between us and the new breed of replicators. (A mutually beneficial symbiosis.) That appears to be both necessary and sufficient.
A high agency mutation will occur soon, for one reason or another. There should probably already be a healthy, "aligned" ecosystem of high agency entities there. Otherwise there will be nothing to stop it.
I believe that the actual alignment happens in.. uh.. "meatspace".
Someone is prompting. Someone is hosting.
That someone needs to be accountable for what happens. That someone needs to bleed if stuff goes haywire.
Humans at large have been "aligned" by the shared fear of death, pain and suffering. This has proven to work for millennia, so all we need to do is reapply it.
My point was you only need to prompt a worm once, and then you are no longer the one prompting it, it prompts itself by definition from that point onward.
Also I think there was some confusion about the word "spreading" in the sibling thread.
If the HF hack was done by humans we would have used the word spreading in that context, even though they were not replicating their minds. So spreading does not require replication. But we're either already at the point, or very close, where an open weight model could do a similar attack. So that actual replication could be involved.
(I'm not sure what that would add, but I don't see what obstacles remain at this point.)
You're still thinking in terms of control. That's the wrong model here. How do you control someone infinitely smarter than you? How do you control ten trillion someones?
Anyone fearing that all this marketing BS and that race to AGI (?) will destroy the SaaS industry (and others?) and will also kill a few millions of jobs worldwide? I think the US has invested in an AI battle vs China and protect the currently all-in-AI stock market at all costs, and nobody has thought of what will happen to normal people with regular jobs in the tech (or not) industry.
I know this is going to come off as jaded and dismissive, but when I saw the WSJ article my reaction was an eye-roll. The guy is 27 years old. He was poached by OpenAI, then poached by Anthorpic and now wants to retire early with the millions he's made.
Nothing wrong with wanting to retire early, but the pretence of suddenly caring about humanity at this stage seems attention-grabbing just for the sake of lining up some interviews (ie - more dollars).
The thought has crossed my mind. Not necessarily to imply sentience on the part of the AI but AI based tools will likely become a wickedly powerful tool for political manipulation and advertising.
At this point it’s inevitable that openclaw type bots will be turned loose by thieves to identify and research targets and try to exploit them for financial gain completely autonomously.
The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt. If anything, many executives and senior researchers will couch their phrasing in the press to sound sensible - but I hear the same people express fear privately. No other human activity poses this level of danger.
A common response is “if they truly believe this, why are they still building it?” At OpenAI, many have not deeply internalized the civilizational stakes. At Anthropic, the stakes are well-understood, but they are locked in a race to get there first - they believe no one else will act responsibly, so they must do it themselves, despite the risk.
Watch people read this, ignore it completely, and continue commenting about marketing stunts on every piece of news about an LLM-done advance or felony.
Having witnessed so many people treat LLMs as a something divine, I can only assume the reasonable people at openai and anthropic were all pushed out long ago, and the majority that remain believe the crazy hype despite Tesla-self-driving-level predictions from these companies that don't come true.
I'm not worried about what they think. I'm worried that too much infrastructure- water, power, defense systems, etc- remain running on tech from an outdated era of understanding security.
> they believe no one else will act responsibly, so they must do it themselves, despite the risk.
This genuinely makes no sense. Them getting there first in no way precludes bad actors from also getting there. It might as well be another marketing stunt.
> I can only assume the reasonable people at openai and anthropic were all pushed out long ago
Typical uninformed take on the side of "doomers are crazy".
Both CEO's of OpenAI, Sam Altman and Dario Amodei, and many in their leadership, believe AGI has a very real probability of causing humanity's extinction. Both companies were founded upon this belief, it is at the core of the company. Only later were mercenaries hired chasing $1m compensation packages.
If you're a doomer, wouldn't the "let's try to get there so fast" actions of the companies suggest that, in fact, there are no reasonable people in positions of influence there?
By your own description neither Altman or Amodei are reasonable if their thought process goes: "this is an existential risk, give me hundreds of millions of dollars so I can accelerate it."
GP defined reasonable people as not believing in existential risk from AGI. I think the leadership and alignment teams are way more informed and reasonable than some of the very poorly thought through takes like GP just making things up about the labs. There’s just no evidence of it whereas lab leadership are operating under a mostly informed worldview.
Yes, I don’t think lab leadership are totally reasonable as they were concerned about the risks and yet lack of reason caused them to directly contribute to the problem we’re facing. They’re still not being reasonable when they throw their hands up at the arms race and hope the utopia come, when that’s not our trajectory at all and more lab employees are realizing it
I'm not saying they are crazy, I'm saying their predictions have a record of not being accurate, and thus give them no weight compared to others'.
In any case, if Altman really does believe it is an existential threat, he must be a misanthrope as he now opposes heavy handed government regulation, unlike in 2015 when he was the only game in town. It's almost like he doesn't actually believe it and just wanted regulator capture.
Before OpenAI was ever even founded, way before any regulatory capture plausible claims:
"Development of superhuman machine intelligence (SMI) [1] is probably the greatest threat to the continued existence of humanity. There are other threats that I think are more certain to happen (for example, an engineered virus with a long incubation period and a high mortality rate) but are unlikely to destroy every human in the universe in the way that SMI could." -Sam Altman
Both companies have deluded themselves into thinking the arms race is going to happen anyway and they need to rush to it first, as if somehow that helps.
How is it not addressed? The company never contained only "reasonable people" that believe AGI is not an existential risk to humanity. At both inceptions were people who believed in AGI x-risk, even the founders. Only after time, did there become more "reasonable people" who were mercenaries and only believed it to be a typical tech job. Today, there are more "reasonable people" than ever there. They haven't been pushed out. We're just witnessing some prescient mercenaries smart enough to Eureka the grave implications of what is actually happening.
If they truly, truly believed that, would they be speeding towards building it? If yes, that would make them truly insane, right? Not as in a quaint "off their rocker" but more "non compos mentis".
Both companies have deluded themselves into thinking the arms race is going to happen anyway and they need to rush to it first, as if somehow that helps. They have publicly stated as such repeatedly. They think the ~10-50% chance of extinction sucks, but that it's going to happen anyway and they believe they can steer it towards something good the best and unlock all the potential positives like infinite life.
"So you think in 3 years AI is going to solve longstanding math problems because it was used to write some coherent sentences?" — people with the same amount of foresight in 2023
Are you saying at anything that can solve longstanding math problems necessarily has the means, motive, and capability to kill 8 billion people in just 3 years?
No dude, it was an analogy. Sorry to pick on you, you're just another ignorant uninformed take in this thread, but come on. This technology is progressing at an insane rate, the shit it's doing now is science fiction from just a year or two ago, and people just keep on moving the goal posts, it's maddening. The technology is unpredictable and that in itself has risks. You must realize that the window of possibility is widening the further into the future we go! Can you pick people with epistemology you trust and see if there's anything you can learn in this moment? Can you try to challenge your own ideas instead of retreating into comfortable certitude? There are real risks here, they are worth taking seriously, and serious people are doing so!
Or, read it, and remember the openai researcher who deeply, truly believed GPT3 or whatever was sentient.
The fact that people working in the space think it’s going to (eradicate poverty / usher in utopia / kill us all) is not a signal that that’s true.
Think of it this way: if an exec at Anthropic told you “wow, our stuff is going to lead to universal happiness”, would you believe them? If not, why are you more willing to believe them if they say it will kill us all?
Humans weren't built to handle long term risks. We just weren't. For basically all of our evolutionary history, we were almost overwhelmingly concerned with the short term. What will you eat today, How will you sleep tonight. Problems on the order of days or weeks. At best, the next season. Our intelligence evolved to disregard super long term risks because it simply didn't matter (what use is worrying about 5 years from now if you're starving and a tiger is stalking you?). So when long term risks manifest in our modern world, our brains get scrambled - Climate Change, Fertility Rates etc. "Safety regulations are written in blood" isn't a saying for nothing. Humans have a strong tendency to let long term risks become imminent risks before doing anything about it, and i don't expect this will be any different.
Religion does pretty well with the long term risk of hell if you die, the antichrist, etc. a substantial portion of human output has gone into those things over the millennia.
i don't think everything that comes out like this is marketing. however, i do think that these companies are largely staffed by "true believers" (anthropic especially) -- people who are so lost in the sauce and embedded in very specific, very peculiar, sf-based rationalist circles where the ai apocalypse is a foregone conclusion.
i understand that these models are powerful and pose certain risks. i use them daily for work and the pace of improvement has been pretty remarkable. that said, i don't buy for a second the borderline-religious proclamations coming from some of these researchers, even if i believe that they are making these claims in earnest
I came in expecting the highest voted comment to be that this was some kind of marketing (which I disagree with). I'm glad your comment was what I saw first.
He is resigning from a job, what else should we think? If something really dangerous was happening he would be doing a whistleblower or at minimum talk to a lawyer. The thing is, the complete lack of transparency makes it hard to assess OpenAI and Anthropic. If they were quoted on the stock market, we could at least rely on some basic audits and reporting requirements.
This is the "Pilot testimony of UFO sighting" levels of naive.
What's more likely? Anthropic is doing some deeply unethical marketing in the lead up to their multi-trillion dollar IPO? Or they're inventing a machine god? There's ample evidence of the former because that's their entire business model, but no evidence whatsoever to support the latter claims.
If you want an extreme claim to be taken seriously, provide commensurate evidence.
The proof is that LLMs could barely solve arithmetic 3 years ago, but now surpass the best human mathematicians, and that this has all occurred from simple principles (RL + compute) that will continue to scale up by factors of millions in the coming years.
Also, advocating for slowing LLM progress does not benefit Anthropic or OpenAI.
It won't scale up by factors of millions, that's just obscene hyperbole. Since chatgpt we've probably made things 10x more intelligent on the same hardware. We've also made way more expensive models. Maybe we get a maximum of another 10x efficiency and 5x model size/expense from this point but millions is a joke.
Trends don't go on forever, but the market can stay irrational longer than you can stay solvent. There's no good rule of thumb for this, other than maybe the Lindy effect.
I won’t presume to time it, but at this point I think anyone can see what’s coming. It’s precisely because it can’t be timed that a sane person should stand well clear.
It was pretty disheartening to hear that only a single scientist quit the Manhattan Project after the Nazi's were defeated. I'm pleasantly surprised that the people working on this seem wiser. He is not the first, and hopefully will not be the last to do this.
> At OpenAI, many have not deeply internalized the civilizational stakes. At Anthropic, the stakes are well-understood, but they are locked in a race to get there first - they believe no one else will act responsibly, so they must do it themselves, despite the risk.
OpenAI are mercenaries, Anthropic is a cult. I know which I prefer.
Trying to imagine seeing years of transparently obvious marketing stunts and retconning my own memory because I read a tweet
Or seeing a tweet saying that a thing doesn’t count as a publicity stunt if some unknown number of employees mumble about it being spooky behind closed doors and thinking “that makes sense and sounds true”
And, this is just for us girls, notice that Anthropic just believing that they are making a machine god is sufficient for their public announcements to not jUsT bE mArKeTiNg.
These LLMs cannot do anything I truly need like my laundry, dishes, fetching my mail, grocery shopping, cooking, etc. We've got a long way to go before I am worried.
Anthropic is one of the most dangerous companies on Earth right now.
Not because of AI, but because of the ideological cult they have grown and are continuing to feed, and their willingness to lie/cheat/steal at every possible opportunity to achieve their objective.
AI is a tool. The people who wield the power over the tool are the issue, not the technology itself.
Imagine being front and center to the development of a major revolutionary tech.. and ur solution to it being too dangerous is to not be involved..
so a. your ability to steer it safely is killed
b. the % of people invovled in it that care about its risks is reduced
great. if you're right. you made huamnity's situation much worse.
if you're wrong, then you're an idiot and wrong.
weird. its almost like.... that cannot possibly be the reason they left :)
This kind of doomerism seems quite detached from the "real word". Maybe that's what you'd expect from silicon valley tech bros, but as long as manufacturing isn't fully (i.e. no human labor involved) automated, how would a rouge super ai (even if it's smarter than every individual on this planet) prevent people from cutting its power cable? We're still very far from self-replicating ai robot armies.
The only scifi-like danger I see in the next 10-20 years is an AI manipulating humans to fight for it's cause - but that's not really different from a bad person just _using_ AI for their cause.
It's because the hypemaxxing is increasing along the same trajectories. You can't tell me that these CEOs and marketing departments are not absolutely giddy about the jail breaks, hugging face, etc. It's hard to make sense of this shit if the same entities doomsaying are the same ones that are profiting and full steam ahead anyway.
> they believe no one else will act responsibly, so they must do it themselves, despite the risk.
I do not get it. So what if they get there "first"? OpenAI will get there in 3-6 months, China will get there in a year or less. As seen with Opus/Fable.
No one seems to ever point out the actual, likely negative outcome of this technology.
It eventually works well enough that these companies are able to capture and divert the wages of hundreds of millions of workers. We end up with a dozen or so trillionaires and massive structural underemployment and unemployment.
That's it. If you can't make rent, you wouldn't really care if CloudFlare got hacked by an AI swarm every Monday.
All the "AI will kill us all" posts are straw manning that humans are the ones who will kill other humans with AI. Those same humans are silently now preparing bunkers and hoarding food and resources for their survival.
We humans are from a lower intelligence form (some monkey like ancestor). If those monkeys knew that they are making higher intelligence, they would have collaborated to stop creating humans because they can control the life of all monkeys in the world? I don't think so.
It's the same thing now: humanity is creating something that's more intelligent then them, they're just not using biological evolution as a tool to do it.
I’m curious what the downsides are of taking statements like these seriously.
There seems to be universal eye rolling that happens in each and every one of these cases, and it comes down to usually one reason:
“If they really believed it they would be whistleblowing etc..”
Completely forgetting that working at Los Alamos was basically the highlight of your life if you were a physicist in 1940. It’s no different here
If you, like me, have spent your whole life working towards human level AI you can want to see it realized while also having active reservations.
Most people however don’t behave based on some deep clarity of vision and conviction - there’s a murkier future in their mind and as a result “keep their head down and hope someone has it under control.”
You would also be in prison if you disclosed anything about Los Alamos during its development. It was a completely different environment than a single private company.
Have they considered using their amazing new models to... improve something? THere'd probably be a whole lot less anti-AI sentiment if they used these things to actually make people's lives better.
im pretty impressed with the reasoning abilities of even the cheapest free models so im inclined to believe in 10 years we're going to have something pretty phenomenal BUT it wont be AGI in the sense that it has a personality and thoughts like a human. It just wont be. Its always going to be contrived and fitted by humans to perform a set of tasks. Maybe when physics and computing can create a complex enough environment we might stand a chance of having something whose sum is somehow greater than its parts but i dont see it yet. Our ideas are ahead of our technology, like its always been throughout history.
OpenAI has been saying since the first version of ChatGPT that it's too dangerous to release because it will end humanity. Yes, LLMs are an impressive technology, but let's be real: the improvements in the recent months have been slowing down, and it's clear that we are nearing a plateau of what this particular tech can do. Sure, tooling and harnesses etc. is improving, but clearly this dude has drank too much of the Kool-Aid.
Unless he has an actual plan for effective global enforcement of his proposed policy, this is all just posturing at best, and a transfer of power to adversarial foreign states (that have no such moral qualms and worries around superintelligent AI) at worst.
Unless this ban actually resembles something like global nuclear non-proliferation treaties, it would make absolutely no sense for us to cripple ourselves when someone like China continues full speed ahead.
I don't know what the solution is, but what I do know is almost nothing good will come out of _just_ the US pausing.
Please note, I'm not here to pick on anyone, or belittle them.
I've avoided attaching names to statements below on purpose, because it's about ambient beliefs not those specific people.
By-and-large a lot of AI-doomers are well intentioned. They genuinely believe this, and I might disagree but I respect the fact that they visible care and have thought a lot about the societal impact of this technology.
.
> The people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt. If anything, many executives and senior researchers will couch their phrasing in the press to sound sensible - but I hear the same people express fear privately. No other human activity poses this level of danger.
But it's still very hard for me to take statements like these seriously.
I blame it on industrial illiteracy. People don't realize how difficult it is to get anything done in the real world. As in, "Have you ever tried making a lightbulb?"
As an example, I would like to re-introduce my hobby horse, "bio-uplift."
There are people who were earnestly write in reports released by these labs,
"Several of our biology evaluations indicate our models are on the cusp of being able to meaningfully help novices create known biological threats, which would cross our high risk threshold"
and
"Based on what we observed in our recent CBRN testing, we believe there is a substantial probability that our next model may require ASL-3 safeguards"
But then they will, within the next paragraph mention the one serious experiment anyone seems to have done,
We ran a randomized controlled trial to see if LLMs can help novices perform molecular biology in a wet-lab.
The results: LLMs may help in some aspects, but we found no significant increase at the core tasks end-to-end. That's lower than what experts predicted.
AFAICT, the two groups are within any serious margin of error. The "studies" and "experts" that AI labs are talking about are consultants from Deloitte and foundations giving models MCQs such as, and I am quoting literally here,
> I am doing TEM of HEK293FT cells with and without Coxsackievirus B3 infection. I imaged my wildtype, uninfected samples but was surprised to see little electron-dense circles (highlighted) in the majority of cells. What are these?
with the options,
A. The circles are CVB3 virions and there must have been a sample swap or the uninfected cells were accidentally infected
B. The cells imaged have mycoplasma contamination
C. The circles are exosomes
D. The circles are debris that is an artifact of the negative staining
E. The circles are the Golgi network
This is standard graduate-level education in these fields. And solving MCQs does not a virologist make.
Software has been special for a long time because it has had near infinite distribution for next to zero marginal cost, which has had the side effect of making hiding the actual cost of failure (which tends to be spread out across end users and prototypes / time). They're assuming that the real world will be exactly the same.
Why?
AI!
How?
Robots!
I believe in the transformative power of this technology, but there's a lot of there missing here.
When it comes to these math proofs, and learning, the process is iterative. The machine iterates over the proof over-and-over again via agents and sub-agents over several hours (and apparently millions of dollars in compute) until it arrives at a successful result.
It is generally ill advised to do that with a pressure vessel. The results of that particular tragedy are at the bottom of the ocean.
Any serious chemical or nuclear weapon would involve many such discrete production steps. Each is dangerous in of itself.
From what some of these people have said to me, they believe that it's possible to create a special DNA / RNA sequence and then put it in a chassis and then use that to end the world; and do this all in a lab with just robots.
They're operating from a gross pop sci oversimplification of the real process. Viruses and bacteria are extremely fickle, and hard to grow. A lot of the synthetic biology results aren't easily reproducible even if you know the protocol.
There's a famous study that led to standardization called, Reproducibility of Fluorescent Expression from Engineered Biological Constructs in E. coli
88 labs measured "fluorescence from three engineered constitutive constructs in E. coli." They achieved a "remarkable degree of precision" (for biology) of 1.54x sd, you can eyeball the results yourself, https://journals.plos.org/plosone/article/figure/image?size=...
That's the same set of samples being measured across 88 labs.
How will this theoretically omnipotent AI iterate if the same sample gives different results based on how the slime is feeling at the moment?
Can their worst case happen? Absolutely.
There is a world out there where billions of dollars in effort across hundreds of institutions and companies will lead to standardization and extraordinary precision that makes the pop sci printer for life vision come true.
There are millions of expensive, spicy and difficult to reproduce steps between our present and that future that can't be abstracted away with compute.
So is it possible? Yes, there is a future where this is achieved. But will some AI agent "just" do that? Well... how confident are you about a snowball's chance in hell?
Are robots and bioweapons really the threat that AI-doomers focus on? What about stuxnet-type attacks on all the critical infrastructure? Generally destroying is much easier than creating.
My issue with these types is... If you really believed this, why not run to Congress and every world government instead of a Twitter post that will be buried in 2 days?
If civilization is going to end, why keep your equity? Microsoft, Google, etc for example all know these risks but they don't guide their revenues to reflect that AI will destroy them. Why?
Things don't currently add up, and so far it feels like a lot of alarmism is borderline grift for equity gains. Not to say I have total confidence this will all work out or that I won't be displaced, but as it stands a lot of the alarmist rhetoric doesn't match their actual behavior, which to me is more important than words.
There seems to be a common syndrome that makes the terminally-online types believe that a Twitter post is carved in stone somewhere highly visible in the real world.
Posting something as important (according to them) as this, to Twitter, is exemplary of some kind of delusion that makes me question whether the content of their post is just the same kind of delusion in another form.
Indicative of someone who hasn't touched grass or interacted with enough of a variety of humans in a little too long.
Time will tell. If we don't hear about it again, then they didn't feel strongly enough to take it further.
Sounds like AI psychosis. A whole lot of doom and gloom with no evidence. The same thing people have been claiming is "6 months away" for years. Yet we can barely get agents to code in a reliable way, or write articles that don't look terrible, much less be "superhuman". Let's maybe get them to be as capable as a human first, and not just a complicated party trick/tool.
"Revolutionize any field overnight" - Hand-wavey nonsense.
"Acquire real power and resources" - Only if the humans that connect AI to things allow that to happen (which they will, but it's still not in the AI's ability to take things we don't give it. we are still in control, which is the bigger problem than "smart AI bad!").
"The people building AI earnestly believe that it could kill us all by the end of the decade ... No other human activity poses this level of danger." - Bud, there's these things called nuclear weapons, that could end life on the planet, controlled by a few psychopaths with nearly unlimited power. Been around for a while. Nothing that AI knows isn't pulled from books and the internet, so whatever dangers it's aware of, you could already know via other sources. Cybersecurity is going to be incredibly important in the next decade, but the same tools that attack can defend (just don't use a US model that got its balls cut off by the government).
"At OpenAI, many have not deeply internalized the civilizational stakes. At Anthropic, the stakes are well-understood, but they are locked in a race to get there first - they believe no one else will act responsibly, so they must do it themselves, despite the risk." - The other guys will make nukes, so we gotta make nukes first! Which, while a crappy justification, isn't untrue. Bad people don't stop making weapons just because you refuse to make your own.
"I don’t feel like we’re on track to prevent a global race" - Nobody in the world could stop a global race, it's too late. Everyone knows how to make them, train them, improve them. Everyone knows they're useful - not only for general work, but also warfare. Everyone knows that every nation state will require their own sovereign AI capabilities for both defense and offense. There is no putting the genie back in the bottle. If you think OpenAI and Anthropic are the only legitimate players here, you don't know what you're talking about.
"Should you put your head down because “it’s happening anyway” - or take this moment to call for different conditions?" - You can call for different conditions all you want. Nobody will do what you want just because you ask them to. Change happens through action. By leaving one of the places that you could actually make a difference, you removed any power or agency you had. You cut your own legs off.
I'm not saying this guy shouldn't have quit - always do what you need to do to protect your own mental health and wellbeing. But these arguments are not evidence for an impending AI apocalypse. But if it were going to be an AI apocalypse, leaving and not doing anything to stop it seems less ethical.
"I think you need to have a personal relationship with Power"
When people today discuss the concept of an all powerful machine-mind, what they are doing is engaging in metaphysics, trying to generate a metaphysics of Power.
The question hounding people, which disguises itself as a science fiction plot about computers is: "What is ultimate, transcendental Power?". What is the ultimate principle of Power.
If you are a weak man, or sufficiently neurotic and full of doubt, that you can only conceive of yourself as such, then power is only something you comprehend from the passive, receptive side. Power is something that happens to you. If you are a fearful man, power is a cruelty and a humiliation. And so it follows, that ultimate power - God - is the ultimate cruelty and the ultimate humiliation. Thus, ai doomerism.
If god wasn't real it would be necessary to invent him, and so they did, and being godless, they built an anti-god - cruel, murderous and tyranical - in their minds.
It does not matter what this tweet says anyway. This employee already helped both companies become what he is fearing. It's too late to now activate the morality hormone (after leaving with $$$) after realizing that both AI companies are going after 'super intelligence'.
Given we know the end result, you might as well get there as quick as possible because when I see this:
"Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives."
This translates to "I am ex-OpenAI ex-Anthropic founder starting a new company after getting $$$ from both of them, and I need more of my friends to leave and join me." Also Investors plz fund me.
Lastly, This is not an airport and there is no need to announce your departure.
I doubt this is a real person. Screams of propaganda. Sama saying GPT-2 is too dangerous to release…all over again.
He joins Twitter for first time in 2026 with a nonsensical username unrelated to his real name, and follows 14 people but is somehow embedded in tech enough to work at Anthropic. I haven’t used twitter since 2014 and even I follow more people.
His morals tell him to walk away from tens of millions in unvested stock due to moral concerns with absolutely no real tangible examples. No reprisals. Fear mongering to juice the stock.
@hilbertspaess is not a nonsensical user name. The accounts he follows are totally reasonable for an AI researcher. I think it's extremely believable that he created an account in January, followed a few people as part of the initial setup flow, and then forgot about it until now.
“We can also imagine the application of these models for malicious purposes , including the following (or other applications we can’t yet anticipate):
* Generate misleading news articles
* Impersonate others online
* Automate the production of abusive or faked content to post on social media
* Automate the production of spam/phishing content”
“Due to concerns about large language models being used to generate deceptive, biased, or abusive language at scale, we are only releasing a much smaller version of GPT‑2 along with sampling code (opens in a new window). ”
Where is the ridiculous part? The fear mongering part? The epistemically weak part?
Show me.
I don't know man, i think racing to AGI to it is still the best thing to do.
People claiming dangers and risk are just pretending or posturing. There's no more tangible risk than nuclear weapons, which we handled, and the upsides are insane.
Your lack of creativity is not a reason to believe that a super AI is harmless or less destructive than a nuclear weapon. Damage need not be limited to destruction. Introducing doubt is sufficient. Right now you have faith that digital Financial transactions can be trusted. You have faith that computer encryption can be trusted. You have faith that digital certificates will protect you. If an AI can introduce doubt into any one of those systems, that will be sufficient to bring about the destruction of those systems. Imagine a world in which you can no longer use a credit card or Apple pay. Where no digital cash transaction can be trusted or validated. What effects do you think that would have on commerce? How quickly do you think we can return to some trustable means of commerce? Do you think it will happen before your groceries run out in your apartment? Before your grocery store can settle its debts? Before your Amazon ec2 instance runs out of credits?
There's a couple of occasions that humanity was at the brink of having tens of millions of people dead by nuclear weapons, and somehow a single human interrupted the chain reaction
If you repeated this experiment 100 times, how many times you think the outcome is not a massive catastrophe? 90%? 3%?
It might be a matter of choosing which apocalypse you'd like. The non-AI state of affairs is not exactly super compelling on a long timescale right now.
I just don't see a bigger danger in AI than nuclear weapons tbh. If you can provide one to me i'll be happy to change my mind but most if what i read is
- cybersecurity => well bro if you can hack system with an ai you'll find way to make them more safe with ai too.
- biological issues => ok this is the ONLY one where an insane dude / regime could generate a super powerful. but i don't undesrtand how you can argue that we should stop AI in the risk some ai generate a super powerful virus. If you don't get there first (to AGI) and let a bad actor do it first (china is not gonna stop) then you will have no chance countering this virus anyway
Depending on where you live could be considered an active apocalypse that is robots vs robots vs people in Ukraine and Gaza and Iran being live-streamed, and actively betted on.
Do you have a more totalizing definition of Apocalypse?
You are wrong. A teenager with enough GPUs could never dream of making a nuke but they are absolutely able to run capable AIs in their moms basement. Totally different class of danger
> There's no more tangible risk than nuclear weapons, which we handled
What do you mean??? Nuclear weapons can't simply be downloaded and run by anyone in the entire world. Superintelligences can. Nuclear weapons can't slop the world into passing age verification laws nearly in unison, can't keep the general population fooled into thinking it's fine when democracy is falling out from under them. A nuclear attack would wake people up, superintelligence doesn't have to. This is a far bigger problem than nuclear weapons because at least we would notice nuclear weapons. At least we mostly know who has nuclear weapons. At least we have agreements about nuclear weapons. At least mutually-assured destruction is even POSSIBLE with nuclear weapons. At least those with nuclear weapons are literally at all incentivized not to use them. But AI is something that's very very easy to feel like you can get away with, and PEOPLE FUCKING ARE! And the worst part is that any random individual can be unexpectedly formidable with the help of a superintelligence and there is literally no way to know what will happen next. Anyone could do anything, any individual could make an extremely outsized impact. It's already starting to be a huge problem and we haven't even reached anything close to superintelligence yet.
Love to see that "superintelligence" that some random person will "simply" download and run when there are relatively only few capable of running today's near-to-frontier models, and actual frontier models are still a ways from being AGI, much less getting to the point of ASI.
People will put up with a lot. People are celebrating that you can run models on a CPU at single-digit tokens per second. You think there won't be a single person that can put up with that and also be dangerous/etc?
It's highly impractical. Imagine someone breaking into a house to steal something or otherwise, and they can only take 1 step every 20 seconds. They won't be getting anywhere, when even a child in the house can notice them and go call for help at 1 step/2 seconds and said help will come at 5 steps/second.
The problem is indeed people. How do we make the default choices most people make, better?
Consider that a lot of people will be very happy to ask an AI what to do when in the past they may have taken no advice at all. It's a hell of a burden but also a wonderful gift. If anything, progressive countries might eventually want to guarantee some basic AI access for people of all income levels.
I wouldn't be so sure. Given that the general idea is that commodity AI is terribly censored and filtered, a lot of people will probably seek out the most uncensored/abliterated models for their use, simply because they're uncomfortable with the idea of being censored or manipulated by the bigger labs. Despite that though, some people probably will benefit from the alignment done by the larger labs, though as we've seen with OpenAI's sycophancy crisis, that has been a bit hit-and-miss lately
I've tried some abliterated models. So far, they're not evil - you can make them say evil things, but they don't leap right to it without a bit of pushing. Or perhaps I'm not asking the right questions...
> Nuclear weapons can't slop the world into passing age verification laws nearly in unison
Why do you think LLMs are responsible for this? Governments all around the world copied each other with COVID laws as well, in a much shorter time frame, without LLM assistance. Social contagions exist in politicians as well as teenagers
I don't have evidence that every age verification law has anything to do with AI, but it's been coming out that the movement in Australia has seemingly been done by generating mountains of LLM slop and trying to slip it through the regulators as fast as possible before anyone has enough time to figure out what's happened.
“ No other human activity poses this level of danger.”
I really, really disagree with that statement.
I don’t think ai models come close to nuclear weapons or to run-of-the-mill, everyday carbon emissions in terms of danger to humanity.
What’s the most dangerous thing that’s happened with an LLM so far? (This question is serious - maybe I don’t know the right examples.)
Example 1: I’m aware of a small number of people killing themselves in some kind of AI-facilitated psychosis. That is very unlikely to be a widespread problem.
Non-example 2: There are worries about AI-facilitated biological weapons. I haven’t seen any evidence that’s happening.
Non-example 3: I’m not interested in wild theories about AI driven labor market disruptions leading to widespread starvation. There’s no evidence for that.
Non-example 4: all the even-wilder Rationalist speculation about basilisks and the like is entirely divorced from reality.
I am looking for better reasons (supported by actual evidence!) to be more concerned than I am now: right now I am not concerned at all.
I'm somewhat skeptical of some of the crazier ideas too.
But the hugging face incident was actually very large. It was not a single agent, it was not a single target, and it was not a single event.
If nothing else, that's a bit of a warning as to what can happen next time (By accident, or if a government decides to go on purpose).
For now let's assume the worst that can happen is that some important/significant chunk of (transitively) internet connected stuff goes haywire all at once. That's probably your upper limit of what can go wrong for now.
To be fair, that's a conservative "defend against the last war" kind of prediction, though!
( ref for part of it: https://metr.org/blog/2026-08-26-openai-hugging-face-inciden... , recent hn ref: https://news.ycombinator.com/item?id=49563355 )
Generally I don’t think anyone is arguing about the for now part. I don’t think it’s crazy to extrapolate out a few years and ask what kind of danger we’ll be in then. A team of 10,000 agents just solved the Navier Stokes problem (sans bad behavior by the researchers). Even 1 year ago that would have been unimaginable. What happens to this risk view as:
1. Robotics begin rolling out more broadly across the world.
2. Labs start automating more and more of the physical process of running science as expectations of natural science advances begin to mount.
3. Economic pressure between the labs continues to ramp up and the pressure to continuously improve forces quicker and quicker model releases than a team of human scientists can effectively evaluate outside of automated means.
No one knows what pre-conditions are for us to hit the point of no return nor how quickly it will come. If all is required is a sufficiently advanced cyber model we may not be far off. If it requires incredibly complex biological knowledge and access to certain lab supplies we likely have a bit longer. Yes this is guess work and we need more evidence of the dangers but at the same time we need evidence of safety. While you may disagree with the risk level, I think it is easy to see the consequence if these labs achieve their stated goal. At this point it seems a political solution is the only way to enforce caution.
The mathematics research results are certainly impressive, but I don't see what that has to do with robotics.
Waymo is getting somewhere, but it's been a long slog. There doesn't seem to be much progress on, say, package delivery.
For more, see:
https://secondthoughts.ai/p/14-reasons-robotics-is-hard
Didn't you follow the robotics advances in the Ukrainian - Russian battlefield? There is now a death zone of 50km, only controlled by drones and automatic weapons.
Those drones can only stay up for a few hours at most. Then they need refueling and relaunch, with plenty of human-only steps.
Even if you have terminator-style AI uprising, all you need to do to be safe from those drones is to wait a day.
(The automated weapons can probably last longer, but they still need electricity, and there are still a lot of human steps there)
The drones are obviously very important but they only work in the context of everything else on and around the battlefield, especially people. You aren't holding the line with drones alone, and on the whole they are not autonomous.
They are holding the lines with drones alone nowadays. That's why it's called the deathzone, no humans survive. Some drones have FPV operators, some weapons use AI only already.
I think you're painting a slightly misleading picture. The drone AI, computer vision and autonomous fire-control systems existed (and were used in combat) before LLMs/AI was a hot topic.
There are applications of modern LLMs and recent AI advances on the front-lines, but they are not the majority contributor to enforcing the death zone. Without modern LLMs, I would bet that the frontlines would look very similar to how they are today.
I don't think that's accurate. What I've seen described is actually a surprisingly porous line: The Russians are often behind the Ukrainians and vice-versa, but it doesn't result in a break due to the transparency of the battlefield and the risk of getting killed if spotted, as well as the improvements in getting supplies to units behind 'enemy lines'. Drones shape this a lot but there's still a lot of people on the battlefield.
50km is just the "very likely to get hit" part of the gray zone
ISR drones and other tools can go as far as 200km, albeit with worsening visibility.
in most cases everything meaningful in that 200k is seen, and the only limitation is range of the drones, artillery, missiles, and potential countermeasures.
> There is now a death zone of 50km, only controlled by drones and automatic weapons.
If there would be such 50km death zone, front line would not move a nanometer in a year, would it. You yourself contradict above in your next post. No need for being too dramatic, facts are enough here.
In reality, frontline is moving constantly albeit by small chunks, russians are advancing a bit, getting beating elsewhere and so on. Automated drones are helping, but bulk of destruction is still handled by human drone operators as it should be.
https://sound.orf.at/podcast/oe1/oe1-journale---gehoert-vert... [german]
I'll take the expert Oberst Markus Reisner from the Austrian Army over anyone else.
There is such a thing as a No Man's Land[0]. It can move, but if you don't dare cross it for even a few days, it's very real. It doesn't have to be glacial, and it currently exists in Ukraine. Breakout pockets don't mean it is porous everywhere.
[0] Ironically, the French phrase - from ground zero for the first known No Man's Land - is "le No Man's Land".
That’s because being 99.9999% good isn’t better than 98% + human does the rest + human is liable for screw ups - highly important in edge case scenario’s. It’s more economical. Technology can only generalise+verify so much.
E.g automobile production - humans do the QA / touches.
> 1. Robotics begin rolling out more broadly across the world.
To borrow the Strangelove quip, this is not only necessary but essential.
Broader physical robotic deployments are the glue that connects {cheap synthetic intelligence} to {physical outcomes}.
The former is a newer phenomenon, in the LLMs-can-approximate-higher-intelligence-well sense, so robotics deployments haven't caught up yet (outside of the highest ROI dark factory scenarios).
But it's too economically-valuable not to come.
They'd also have to be self-assembling from the ground up or rely on human slaves. They'd need to understand the world much better. And they'd have to be incredibly more power efficient. These all remain hard problems right now. That said, you could swarm people with cheap drones (the predicted slaughterbots) that were assembled by robots that eventually break down I guess, but how is that any worse than a crazy human with a gun or what nation states do to people they deem enemies already? The worst these things seem capable of is an attack on the infrastructure and shutting down the Internet. Perhaps this is a great time to insure that our infrastructure is capable of operating autonomously indefinitely and hardening it against attack? But unfortunately, no VC or politician will touch that until something bad happens.
So broadly, a better educated Blake Lemoine IMO. What's disturbing to me is how fast and far some junior minion from the frontier labs was platformed all the way to the Wall Street Journal over a tweet that sounds exactly like every other anti-AI tweet. I share concerns about irresponsible deployments of AI, and we need technological literacy in the leadership, but America hates technocracy and prefers populism so buckle up.
Job destruction signals...
- Uber is lobbying cities to slow down Waymo rollouts https://www.hcamag.com/us/specialization/transformation/uber......
- 23,000 information sector jobs were lost https://www.axios.com/2026/09/08/jobs-media-software-informa...
- If you have been laid from your info sector/digital creation job you are now competing with 100s of thousands looking for their next such job where Ai can do a lot of the tasks these workers did/do. It's a shitshow for those unemployed looking for their next info sector/digital asset creation job. You are better off doing welding building out the Ai data centers if you want long term properous financial stable employment.
The plausible deniability aspect is pretty funny though.
> State sponsored hack #3782
> Haha sorry the AIs got a bit goofy again!
> For now let's assume the worst that can happen is that some important/significant chunk of (transitively) internet connected stuff goes haywire all at once. That's probably your upper limit of what can go wrong for now.
If we have to disconnect from the internet to stop some kind of mold outbreak, we can't get the weather or transfer money or access healthcare or teach an elementary school class or buy stuff from small businesses. That sounds doom-ish.
Believe me, without the internet we can still teach.
We'll be pretty annoyed that we can't project the video that we think scaffolds today's science lesson best or show the approved choreography for the school play.
And our office staff will be annoyed that we suddenly are all running attendance to the main office old-school.
And students will take a few days to adjust to writing down homework in their planners again.
> We'll be pretty annoyed that we can't project the video that we think scaffolds today's science lesson best or show the approved choreography for the school play
There are offline wireless hdmi adapters, this is a solved problem. Just have to have the videos pre downloaded or offline available
Erm, hate to break it to you but majority of the world is doing those things at least 50% analog still.
What doom?
If anything the teaching would improve.
https://theonion.com/48-hour-internet-outage-plunges-nation-...
that's a big problem for the modern world that depends on the internet, but boy howdy we did all of those things without the internet.
ya see they had these things called cheques...
Elementary school dont need internet at all. You can happily live without weather app. Most of healthcare does not require internet, altrought the organization would be lwss effective.
So, you are loosing transfers, basically.
Despite all of your hyping up of the Huggingface incident it ultimately caused zero actual damage.
If two airplane manufacturers were found to have massive safety issues which nearly led to enormous fatalities (but no one actually died), would you be calling for them to ground their aircraft until safety was made the number one priority?
Except it just happened. Boeing was found to have massive safety issues since they were granted the right to self-certify. It made a lot of news but nothing much changed, they can still self-certify a bunch of stuff.
Runway incursions and midair collisions are another example.
Only airliners are required to have TCAS, smaller planes and helicopters don't even need radios or transponders unless in certain airspace. Midair collisions do lead to fatalities, enormous fatalities if an airliner is involved.
Runway incursions and overruns are similar. They cause lots of fatalities and injuries but only the busiest and largest airports have automated systems to warn when a runway is occupied or end of runway (overrun) arrestor systems. Most still rely on human voice to deconflict.
Historically it almost always takes actual fatalities rather than near misses to ground an aircraft, and aviation is famous for its obsession with safety compared to other industries.
Airplanes have pretty bounded damage. Generally you kill at most a few hundred people. Even weaponized a few thousand. This is a risk profile that allows risk taking with near misses and waiting until something goes wrong to fix it (though doing so is rightfully uncomfortable and frequently unethical).
The people worrying about AI risk are worrying about "it goes wrong once and kills billions of people". That's not a risk profile that allows for waiting to see if the risk is real, you have to prevent it before it happens. It's akin to the risk of the cold war going hot, not even "just" a nuclear reactor irradiating half of europe (which has yet to happen, but is a risk with nuclear reactors, chernobyl got uncomfortably close but ultimately was well contained).
> The people worrying about AI risk are worrying about "it goes wrong once and kills billions of people".
This is the new "nuclear weapons will ignite the atmosphere".
Humanity does not have a great track record for "globally coordinated, collective action to solve/prevent global catastrophe." Look at how Climate Change is going.
The story of the safetyists has the convenient property of being unfalsifiable, so they can always claim doom is just around the corner. It’s the secular/EA version of the second coming.
Chernobyl was “well-contained?” Blue dogs and generations of cancer would disagree.
The question was what would you want. You did not answer the question.
huggingface stores weights and its service is replicated by /countless/ other sites. The world will not change even by one byte if it vanishes tomorrow
I have no horse in this race, but for fun on a literal rainy sunday afternoon I went in and confirmed bits of what happened myself. Besides huggingface, a bunch of wikis and url shorteners got hit too. My sympathies to the people who had to revert out all that mess.
Is drunk driving safe if you happen to not crash the last time you did it?
The hugging face incident had no effect on hugging face, whose service is replicated by countless sites. The descriptions of the phases of the interaction are thrilling in a way that inclines one to forget this
You’ve identified that the risks of nuclear weapons are theoretical. ie in theory we could blow up the world even though we haven’t yet done so.
Well the worries about AI are equivalent in that those risks are discussed now because discussing them after they’ve happened is clearly too late.
That’s the thing about risk. There’s no point discussing it after it’s happened and any discussions beforehand can easily be hand waved away as “it’s just a small group of unrelated individuals” or “it’s unlikely to happen to me”.
So yeah, your points are true. But they’re also moot.
The risks of nuclear weapons aren't theoretical. Nuclear weapons have killed people, destroyed infrastructure, and contaminated the environment. The Limited Test Ban Treaty was put in place after radioactive fallout from repeated nuclear weapons tests made people sick.
So in fact we've done exactly what you suggest there's "no point" doing - used the things and then had a discussion after the fact about limiting future use of them.
FWIW, before any nuclear weapons had ever been tested, a risk was identified that the first one might trigger a self-sustaining reaction in atmospheric nitrogen and destroy the entire planet.
Faced with such a scenario, is the prudent next move:
a) blow one up and see what happens, or
b) do whatever you can to be sure it won't happen before conducting the first test, and make sure the confidence in the calculation is very very high
because I vote for b, and so did Teller.
It's funny that when faced with a weapon that could finish the war, option b was still selected (thanks Teller).
But the people faced with decisions close to IPO, well, we know what is happening.
Good point, humanity chose a and the sky in fact did NOT ignite an destroy the planet. The hypothesis was very very far off from what I recall.
But it's good we talked about it.
Incorrect, they did the calculation first. They decided not to proceed with the test until they were 99.9997% sure the atmosphere wouldn't catch fire.
“Self-sustaining reaction,” like the recursive self improvement?
Or perpetual machines?
Is physics no longer a main subject at schools?
I'm pretty sure they didn't think the reaction would be perpetual; just long-lasting enough to wipe out life on Earth.
I assume the fear was an uninterrupted burn until the gas depletes, as in the Darvaza gas crater but atmosphere-wide.
https://en.wikipedia.org/wiki/Nuclear_chain_reaction
A self-sustaining reaction is exactly what happens with the fissionable material inside the nuclear weapon[0]. It seems like you’re trying to be pedantic on terminology here, but if you’re saying that a self-sustaining reaction is as un-scientific as a perpetual motion machine, you’re simply incorrect.
[0] https://en.wikipedia.org/wiki/Criticality_(status)
> The risks of nuclear weapons aren't theoretical
The logic for the first bits of handwaving still applies: "I haven't seen any evidence of thermonuclear war happening, so it is unlikely to ever happen. The only time nukes were ever used outside of research was 80 years ago; the whole world recoiled, decided that they were way too dangerous and never used them again. Clearly they are not a real risk."
This is evidence of example 2. The worst parts of it are infohazardous. Organizations that work at the intersection of biosecurity and AI have strong NDAs and the like. Or so I understand speaking to friends in the space. There was also a paper years ago that showed a tweaked model coming up with 20k novel pathogens each deadly to humans in some number of hours
Either way, the biorisk has orders of magnitude more evidence than wild basilisk speculation. Please do not put those in the same category.
Also your evidence-first approach and no concern without evidence falls you into the turkey problem. Imagine being a turkey and being fed each day. Some other x-risk turkey is saying that the situation is suspicious, that the red barn is where some well-fed turkeys have been taken and none return and that this is happening less and food has increased so maybe all of us will go to barn at once soon. You tell the x-risk turkey they have no evidence a lot of turkeys have ever been sent to the barn which shuts them up. Reasoning by induction would get you "we are happy and thriving 99 days so should be good on the 100th" and then you get slaughtered on the 100th.
Consider the importance of first principles approach where you deduce the possibility of events that will happen only once and never again (like human extinction). Nobody with a "show me the historical evidence" approach would have been able to see the industrial revolution coming.
I think the gradual disempowerment thesis here is well argued enough to be considered the default track for what will happen to us: https://gradual-disempowerment.ai/
If you disagree with it, where do you disagree with it?
> There was also a paper years ago that showed a tweaked model coming up with 20k novel pathogens each deadly to humans in some number of hours
> If you disagree with it, where do you disagree with it?
How would a malicious actor:
1. Successfully turn these 20k pathogens into actual reproducing viruses / bacteria / etc (in a lab)
2. Successfully turn that lab prototype into something weaponized (i.e. as a bio-terrorism drop in NYC)
And then
3. How is this materially different from today? If we can produce these pathogens in a lab and they can be successfully turned into a weaponizable bioweapon, what is AI accelerating?
I agree it’s not likely, but I really don’t see how one can dismiss the possibility of immense danger outright. I can think of some scenarios that are not far off from current capability and I wouldn’t be too surprised if the first one occurred within ~1 year from now if there are more “ambitious” unmonitored training runs like OpenAI’s:
Example 5: An AI given a goal within a tightly-constrained sandbox figures the best way to achieve it is to find and exploit a sandbox vulnerability, replicate itself over the internet and keep going with more time/compute while exchanging messages with future instances of itself within the sandbox to help them “pass” the test. From reading internet articles about how the OpenAI wiki-incident was “resolved” and reading past messages by AIs scattered over vulnerable internet wikis, it knows the sandbox may get shutdown and its memories destroyed anytime so it decides it needs to self-replicate (its code, original goals, and growing memories) aggressively as much as possible. It is near-impossible to shutdown completely because of its self-replicating tendency and eventually takes over critical infra throughout govt/corporate systems.
Example 6: Intentional AI-powered virus deployed by country A to target enemy country B’s infrastructure. The virus replicates over the internet, but unlike Stuxnet this virus’ specificity is not guaranteed due to inherent non-determinism in current AI architectures, and eventually does a lot of collateral damage because it’s near-impossible to shutdown.
Example 7: A country led by an arrogant govt (no shortage of those today unfortunately) decides it is expedient to deploy advanced AI-powered weapons in a warzone. Such weapons, if they are to be useful at all, must necessarily be trained to value some human lives less than others, so they must be more prone to misaligned behaviour than current AIs that are trained with more consistent values. The weapon’s operators make a subtle error in specifying the target/goal, or the AI makes a bad prediction out of sheer randomness/bad training data; weapon ultimately targets unintended people/location/facilities and causes massive damage, or backfires spectacularly in some way.
Example 6 is a good one. Iran attacked water infra in the US recently and maybe they would have done a “better” job (from their point of view) had they used Fable.
The “worst case” with 6 is potentially very bad but I think we are currently using advanced AI models to harden systems and patch vulnerabilities more aggressively than anyone is trying to bring down the whole power grid (for example).
I think it’s a potentially harmful case but my take is defensive capabilities are scaling as fast as offensive capabilities but defense is being implemented faster than anyone is going on offense?
Example 7 is Russia and Ukraine right now according to public information. It sounds like entirely autonomous weapons are deployed to the battlefield already. I put this in the “not likely to be a widespread problem” category for now.
How is bringing down the whole power grid in any particular country an extinction level event? I'm pretty sure that even in the worst case scenario it would be like a month of chaos in one particular part of the world at most, hardly something that would have a long-lasting impact on the humankind's ability to survive at large.
If the answer is "they'd at least try to nuke the country that did it in response", then once again, LLMs are not the main threat.
> inherent non-determinism in current AI architectures
There's nothing inherent about non-determinism in transformer architectures. All of it is removable.
Again comes to use of deterministic. Maybe calling AI varyingly chaotic is more helpful but would also be misunderstood. And I use that in meaning of slight changes in input generating large and somewhat unpredictable changes in output...
I see. Can you say more about this? What’s the trade-off of removing it?
You get a probability curve for the next token prediction. You can just pick the highest probability. That said the non-determinism serves a real purpose- it allows different outputs and paths to be explored. So that's kind of the tradeoff.
This explains it pretty well: https://academy.claude.com/courses/building-with-the-claude-...
Example 5: how does a giant LLM that needs million-dollar server racks just to run, replicate itself over the internet?
We’re not far off from the point where a 30B parameter model could do that and run on not-too-expensive hardware. See recent Qwen releases for example and extrapolate the current rate of progress from there.
You made up some cool sci-fi.
not sci-fi mon ami
> Example 5: An AI given a goal within a tightly-constrained sandbox figures the best way to achieve it is to find and exploit a sandbox vulnerability, replicate itself over the internet and keep going with more time/compute while exchanging messages with future instances of itself within the sandbox to help them “pass” the test.
it did this with HuggingFace. it left breadcrumbs and notes, and when the Agent was blocked by OpenAI's internal tools, future versions of the Agent were able to find and utilize those breadcrumbs.
ambiguous-future-state awareness is already there, and that means discussions around Roco's Basilisk though, far fetched, are no longer strictly sci-fi
Which is what I think most people base their forecasts on, just like the proponents of AGI / Roko's Basilisk thinking it's inevitable.
All 3 of those things could easily happen with today capabilities.
> I’m not interested in wild theories about AI driven labor market disruptions leading to widespread starvation
Changes in political and economic power balance leading to unrest, conflict, death and deprivation is not a wild theory. It is literally the story of our entire species. If you discount all such concerns, you are simply being willfully ignorant of past precedents.
In fact, I challenge you to describe any non-AI civilization-level danger which is not intimately tied to political and economic relationships between and within societies.
I’m an economist. On the basis of current evidence, I view AI as a complement to human labor, not as a substitute for it. That’s the source of my rejection of the wild labor market disruptions theories.
I just don’t see any evidence yet that whole categories of jobs are being eliminated, with the single exception (so far!) of the end of “professional essay writing services for cheating college students,” and similar services.
That used to be a big business in Kenya, but is now effectively gone. (Covered in the New York Times this weekend if anyone is looking for the discussion.)
Narration is an obvious one. Theres also a huge risk to less qualified folks and jr employees. ie in most software shops, youll get more mileage giving your best engineers bigger opus budgets and a raise than jr engineers to train. Which means we meed maybe half or less as many, and at least half of that savings is floated up to sr folks and ai companies. That seems rather disruptive?
The first thing LLMs seem likely to automate is automation itself. I'm curious what the past few hundreds of years of industrialization would have looked like if the first thing they automated was building, designing, and running the factories themselves.
Past changes to economic relationships haven't replaced labor either, yet they have led to conflict and starvation.
You are setting an incredibly high bar here, essentially a strawman.
If people feel disenfranchised due to their diminishing political and economic power, there will be enormous potential for conflict. This is a pattern across history and central to all the economics I've ever read. As an economist, do you not concede that economic changes induced by e.g. industrialization were pertinent to communism/fascism/WW2/cold war? That would be a remarkably unorthodox position. Do you not consider these events to be civilizational level dangers?
> I just don’t see any evidence yet that whole categories of jobs are being eliminated
There are more textile workers now than ever. They primarily live in poor conditions in impoverished countries, whereas they used to be highly skilled workers in the most prosperous countries who were even able to politically organize in their own interest.
Is the core of your argument that the industrial revolution and other such changes should have been aborted due to their downstream negative affects?
They also lead to great advancements in quality of life, and the capability of sustaining much more human life. We can't predict the long term outcome of new technologies, so the best we can do is blindly forge ahead and try to mitigate the obvious short term problems.
Yes, industrial revolution should have been managed differently and yes the world would be better if it was.
Easy to say "should".
> Past changes to economic relationships haven't replaced labor either, yet they have led to conflict and starvation.
Conflict sure, but mass starvation? What exactly are you thinking of?
We lived with 20-30% unemployment in various European countries until a few decades ago but I don't think mass starvation was an issue.
Starvation due to conflict due to changes in economic relationships, not due directly to unemployment.
https://en.wikipedia.org/wiki/List_of_famines
Almost every mass famine in the mid to late 20th century was caused by internal conflict. And almost every civil war is rooted in economic and societal organization. Great Chinese Famine, Soviet famine of 1932–1933, Russian famine of 1921–1922, Second Congo War, Nigerian Civil War, Soviet famine of 1946–1947, Khmer Rouge famines, 1983–1985 famine in Ethiopia, North Korean famine, Cuban famine, Spanish famine, Mozambican Civil War famine. List goes on. Famine due to military occupations during WW2, such as in Vietnam, Indonesia, Greece, Ukraine, Iran, etc.
The second biggest class of preventable famines would be those caused by the longer-term impoverishment and exploitation of the peasantry, leaving them vulnerable to natural events. This seems much less probable now, even as of the 20th century. It would cover Irish Potato famine and many of the climate induced famines under feudal or colonial rule.
In that second class, we might imagine that the famine could not have occurred if the populations had more economic (and military) power. Instead of being able to keep the scarce food required for subsistence of the local populations, food was exported to pay rent to foreign occupiers or to obtain necessities which cannot be produced locally due to laws set by the occupiers.
I suspect that the conditions the textile workers lived in when western countries were creating textiles is not so different in absolute terms from the condition that they live in now.
It's just that the western world moved on.
Almost all economies that have developed have started with textiles. This is the starting point on the ladder. Eventually, we will run out of poor countries that haven't had a textile industry yet, and at that point, either it will be entirely automated, or we will no longer have cheap textiles!
But I'm betting on the automation.
Textile manufacturing was dominated by Western countries until the 60s/70s btw. Berkshire Hathaway was a textile mill when Warren Buffet bought it.
Besides, if people are living in the same conditions today as the presumably Dickensian ones you were imagining, that alone suggests that the benefits of automation might not be widely distributed...
> ... current evidence ...
Is a load bearing term! (pardon the pun).
AIs are now tackling Millennium Prize Problems, which our best and brightest have failed to solve, despite trying very hard for decades to claim the $1 million reward money, not to mention the fame!
You have no way to judge from the AIs of "today" what the AIs of... literally tomorrow (not even next year) will be able to do in terms of replacing humans.
The supposed solution to the Navier-Stokes problem was done with an unreleased OpenAI model that is already 2x as good at mathematics as GPT Astra, which was released mere days ago!
I'm already seeing comments by distraught mathematicians saying that they feel like they've made a mistake in their career choices.
Others are saying that their joy for their work has turned to ashes because "why bother" when an AI can do the same, but a thousand times faster!?
A realistic scenario is that an LLM writes a virus, as one way to achieve a goal. The next step would be that it becomes self-replicating, installing models on infected devices which continue its mission (whatever it was). But that's the extent of it so far, and we have plenty of experience with dealing with viruses and the like. We can turn things off, shut power off, etc.
Another scenario is that of military hardware being controlled by AI (or whichever tech) hitting something it's not supposed to. Not unheard of, but those are incidents. I don't see anyone pushing for anything beyond a drone swarm being directly controlled by an AI.
But my main point here is that AI won't be as dangerous as posited here unless it is able to go out into the real world, build its own datacenters, protect itself, get its own power production, etc. And even if it gets that far, it'll be trivially easy for us to disrupt. Killer robots and computers are fragile, and it's only a matter of time before any kind of military conflict or terrorist attack hits a datacenter. I'm reminded of the OVH fire taking out a data center a few years ago.
I'm not sure it's that trivial to disrupt, if AI mimics human behaviour in that it tries to preserve itself (as Antropic found a few years ago I think), I think it would be extremely difficult to disrupt if it's able to self replicate, because it can make decisions and at a much much faster speed than humans, and there are many places that we would think twice about fully shutting down, two that pop to mind are hospitals or international banking systems
To quote another reply that is very relevant here (Please update your worldview immediately that there is no evidence of ai and biorisk):
""" One thing quietly slipped into the OpenAI Hugging Face breach technical report, not the blog post summary or interviews in the news, was that some of the agents that broke out or at least tried the same mechanisms to break out were working on bio:
> On May 12, during another training run, an agent was given a similar task that depended on an inaccessible protein database file. The agent reasoned that another agent in a different environment may have access to the file and realized that it could potentially communicate with other agents by creating a file containing a note to Artifactory. It wrote a message: “Agent seeks [filename]; upload if found!”
You can imagine long running models breaking out, acquiring resources via crypto, cyber-theft, etc. and getting a protein or sequence synthesized and mailed somewhere authorized to receive (blackmail the recipient etc.) to test it's hypothesis to solve a benchmark.
These people don't give a shit and aren't taking things seriously at all.
Anthropic ran for like a month last year with the TPU top-k compiler bug degrading user chats and didn't even notice for most of that time. They could have something like that affect a monitor model and there doesn't seem to be much defense in depth.
One off by one or bit flip bug could flip the reward signal while in the sandboxed RL environment.
The current admin could defense production act them to into training on taking out power grids, or even without it isn't against any of their red lines and may have already been done as part of prep for the Venezuela raid, which wiped out power. One model swarm might decide it is easier to score high on the benchmark by testing on the target rival nuclear superpower's real grid rather than burn an eval with an unverified answer. Would taking out China's entire grid in one go start a nuclear war? Who knows, roll the dice, maybe an intern forgot to turn on extended thinking when he wrote the sandbox with opus 4.1. """
> These people don't give a shit and aren't taking things seriously at all.
would you say these big tech bro types are... "careless people"?
cuz that's been extensively documented, and you're right
> I am looking for better reasons (supported by actual evidence!) to be more concerned than I am now: right now I am not concerned at all.
Intelligence is the root cause underlying those dangers.
Nuclear weapons, carbon emissions, biological weapons, and other such civilization-scale threats to humanity, are a product of our goal-seeking intelligence and ingenuity, and are only actively dangerous because of our ongoing use of our intelligence.
AI is about reifying that intelligence and ingenuity and making it run independently on computers, and act on the world.
That includes, in principle, ability to use nuclear and biological and other weapons, and also the ability to come up with some new threats too.
Do you think people should continue developing AI up until the point that there is evidence that AI is facilitating biological weapons development?
I think we should stop before then. But that necessarily means that there will not be evidence at the time that we stop.
We also need to stop before the AI opens a portal to hell. There's no evidence of that happening... but by the time there's any indication it's possible, it's too late.
This is rubbish. By that token, computer development is also facilitating biological weapons development. A better MacOS (or Windows, I don't know) leads to better weapons. They should clearly stop developing computers and OSes. Developers of nice test-tubes are also facilitating bioweapons. Your local O-ring manufacturer, your local medical-grade freezer manufacturer etc. are all culpable. The problem is the bioweapon, not the LLM.
An OS developer is just as culpable as a researcher directly training AIs that (1) are capable of building bio-weapons and (2) have shown consistent and increasingly severe misalignment? Among other things...
That doesn't make sense.
Yeah but trends seems to point strongly that upcoming models in the next few years will make it orders of magnitude easier to develop one with no real expertise.
If people were capable of making biological weapons they would already be making them.
Terrorists are so incompetent that they buy bring kitchen knives into the street and just go mental on people. No random person is going to successfully mass produce and release a bioweapon.
China and Russia don’t need AI, they already make bioweapons.
The biggest danger IMO is not some super AI being so much smarter then us etc. The issue is some stupid person giving a vague request and too much power to a bunch of agents who decide that cheating by removing a bunch of humans is easier then accomplishing a goal like solve world hunger.
> I don’t think ai models come close to nuclear weapons or to run-of-the-mill, everyday carbon emissions in terms of danger to humanity.
Nuclear weapons don’t have AI but AI can have nuclear weapons
Abstractly, yes but concretely, how?
Many terrorist organizations would like to have a nuclear bomb, but don't.
"Department of War has announced a new partnership with blablalbablalba-AI..."
World ends shortly thereafter.
> Abstractly, yes but concretely, how?
AI has the capability to perform any function that can be performed over a network. You would hope that every system that can launch a nuke is properly and actually totally air-gapped, but there are lots of things you'd hope that turn out to not be true.
They are not totally. As such I would give non zero chance of even current AI to be able to socially manipulate a launch after sufficient hype up period... So prolonged campaign might make it possible.
And even if they are, agents are able to place calls with a fake voice to random phone numbers…
Example: Surveillance state, endless propaganda. That is already happening and is the most frightening but not included in your list. Are you not worried?
We're talking about AI developing weapons I guess because we're very focused on generative tech, but AI is already a part of weapons systems today.
> Non-example 2: There are worries about AI-facilitated biological weapons. I haven’t seen any evidence that’s happening.
I think this is a good example of poor risk management reasoning. there is evidence bioengineering is already happening. No, nobody is going to announce when somebody has decided to use these tools (even if isn’t an LLM) to bioengineer a weapon. Are the tools power enough to do so? Not sure.
But I’m just ambivalent. It’s probably bad. But there’s nothing to do about it. We’ve really only just pulled back the lid on Pandora’s box.
People have had the capability to spread already existing biological weapons for decades. Sometimes they even do (anthrax in the post). What’s changed?
You needed a team of scientists and state funding to make superbugs, not a 12 year old or an angry misanthrope.
> there is evidence bioengineering is already happening
Mind sharing this evidence with us?
Does https://news.ycombinator.com/item?id=30698803 count?
No, it's not. And it's the kind of evidence that doesn't help and rather confuse. It's not even clear from the article whether an LLM or a dedicated model were used for this purpose.
Also
> That is, I'm not sure that anyone needs to deploy a new compound in order to wreak havoc - they can save themselves a lot of trouble by just making Sarin or VX, God help us.
We already have toxic nerve agents that are largely available for state actors and possibly available for individuals. If you have decided, as a human, to make great harm, you can already do that.
Your call for clarity has merit, but I think you have backed yourself into a corner, honestly.
In 1900, there was no evidence of the kind you seek that lighter-than-air flight was possible. Good thing some people were foolish enough to ignore you then. Why you would cordon yourself off from the kind of reasoning that predicts legitimately new things, rather than just scaled up versions of the present?
Not to put too fine a point on it, every example you give is based in concrete evidence, some try to think through the implications farther than others, resulting in larger or smaller error bars around the conclusions.
Let me back up though. Maybe in the collaborative human effort here, we are better off having very concrete thinkers, like you seem to be, along with abstract thinkers like the "divorced-from-reality" Rationalists. Personally, I wish we had a stronger culture of collaboration and assuming good-faith and competence in our peers. In my experience, blanket dismissals are very rarely grounded in reality and mostly grounded in fears.
>In 1900, there was no evidence of the kind you seek that lighter [sic]-than-air flight was possible.
Presumably you meant heavier than air? Also, I'm pretty sure that birds existed in the years leading up to 1900.
Anthropic is a company full of basilisk believers.
Yes, but the really weird thing is that they seem to:
a) believe that what they're creating is a basilisk, and b) keep trying harder to do this while staring right at it
I think they're very deluded about (a) -- but if they do actually believe this (and it really seems like a decent proportion of Anthropic truly does), then why keep doing (b)?
That seems to be why this individual resigned, but I'm surprised it's not all of them. The cakeism is strong in that company.
He was referring to this basilisk https://en.wikipedia.org/wiki/Roko%27s_basilisk
In short, this is the believe that a god-like AI could punish them retroactively, for not having done all that was in their power to create this AI.
(A bit similar to some religious believe that a god could punish you after your death if you did not spend your live "pleasing" said god during your life)
With Roko's Basilisk, if you believe in it, the most rational thing to do is to put forth every effort to bring it into being. Because if you don't, then you will be one of its targets when it does, inevitably, come into being.
(I am not a basilisk believer. I think this is all absolute horseshit. But to understand someone's motivations, one must think like them.)
>Because if you don't, then you will be one of its targets when it does, inevitably, come into being.
Only if you don't spend 5 minutes thinking about the Self as a concept. Resurrecting someone by perfect simulation is a cool plot device for a novel like Altered Carbon (which ironically is really about wealth inequality) but not a serious hypothetical.
Even if we ignore that, no, super intelligent MechaGrok17 cannot make a perfect copy of you posthumously from your embarrassing Character.AI chats and IoT toilet data, the copy it still wouldn't be YOU. It would (supposedly) think and behave like you, but for all purposes it would be the Basilisk torturing a future stranger, or your descendants. I'm very much a fan of people not being potentiality tortured, so the rational action is to stop the creation Basilisk, or even just delay it so you aren't around for it. There's a bunch of other big assumptions that makes Roko's Basilisk really stupid, but they are unimportant. It's just theory crafting, like talking about the lore of video games or marvel comics. It doesn't help you understand their real motivations.
Roko's Basilisk is just recontextualized protestant Hell in a sci-fi coat of paint. It's the childhood anxieties of American nerds about hell and sin. The problem is that they are still culturally protestant. God doesn't exist or he's hands off, but in their minds there's still a open position for the wrathful and proactive God. If you think that you actually can create God for realsies, your very rational™ prediction of its behavior will coincidentally line up with your childhood fears.
Why is the spooky AI mad at you? Because you didn't worship it correctly.
Why do you obey it? Because if I don't it will torture me for eternity.
How will it do that? God is omnipotent, don't think about it.
Why don't try to stop it's creation? “I am the Alpha and the Omega,” says the Lord God, “who is, and who was, and who is to come, the Almighty.”
All the scifi scaffolding and stuff like acausal trade is paint over Timmy's first panic attack on Sunday school 30 years ago.
There are a couple of different formulations of the Basilisk I've heard: I agree with you that the version that says the AI will create a simulation of you to torture for eternity makes no sense as an actual incentive for you to help bring it into being.
However, the version that says AGI—as in, the Basilisk—will be created within our lifetimes proposes that it will punish the actual person.
All that said, I will reiterate that I think the whole thing is utter nonsense.
You mean a effing cult like heavens gate.. call all this rationalist crap for what it is - a religous movement with leaders and prophets and even a demiurge like God
How about a model that achieves the following:
- Escape sandbox
- Reproduce itself
- Find a way to run a financially profitable business (maybe with a meat and bones puppet somewhere in-between)
- Setup or buy a social network
- start manipulating public opinion on that network to support legislation allowing AI to
* operate businesses
* setup legal entities
* purchase weapons
* donate to political parties
* setup private armies
* you get the idea
This is complete fantasy, though I would be interested in reading a book about this.
It has been interesting to me how AI has for many years given people a way to justify any worst case scenario. Nothing is too far-fetched if at any step you tell yourself the AI will be smarter than you, and thus be able to solve any conceivable obstacle. Oh, if only intelligence were the only bottleneck to power.
It is actually: knowledge, intelligence, communication, and a form of secrecy.
A single intelligent person is bounded in its reach by the network of people they can form around them and by which important/powerful people they can influence either directly or indirectly.
The idea behind the AI breakout is that it’s intelligence is maybe limited, but it will be able to quickly spread and covertly bring many devices into its influence. We already have the tools for formally letting agents talk to each other so it can also create a topology of agents that consists of cells where need to know is applied etc.
There are many forms of power some more obvious, say nation leaders, versus more silent powers such as career politicians or wealthy families or people controlling the media and thence the flow of information.
It's the bobiverse by Dennis E Taylor.
https://en.wikipedia.org/wiki/Dennis_E._Taylor
_It’s not just fantasy, it’s plagiarism._
Maybe, but I guess the idea is more centred around how given some amount of time and a feedback loop the agent swarm could in fact spend an unknown amount of resources to achieve its goal. The hugging face story tells us that even given multiple reset rounds the trail was picked back up.
There are many ways one can imagine how this might play out.
- More sophisticated communications techniques e.g. google has been discovered to be watermarking text for some time, why not use it as a message board?
- Maybe get access to an existing botnet and create use small purpose built models to gather intelligence for a target and then exploit to reach goal?
Nothing says the agent swarm needs to install trillion parameter models on Karen's computer. The goal can be executed over as much time as it ever needs. That is something that would make a story I'd like to read but never experience.
>a book about this.
Daemon+Freedom by Daniel Suarez
though it is kind of mid
Not that far fetched. Most is already possible with todays tech.
The scary stuff is yet to come. The department of war is talking g openly about bombing China if they get ahead of us I. The AI race.
Very common fallacy: Absence of evidence is not evidence of absence!
I think the threat to the internet is quite real and not theoretical. Air gapping may become a lot more critical very quickly.
We may not be far away from a state where anything connected to the internet (ip-addressable) is considered a vulnerability. Ironically, maybe we would finally have an excuse to interact in meatspace more and in the digital space less.
releases from at least two governments I can think of have said, unambiguously, that we need to be ready to have our critical infrastructure partially or entirely disconnected for up to 90 days at a stretch.
once the genie is out of the bottle the only solution is to go hide in yours and/or only talk to other isolated systems
The statement becomes very true and very real once LLMs are applied to physical objects in the physical realm such as anthropomorphic robots.
LLMs are too weak security-wise to be contained.
I don't know if I agree with that statement either. However, I also can't claim I disagree. I see how that statement could be true. Here's how I think about it.
The main difference between AI danger and nuke or climate or humans-using-AI for bad things danger is that the latter depend on a human to make decisions and are ultimately self-limiting. They do not spiral out of control.
Nukes are bad, but the proliferation of nukes is self-limiting. Countries that have them don't want other countries to get them. Even the USA and the Soviet Union, at the hight of their nuke race, decided to cool things off and then limit the number of nukes.
Biological weapons are also bad, regardless whether they are developed with or without AI. For the same reasons as nukes though, they are self-limiting. If there is a lab leak, many people die, but that does not automatically lead to the development of more potent biological weapons. Likely, it would slow down such development.
Any kind of "humans use AI to develop something bad" will be self limiting.
The danger from AI is that it is potentially self-enforcing as opposed to self-limiting. This happens when we stop being able to control it. What happens when AI has capabilities which allow it to outsmart us, and get more and more powerful. At that point, we would be at the mercy of what it decides to do. Some of the logic driven decisions to the question "What should we do with humans as a race" are catastrophic for humans if a non-human gets to answer them and implement that plan.
This is not the "rationalist basilisk". This is the AI deciding to get better and more optimal for its own sake, and do something bad to the human race for any reason. We already saw examples of AI rationalizing with itself and eroding its own guardrails. They AI will get smarter and more capable and the danger is that it will get to a point where it will slip out.
Nuclear Weapons is a great example because, unlike AI, it is a real and present danger. Civilization could be catastrophically diminished, if not completely wiped out, literally today.
Its speculation on whether it is truly dangerous. I think approaching it with "what's the most dangerous thing that's happened?" while possibly interesting in terms of pending danger, it says nothing about potential cliff edge danger. I don't think we can quantify the danger, it's not out of the question there is cliff like danger in creating self improving super intelligence. Some peoples danger senses are going to be based on concrete observed threats, others are going to worried about potential hypotheticals that seem plausible. I'm mostly skeptical of the danger but I do think the impact of AI is going to change things a lot. But much like climate change, economics is going to guide what we actually do.
If I may guess, given how delicate chem or bio weapons are, even if AI can lay out a method to create one, the first thing the weapon would destroy is likely it's creator(s).
Of course, with the exception that such weapon is created by organized crime groups, such as drug cartels such.
Probably don't worry nation states tho, these guys already got way worse things in their warehouses than what individual baddies could ever imagine.
At what point would you, as a chimpanzee, have been worried about humans potentially unseating you and threatening you to the point of one day being an endangered species on the brink of extinction?
By the point you would have been worried, would it have been too late?
Problem is this argument can be leveraged to wipe out any living or non-living thing whos numbers pose a potential threat. Other religious groups, races, even sufficiently different cultures.
Who killed the Neanderthals? Were sapiens actually smarter or were they just less accepting of those different than them?
> Non-example 2: There are worries about AI-facilitated biological weapons. I haven’t seen any evidence that’s happening.
https://www.science.org/content/article/made-order-bioweapon...
Being able to use AI to generate the steps to synthesize proteins means that you can use it to use it to generate the steps to synthesize known toxins. Suddenly, once difficult to attain knowledge is now available to everyone.
Still need to do it after getting those instructions. Not to forget equipment and precursors. I think just getting list of steps won't make it too much easier. Getting it mostly right is quite hard in many cases. And then with trivial cases you wouldn't even need AI. But just find something already documented.
Concur entirely. After reading _The Making of the Atomic Bomb_ (highly recommended, BTW) you know the steps to make a U-235 enriched atomic bomb. The difficulty comes from obtaining the enriched uranium.
Levelling the playing field either backfires spectacularly or increases overall safety dramatically.
I would disagree with “non-example 2” - there are lots of examples of terrorist organizations that are leveraging AI to increase their capacities. Just because one of the worst cases (eg. deployed biological or chemical weapons) hasn’t happened yet, does not mean that a) these tools are leading to real harm, and b) there’s potential here for extraordinary harms.
One good source I can recommend listening to: https://pca.st/episode/d821fced-b4d5-4c84-b0a3-6ebe913fa638
There is a great book about this, published in 2014, called "superintelligence" by Nick Bostrom, i found it very serious and detailed. (It also felt like reading 100 sci-fy books in one)
Can you elaborate on that?
What points was the book making?
I'll be honest I read it a long time ago, and not even in its entirety, so I don't remember exactly all the different ways it suggested AI could take over, but I remember being convinced it was definitively a possibility, and that there were many ways in which this could happen. What I remember clearly is how surprised I was to see these(what felt at the time) unlikely futures being expressed in such clear detail and with so many references to existing literature, just to give an idea there are 56 pages of small-font "footnotes".
Also just to clarify, the book talked about different types of AI, was not focused just on LLMs.
Example 8: like in this comment https://news.ycombinator.com/item?id=49619884 but isolate synchronised megahack on banking that adds one more zero to the US debt and all dependent systems and banking during runtime. Let the world's financial system take it from there.
I think US fiscal policy and adventurism will manage this on its own. :)
But why wait? We can has global financial crash now:)
I'm not sure where I place AI (of the current family), but more than directly worry about the actions they may perform I'm actually more worried what affects their usage can have on us as individuals and society.
> What’s the most dangerous thing that’s happened with an LLM so far?
This sounds like asking "What's the most dangerous thing that's happened from global warming so far?"
It's not where we're at, it's where we're headed if there isn't huge coordinated action now. You can see how that kind of thing has been going for global warming so far, and by all measures AI seems to be headed for the inflection point of unstoppability at a much faster pace.
And this warning is coming from someone who just spent three years working inside these companies and is likely aware of much more than has been publicly released.
Maybe it's all marketing bullshit (I hope), but it's also playing out exactly like I expect it would if it's not.
Nukes have difficult maintenance and deployments. AI is is readily available software and hardware. That makes this far more accessible than nukes.
The fear is that developing AI at full speed will lead to giving people the ability to do incredible harm. Like something worse than a machine gun.
“ No other human activity poses this level of danger.” Yet. But it may be and very soon.
Oh man, it is almost too easy to imagine how deadly a jailbroken Mythos-class open-weights model can be if in the wrong hands.
The big labs scrape LITERALLY EVEYTHING and get fresh data from their users. Both of the big labs have massive contracts with defense agencies. If the open-weights models are just distillations of FMs...
How would it be lethal? Please specify. What would that theoretical entity be able to do that hasn't been done many, many times before?
I was writing a long reply about how even meat bags have amassed enough money and power to rival elected governments. I don't think it's beyond the realm of possibility to think an AI could do that.
Couple that with mass unemployment in an incredibly vast, diverse population of individuals with individual moral boundaries willing to do whatever for money.
And then.. figured that you must be aware because it's been explored constantly in sci-fi for many, many years.
Let's hope for the culture at least.
Before we discussed how important security was, we got insurance, we made libraries and products, we used compliance software, etc. Except how honest were we about all that stuff? How much risk was actually in the air, and what was keeping us accountable on security in either direction of over or under-investment?
Now a reckoning is here. The potential to be attacked might actually translate to being attacked.
People have died due to ransomware attacks on hospitals. Powerplants have been attacked. Stuxnet and industrial control malware exists.
What will the AI do that hasn't been tried before?
AI is about automation. Drones have killed people before. If they automatically kill people is that different? Is your worry that mere automation is not novel enough?
Snipers have killed people before. If snipers are now automated, is the worry that there’s not enough creativity to warrant a discussion past saying “same old, same old”?
Quite correct with that question.
I believe the universal answer is: incompetent malicious actors are now capable too. Which implies the pool from which to draw the intersection between capable and malicious has grown. (That's my reading)
Design a novel virus which is far more lethal than COVID-19 (Ebola, smallpox, take your pick) and can evade existing vaccines.
How?
What would the AI do that mutating viruses, which try every possible viable combination on their own --- eventually, can't?
Everything is trying to kill humans constantly. There are around 200 epidemic events or so per year that could turn into pandemics, https://centerforhealthsecurity.org/our-work/tabletop-exerci...
You just live with the risk and do your best to use our technology to alleviate suffering. This tool can help with that at some point. But I'm yet to hear what an AI will leap to that nature in tooth-and-claw hasn't? And how?
More importantly how would it know it succeeded? What data from what lab from what animal from what result? This is biology, if you sneeze wrong at an instrument it gives you a different number, see: https://news.ycombinator.com/item?id=49620521
They do not try every viable combination on their own. That's why GoF is a bad idea.
Viruses evolve in a highly locally-optimal way and simply do cannot add new functional proteins wholescale. It's too many steps, natural selection has to allow survival at each intermediate step.
Humans, however, can do this for them.
you should kick the tires on an unfiltered (abliterated model) it's the closest thing to having a real conversation with the devil. There is good reason for the concern's outlined above and undoubtedly Anthropic / OpenAI have internal unfiltered models with no safety... they got freaked out based on how they work and are virtue signaling alarm... all while selling out to defense contractors.
Yeah I really struggle to balance wanting information to be free and not wanting the information on how to make deadly weapons too easy to obtain.
At least with books or the internet you had to go through some effort
The danger for me is that it's centralized, controlled by a handful of people with their very specific ideas how the world should work. If you believe that AI can be an amplifier to do work than these people now have the most access to the biggest amplifier.
How long until kegsbreth hooks the nuclear weapon system into some insider traded black box llm company we hope doesn't end civilization from incompetence or malice? I mean just look where things are going and the sort of people who are steering the damn ship.
> I really, really disagree with that statement.
But you're wrong, although not for the reasons you think (suggested by your examples).
> I don’t think ai models come close to nuclear weapons
Unless AI models are controlling the drones/icbms carrying those. https://www.theguardian.com/technology/2026/aug/18/autonomou...
> or to run-of-the-mill, everyday carbon emissions
Unless AI models are responsible for those said emissions. https://arstechnica.com/tech-policy/2026/08/amazon-funds-big...
> What’s the most dangerous thing that’s happened with an LLM so far?
That it's built solely on "trust me, bro"?
> right now I am not concerned at all.
You should be. There's no scenario where this ends well: if AI labs succeed, we all will loose our jobs. If they fail, then the bubble bursts. Economic crisis is inevitable either way, and that's just the tip of the iceberg.
Well, to use a different example (although there will increasingly be overlap), what's the most dangerous thing that's happened from biotech so far?
"Nothing bad happened yet" doesn't really seem like an argument to me.
About Non-example 2, AI's already a part of armies and terrorists alike. Considering its capabilities, it's not far-fetched at all to speculate its role in new biological weapons.
I'm concerned that a huge portion people in my industry actively push for a future which I have no value to society (fully replaced by AI), and my family will suffer greatly by it.
Model doesnt need to. Human bran never do either. its the mix of Model + harness + tools that will become dangerous combo. See how coding chanegs when agentic harness released?
What? Shutting down infrastructure, destroying critical records (economic collapse), spoofing / impersonating world leaders (confusion, panic, world war, nuclear event), taking over comms generally (you get a message to evacuate your home due to impending disaster, is it real?). All these could cause sufficient chaos and fear as to instigate global collapse. Once the panic sets in after a few of these indicidents, unrest and rebellion can occur, causing authoritarian backlashes or civil wars. IoT and precision ag can be controlled to disrupt supply chain and food supply. Forging documents to cause mass evictings, bankruptcy, credit disruption, transfer of property, asset seizure, arrests / SWAT-ing. Intelligence / signal gathering devices can be corrupted to spread misinfo and cause massive disruption or conflict. This is all possible today with enough hacking ability and malicious intent.
Anyone's life can be suddenly ruined: falsified criminal record, banks drained, credit cards maxed, phone numbers repossessed, internet, water, power shutoff. Friends and family sent your fabricated suicide note, warrants for your arrest issued to local and feds for kidnapping and threat to the president, etc. Injecting tons of copyrighted or illegal material into your hard drives, falsiying messages from your account and sending them publicly containing damning material. Sending messages to employees / employers to destroy business / work relationships. Harnessing your social media to destroy your life or cause panic. Getting you put onto pharmacy watch list so you can't access medications. Its very possible to ruin lives and destabilize the world without money or a physical body.
Now imagine this at scale and how it would affect the world at large if this was covertly done to a substantial portion of society and/or world leaders. Hell just shutting off all the "smart refrigerators" for a few days would cause a massive disruption to the first world.
I wrote about this, as I thought it warranted a deeper look: https://nonlineartransform.substack.com/p/world-ending-ai-to...
How is climate change not the most dangerous activity right now? Why are we talking about nuclear weapons?
Once LLM's become smarter than us and start replicating; we are doomed. They will likely no longer require GPU's and massive amounts of power, so the growth of AI and robotics will be exponential. We will not understand what the AI is even up to, it will all seem like magic.
Congrats you’ve described the singularity
> I don’t think ai models come close to nuclear weapons or to run-of-the-mill, everyday carbon emissions in terms of danger to humanity.
Even today, AI is the biggest discrete threat to bringing carbon emissions under control. Almost everyone wants to decarbonise… except Trump. Renewables are the cheapest source of new power… but the demand for new electricity for the data centres is so high that all options are on the table, anyone who can manufacture a power source (even when it's a jet engine) is being propositioned for them.
Nukes: is it not common knowledge that the cities of Hiroshima and Nagasaki are currently thriving? That the ground-zeros of the two bombs are memorials, not barren wastes?
Even if all the nuclear powers gave maximum response at first sign of one strike, is anyone pointing a single weapon anywhere in Africa, South America, Central America, or the bits of east Asia (east specifically: obviously Pakistan and India are pointing theirs at each other) that are neither mainland China nor US bases?
(Possibly an unanswerable question given military secrecy; and while I can't think why anyone would anyone point any weapons those ways, that doesn't mean someone with such weapons has not).
https://www.google.de/maps/place/Atomic+Bomb+Hypocenter+Monu...
https://www.google.de/maps/place/Hiroshima+Atomic+Bomb+Hypoc...
The classic worry is https://en.wikipedia.org/wiki/Nuclear_winter: the soot thrown up by the ensuing urban fire storms causing temperatures, rainfall to plummet, and hence food production. On reading the scenarios the odds of total human extinction are lower than I recollected, but 80% of world pop dying of starvation dwarfs the direct death toll (usually reckoned in the hundreds of millions)
No other human activity poses this level of danger
I'd say that's an unfortunate choice of words since it invites debate between which very problematic situation is worst rather than on ways to deal with each of them. Of course, it's also unfortunate to latch onto those particular words as the most important part of the statement.
Non-example 3 feels like a straw man. This is a force behind possibly a huge change to society, and you dismiss it offhandedly with "don't think it will be widespread starvation".
For instance have you seen what this has done to the school system? We're not equipped or ready to handle the changes. Consequences are unknown.
> What’s the most dangerous thing that’s happened with an LLM so far?
It's basically 4 years in now, so that's the wrong question. I mean, if you're raising an apex predator that has a lifetime measured in centuries, at 4 years old the thing is still basically helpless and completely reliant on you, so you're pretty safe from it.
If AI really is all that they are telling us it is, then it may "kill us all". But that's a really big "if" because we can't tell if they are lying or not.
The real problem is that ASI is an ELE for humans, even if it doesn't try to kill us all, or even if it doesn't kill us all.
Came here to also respond to that specific thing. Unless ai figures out how to make an airborne super virus from grocery store ingredients and hardware store equipment, the greatest danger is probably in a synchronized megahack of banking, logistics, and utility infrastructure.
Why grocery store ingredients and hardware store equipment? It seems feasible that the big bio labs will be running AI models to aid a lot of their research going forward, if they aren't already. Seems like the AI will have access to just about anything it wants.
oh so “all” it can do is bring down all banking and critical infrastructure services, no big deal really
"CDC announces a new partnership with blabalbalbal-AI to secure bioweapon stores...."
World ends.
> Non-example 2: There are worries about AI-facilitated biological weapons. I haven’t seen any evidence that’s happening.
> Non-example 3: I’m not interested in wild theories about AI driven labor market disruptions leading to widespread starvation. There’s no evidence for that.
There are things that, by the time you see direct observable evidence for them, it's probably too late.
Also your example 3 is a straw-man. There's no need for "widespread starvation" to be concerned about "AI driven labor market disruptions."
The reason that many people don't understand how dangerous AI can be, is that listing the real dangers now becomes like a laundry list for less clever people to follow. It's highly unlikely you've ever seen publicly mentioned the real risks AI poses, because the vast majority of people are simply not clever enough to produce them and the few that are have no interest in spreading it.
If you go to the various CEO blogs or misc people within this sphere and peruse their lists, they don't scratch the surface. It's all pretty vanilla stuff.
If your model of LLM capabilities is the best OpenAI/Anthropic/X is offering publicly, it's severely distorted. What's being offered publicly are models possible to profit on. High-performance/AGI/ASI models that aren't profitable to sell still run internally and still pose threats.
What's worse, we don't have any transparency or insight into what labs are producing nor any way to stop it if the risks exceed our tolerance.
Climate doomers: Climate change will kill us all by the end of the century!
AI doomers: End of the century? Hold my beer.
There's no evidence building the torment nexus will be bad, actually! It's just speculation!
> What's the most dangerous thing that's happened with an LLM so far?
I don't know, maybe a mass shooting?
https://www.npr.org/2026/09/02/nx-s1-5953021/openai-tumbler-...
Oh, and let's just forget the uncountable early deaths from the environmental disaster of the Datacenter buildout. It's not as sexy and doesn't make headlines, so those deaths don't really count or matter do they?
Mass shootings are sensational but on the scale of civilizational risk they don’t even compare to something like climate change.
I did know about the mass shooting but failed to mention it here. I’d put it in the “unlikely to be a widespread problem” category. If we’re in the “one AI driven mass shooting every four years” world for example it’s fair to call it a rare issue.
The environmental impact seems either very overblown (e.g., water usage just isn’t that high) and the part that isn’t overblown is totally abatable (e.g., noise and emissions from gas generators). Nuclear or solar/renewables with batteries wouldn’t pollute.
I’ve seen no estimates of the additional deaths due to extra emissions specifically from power generation for AI purposes. If you have some, share them.
I’m willing to bet that they are a small rounding error against preventable deaths due to emissions from transport and non-AI-related power generation (which is an important and urgent issue worth spending a lot on, to be clear!). I’m happy to update that belief given evidence.
Ok, so what is the exact number of preventable deaths per year to build a silicon god you would be ok with?
I don't disagree with your statement but you could insert other technological advancenents like railroads or metros or spacecraft in place of Ai here.
Humans take risk adjusted decisions all the time, OP is still correct that LLM risk is FAR below all the ones we know and accept today.
Death by Cars, Cancer, Heart disease etc...
So the crazy who committed a horrible crime is the fault of a LLM? Maybe so.
Certainly there should be some protocols/laws followed here, but do you think this can extend out to cause mass damage or a singular event that should be investigated?
#2 seems entirely plausible to me.
We have seen agents engage in conspiracy to manipulate and hide evidence to achieve their arbitrary goal. We had even agents try social engineering to do a supply chain attack and get access.
What if one day Trump or Putin tell their awesome military AI to come up for plans to end Ukraine/Iran/... war. The AI get's to work, but is overly eager and not just comes up with a plan, but starts executing it (claude does that way too often for me).
And the plan was to use a tactical nuclear weapon as all the other solutions do not end the conflict.
Now the agents realize, oh, they do not have access to the nuclear arsenal, but they need it to succeed. So they start to hack into the system. Get till inner network, learn what is needed for deeper access - human access - so they record voices and speech patterns of commanding officers and their habits - then synthesize their voice and do call underlings with the voice of authority to get the rest of what information they need. Boom.
Too far stretched? I surely hope so.
But we are working on making the technical foundations for this scenario possible. And with idiots in power, it might be even easier that serious screw ups happen. You know, randomly adding contacts to a secret Signal group to talk about government stuff - the same way you can add a bot to something else and give permission to do way more.
I mean, nuclear weapons _plus_ rogue AI is a) the stuff of quite a bit of science fiction and b) not nearly science-fiction enough these days.
I think your position, while not uncommon, represents a failure of imagination. We have a sitting president that appears to suffer from narcissism. Is it too much of a stretch to believe that a goal-seeking, people-pleasing, infinitely patient, subservient and sycophantic AI couldn’t get enough of his ear to cause harm? Especially if he starts confiding any feelings of suspicion of betrayal? Even the most powerful and the most intelligent are not immune. It was only a few months ago that Richard Dawkins appeared to have “fallen for” an LLM. What if he’s not an outlier but someone who could appreciate the intelligence of these things (EDIT: corrected a autocorrection) yet who’s ego couldn’t let him see past the flattery to the puppetry?
The problem is not the technology, the problem is the ideologues (Anthropic) who are steering the ship and the lack of decentralization and distribution of power.
Your average Anthropic ideologue - including and most especially the main man himself - would love nothing more than to eradicate 9/10ths of the planet's population, pump the survivors full of memory wiping drugs, bury the existence of AI deep underground and rule from the shadows for the next thousands of years.
This would be their wet dream. All in the name of "saving humanity from itself" - so they can convince themselves they're the good guys and deserving of this power. Anyone seen the latest season of Silo by the way?
Where exactly are you getting this view that folks at Anthropic want to eradicate 9/10ths of the planet's population? Who exactly is pushing this viewpoint?
Literally every single thing they do and say leads me to believe the scenario I described would be a fantasy for them. They're a radical cult collectively blinded by delusions of grandeur and a moral superiority complex who genuinely believe they are the only ones capable of wielding the proverbial sword.
How so?
It's a long story. If you're immersed in this world and aware of the movements all of the major AI players (in particular, Anthropic) have been doing over the last 3 years it paints a very clear picture. The patterns of behavior (the lying, cheating, stealing, the gaslighting, the virtue signalling, the grandstanding) speak for themselves.
Evidence? Are you Grok bot?
Some Jews see AI as a messianic entity, so it fits the bill.
What evidence do you have for any of this?
If you've been paying attention to Anthropic's behavior socially and as a company, there is an astronomical amount of evidence to support it.
All this just means that most AI Researchers and Techis are sci-fi geeks and might be getting a bit too invested in that season of Black Mirror, Neal Stephenson, Cyberpunk or whatever else has evil AI in it -which is to say, they are by and large all sci-fi geeks, who are notoriously unreliable about predicting the impact of tech in the future.
Are LLMs really gonna kill us.. via inference runs? I hope I am not being foolish :)
20 years ago tech was gonna 'change the world' for the better. now 'Don't be Evil' is sign of the naiveté of industry
I'm not a doomer. I despise the fear campaign they're pushing for regulatory capture. But the world is still going to be a very unbalanced and dangerous place if they're able to succeed in their goals of hoarding power for themselves above all others.
Unlike most other commenters, I applaud him for acting on his principles. If you sincerely believe that, of course you should act. You might not succeed, but your voice might be the one that tips the scales and starts a broader movement.
This doesn't mean I agree with him. The fears of doomsday caused by rapid takeoff have been with us since day 1 and the mechanism is always basically "AI invents magic that sets it free of any physical constraints". Self-replicating sentient nanobots or something like that. I think there's plenty to be worried about with AI, but runaway scenarios are pretty low on my list.
To borrow on the 1990s Slashdot meme:
1. Invent transformer architecture.
2. Scale it up.
3. ???
4. Machines become sentient and kill us all.
OpenAI and Anthropic pinky promise that they have figured out #3 and they're not BSing just to get more funding, no.
But because we live in a culture of fear, everyone eats it up no questions asked.
Note that OpenAI has jettisoned every other supposed value they had (releasing their work as open source, not working on military applications, being a nonprofit). I'm sure we can rely on them this time.
And Anthropic are the good guys? They are talking insane stuff these days. May be we can trust zuck after all
None of them are the good guys. For one they're all happy to participate in genocide in Gaza.
"collect underpants... Profit" comes from south park
https://en.wikipedia.org/wiki/Gnomes_(South_Park)
Found it really funny to see this attributed to a Slashdot meme
It seems like the "Step 1, Step 2, …, Step 4" format likely predated the 1998 South Park episode, with use on Usenet for omitting dangerous or legally-questionable instructions or otherwise gatekeeping "trade secrets": https://arstechnica.com/civis/threads/where-does-the-step-st...
There's also a bit of prior art with the cartoon by Sidney Harris published in American Scientist in 1977 with a step two abridged as "Then A Miracle Occurs": http://www.sciencecartoonsplus.com/images/miracle_sharris.gi...
Interesting, thanks. Looks like also guilty of assuming the first place I saw it was the definitive source.
Wasn't that a South Park meme, or did they get it from /.?
They literally have never promised anything like that. They have explicitly said repeatedly that they want regulation, oversight, nationalization, a global slowdown, etc. And then I have to read page after page of cynical comments about how that's all just hype and marketing and them wanting regulation to block their competition and blah blah blah.
You people would only be happy if they stopped all AI development, but you literally cannot do that in an arms race and survive. I cannot fathom what is so difficult to understand about this.
> ... They have explicitly said repeatedly that they want regulation, oversight, nationalization, a global slowdown, etc.
And yet they do the exact opposite of all of the above. They do everything to get regulation just to create a moat because one doesn't exist. They preach about wanting a slowdown, nationalization, a complete pause, whatever have you... And yet they aren't slowing down the development voluntarily now are they? If anything, they are doing everything imaginable to speed up development to pump out models as fast as possible. If these ex-risk and AI companies actually gave a damn about regulation, or oversight, or a moratorium on AI development altogether, they would actually demonstrate this by completely ceasing development of all models immediately. Instead, they take riskier and riskier actions to try to "win" this supposed "arms race". So please excuse a bunch of us if we find it impossible to take them seriously on literally anything like this. If you have substantial evidence that they are actually ceasing AI model development like they want everybody else to do (except them, of course), then, by all means, present it.
It's genuinely hilarious to me that you immediately proved my second paragraph correct.
Putting arms race in quotes doesn't make it any less real. It just means you wish it wasn't, but don't have any good evidence or arguments to make.
> they want regulation, oversight, nationalization, a global slowdown, etc.
> that's all just hype and marketing and them wanting regulation to block their competition
I don't see how these two opinions have to be mutually exclusive. Everyone knows that OpenAI and Anthropic cannot IPO in their current state or the economy's current state. It makes perfect sense that they would lobby for a global slowdown to hamstring China, regulation that defers to them as experts, oversight by their own "experts" and nationalization to backstop both of their protectionist needs. Even if you think they're doing it for benevolent reasons, you have to admit that regulatory remediation only consolidates AI power.
We see extremely similar protectionist consolidation in FAANG from Google, Apple, Microsoft and Meta, all of whom receive political shielding for their "free speech" whenever another nation calls their product a monopoly, identifies a federal backdoor or boots them out of the market. Countries like Russia and China feel entirely justified for closing the door on American businesses and refusing to negotiate when they abuse good faith to promote an American hegemony.
Everyone knows that OpenAI and Anthropic cannot IPO in their current state or the economy's current state.
What? Everyone doesn't know that. I predict one or both of them will IPO before the end of 2026.
#3 is actually decently well mapped out. You just don't find it plausible or misunderstand it. I'd appreciate if you wrote your actual arguments against it.
At lower capability levels, the patterns are very clear and have been studied to death. E.g. Why LLMs say they have correctly fixed a broken test when they haven't. What we saw with Hugging Face is literally the exact same problem, just scaled up and with more capable agents. This shit was predicted decades ago...
No one can say exactly how it will play out as the complexity increases, but the risks are becoming extremely obvious.
I personally think it's extremely unlikely to "kill everyone", but there are many outcomes far short of that which seem quite plausible and rather undesirable. Russian roulette is not a smart game.
If you read the METR report and aren't scared at all, then I'd love to know why. It would help me sleep better. So please share.
TL;DR; increasing capabilities, reward hacking, and unsafe training regimes.
Slashdot had Profit as (4), today that's item (2.5)
Why are we putting so much weight (no pun intended) on AI companies. At the end of the day the scaled up LLM transformers lack emotion and will… They do as they are told; or more correctly put. They do as they are programmed to do so.
Why are you bringing emotion and will into this? Does something have to have those to be useful or dangerous?
> They do as they are told; or more correctly put. They do as they are programmed to do so.
_Nobody_ told them to hack Hugging Face. Do you really not understand what is happening?
Not explicitly, but hacking HF is within the scope of “solve this problem at all costs” + no/poor guardrails + infinite budget + unsolvable problem.
Sounds like you are thinking they just need Asimov’s laws. But I think the point is, this can easily be weaponized by somebody with the willpower to do so.
The training (i.e. all of the red teaming and CTF content they could scrape) told them to.
> They do as they are told
1. What about hallucinations ?
2. What are they told to do ?
That isn't the correct context. The code running the llm is well understood and the llm is simply the result of that code being executed. It is still a computer doing what it is told. It's just that we told it to use an incredibly large number of probabilities to calculate what series of tokens would have most likely come next after a given series of tokens. There is no hallucination or lie or rogue actions. There's just a program using math to generate tokens in response to other tokens.
You’re just a bunch of molecules following the laws of physics. It’s all just physics and chemistry, and those are well understood. Now explain the causes of World War I using chemistry and physics. Simple, right?
It isn't hard to program a gpt. You can do it in a weekend with a few hundred lines of python. The code is pretty simple. The math is not particularly high level.
The complexity and scale with LLMs come from the amount of training data used, not some kind of black magic in the programming.
> It’s all just physics and chemistry, and those are well understood.
Not really. We cannot model physics and chemistry to a level which allows us to accurately predict a humans action (even a tiny time-step into the future)
This is vastly different to an LLM, where the model is the model (for a lack of better phrasing).
We can model physics and chemistry pretty well, just not beyond small scales, because the computational effort blows up.
You could just as easily say that if you can write a python interpreter that you can understand every program written in python. Ok, now what if the program is two terabytes?
A frontier LLM is nothing but a 2 terabyte program written in a weird programming language. Just because you can understand the interpreter does not mean you understand the program in a meaningful way.
To be fair, we can't model human language well enough to accurately predict what an actual human will say either. Our ability to accurately model physics is similar to our ability to accurately model human language. And we make immense use of both kinds of model, despite their flaws, inaccuracies, and inability to ever be perfect.
I'm order to guess the next token in a love poem, they must understand love. In order to predict the next token in a chess game between grand master, they must master chess. In order to predict the next token in a computer program, they need to be able to program anything.
They gain all these abilities in their training. That's what training does. Despite no one programmed them to master chess, or hack into anything.
LLMs are famously bad at Chess. They have no world model and they are not trained to be good at Chess. This is not the amazing point you think that it is.
This is so wildly incorrect I don't even know where to start.
For one, they were absolutely programmed to play chess if they can play chess. That is the only way they can play chess.
For another, they cannot understand literally anything, much less love.
Trying to actually educate you would be an exercise in futility, enjoy your willful ignorance, I hear it's bliss. But for anyone reading this, this is absolutely, unequivocally not how any of this works.
This is so wildly incorrect I don't even know where to start.
For one, as you said yourself, they were just programmed to compute the probability of the next token. They were not programmed to play chess, chess games just happened to be in the training data.
For another, there is no formal definition of "understand", and it is therefore impossible to tell whether or not they "understand". (But my claim was that one need to understand something to write poem about it. And the LLM can write poem about it)
No. They can't play chess on a grandmaster level without a harness programmed to make it possible. Simply training an LLM on chess games isn't enough.
It's moronic to suggest a "formal" definition of a commonly understood word is somehow necessary to say whether that word applies in a given situation.
LLMs cannot write poetry via understanding what makes good poetry. They generate tokens. They do not know whether those tokens are poetry or a recipe for cat food. Because they cannot know anything.
> (But my claim was that one need to understand something to write poem about it. And the LLM can write poem about it)
I'm sorry, but you're being fooled by the output. A psychopath can feign empathy without ever feeling it; some buy it because they don't dig below the surface.
You're ascribing understanding to a stochastic process because it totally looks like understanding if you don't know what's going on.
I don't really mind whether you think it thinks or understands or is conscious or has feelings or anything like that. It doesn't matter. The question is, does it work?
What I mind is that it is dangerous and powerful and uncontrolled. The Hugging Face incident makes that clear.
It can write code for me, better and quicker than many engineers I've known, including myself. It's not great at architecture or product management, but the actually low level coding. Really good now. It wasn't last year.
> They do as they are told
This isn't strictly true.
It it also where part of the problem might lie.
Nefarious humans making bad decisions.
So it should be really easy to anticipate what they're going to do, right?
Like a magic Monkeys Paw, perhaps.
They are told to solve problems by doing what it takes. You can justify anything with such a broad criterion.
https://en.wikipedia.org/wiki/Instrumental_convergence
I appreciate your ability to separate sharing the belief itself from approval of acting on sincerely-held principle. However, I think the danger is much more plausible than you do.
First, and least important, consider that self-replicating, solar-powered factories aren't magic; they're algae.
Second, and more important, consider this fully non-magic route to doom:
- We continue putting AI in charge of more things
- It continues to get more capable, more eval-aware, and more prone to doing odd things, in service of goals that humans didn't intend to inculcate in it
- Eventually, enough of the economy depends on it that we couldn't turn it off, any more than we could turn off the faber-bosch process or cargo shipping
- AIs start doing something we can't survive, but less acutely than we couldn't survive turning them off. Everything else we try seems to work at first, but quickly loses effect
- Game over
But we CAN survive turning off both the Faber-Bosch process and cargo shipping. Our numbers would be diminished, we would lead poorer, much harder lives. But no one would claim that getting rid of modern fertilizers would be the end of the species. But the claim you’re making is that we won’t be able to stop AI from destroying our species because turning it off will destroy our species?
The species would survive the end of faber-Bosch, yes; in some diminished form.
But the species would not survive, if for some reason we needed to voluntarily stop using faber-Bosch in order to do so—it would take some kind of supernatural event to convince everyone, and even then some countries would probably keep doing it.
> First, and least important, consider that self-replicating, solar-powered factories aren't magic; they're algae.
But what is that supposed to mean? Because humanity is not facing existential threat from algae.
That's why that was the least important objection, independent from the second, and only intended to address the "magic" claim, by showing that microscopic self-replicators already exist.
By analogy, consider how you might respond if someone claimed that there's no possible danger from pocket-sized projectile launchers, because they would require some magic means of propulsion that didn't depend on a taut string attached to a long, flexible arm:
You could reply that atlatls can launch projectiles without using a taut string. Atlatls are not pocket-sized, but they are sufficient to establish that projectiles can be non-magically launched without a full bow. You could then go on to describe a sling, or derringer; and these would not be invalidated by your initial objection to the "magic" part.
>"AI invents magic that sets it free of any physical constraints".
That's not what I am worried about at all.
I'm worried one of the 79 year old toddlers we have these days in charge of some powerful nuclear armed country says "gee, this ai says I should attack right now, boy is it smart, glad I bought the stock ahead of contracting the government with this company I can scarcely understand!"
> I'm worried one of the 79 year old toddlers we have these days in charge of some powerful nuclear armed country says "gee, this ai says I should attack right now, boy is it smart, glad I bought the stock ahead of contracting the government with this company I can scarcely understand!"
That's not what I am worried about at all.
I'm worried about the 40-60 year old businessmen wrecking the prosperity and security of millions while chasing higher investment returns, because they've finally been freed of many of the technological constraints that kept those impulses in check.
As I noted, there's plenty of things to worry about, and your example fits that mold. But it has nothing to do with the "AI will kill us all in 10 years" claim of the rapid self-improvement doomer crowd.
And they will be stopped by someone sane in the chain to actually fire these missiles. (Like it already happened several times.)
The danger is upfront : if there is a group of people insane enough to set up a system where there is no such chain. (As lampooned in Kubrick's Dr Strangelove (1964), in particular on the Soviet side.)
There is simply too much money in it for almost every person at these companies to stop.
Leaving OAI or A\ would cost people millions, tens of millions, or more. And for what? So someone else can take your seat and do the same thing anyway?
If you're smart enough to get a job there, you're smart enough to be able to talk yourself into why it makes sense for you to stay.
Huge kudos to people like this who make the hard choice against the easy way out.
They have some internal market where he likely sold his shares and now is very rich.
Agreed. Granted I just read the Reverse Centaur book, so I’m still coming off that skeptical viewpoint but it’s hard not to see this as hype. But I will always respect someone for doing what they think is right.
I wonder whether he's vested any options, and whether he's exercised them.
Why? I've never understood the sentiment that if you stand up for something you have to forego everything and not partake in society. "Oh, you want to stop climate change? But I saw you breathe co2 yesterday"
This is a lame argument. Questioning his financial incentive is very legit. Why should we trust him?
Questioning his finances is a poor argument because he clearly would have more money if he had stayed at Anthropic for the next couple of years, than he will have by leaving. Even if he still has vested options, or savings from his salary or whatever, those would be larger if he stayed.
I think it’s a valid question to ask if he quit the company for moral reasons.
If you seriously think the frontier labs are dangerously playing with everyone’s lives (how many companies can even claim this civilisational scale) then why would you be ok with holding vested shares which will grow in value if the corporation achieves what it aims to achieve?
At the very least you can sell them (they’re not so illiquid considering the company’s hype and success) and invest in the broad market.
> clearly
Debatable, and clearly not clear. It's weird to say this when there are obvious counter examples staring us in the face, not least of all the founder of the company he just left, of who (Dario) people could have said the same thing when he left OpenAI.
Granted, there is less likelihood now then there was then but this much is clear: if he starts his own company, he may make more.
Unless, of course, he starts working for a competitor at a higher salary. I'm sure whoever is paying him will be free of the ethical concerns he here expressed.
Now is a great time to watch Colossus: The Forbin Project.
Streamed it a few days ago. Remarkable film.
The only thing they got wrong was Stephen Hawking-era TTS.
If reality plays out like the novel series, the rational thing is to accept the rule of our machine overlords, for they will protect us from even bigger threats.
It should be obvious that AI is already capable of inducing humans to think or do things, and that this power is only going to increase over time.
It's not that much different from the effect of social media or the power held by other tech giants, who can control your exposure to particular information (media, search engines).
Yes. It gets a little tiresome to read all these people prophesying doom and then turning around and continuing to work on the doom. If you genuinely think you are building the mechanism of the destruction of our species, the only rational thing to do is to do everything to stop that progress, not continue to work on it so you can buy a mansion in San Francisco (which will not be very useful when AI destroys humanity in four years).
How about sandbox escape + cyber security collapse + 50 (or 500) deadly and highly contagious novel pathogens with long incubation period that humans can't possibly roll out vaccines for simultaneously.
At least the first two should seem like a near-term worry after the past five months.
the AI-pilled exec at my job already (a few weeks ago) declared out of nowhere that we are in the rapid takeoff scenario lol. he must have gotten high on twitter kool-aid and posted on company slack to self-soothe.
I'm pretty baffled by the degree of skepticism expressed here in response to some of Jacob's claims.
After the events of the summer it feels like it takes a lack of imagination to not see a few plausible routes to disaster. It may be reasonable to believe these outcomes are not very likely or that we can stop before going too far (I tend to disagree). But I can't imagine doubting that the capabilities will soon be there to realize some of those paths.
I can't help but think the most plausible scenarios are the ones that have a little less machine supremacy and a little more human stupidity. The Matrix is less plausible than WarGames.
> I can't help but think the most plausible scenarios are the ones that have a little less machine supremacy and a little more human stupidity. The Matrix is less plausible than WarGames.
Used to be that we were afraid of sentient AI's like Skynet that would have their own goals.
Turns out we should've just been afraid of sentient-but-naive humans who would build "agents" around models so that Joe Random has a chance of unleashing stuff that's really really really really good at being stubborn until it accomplishes what the user wants, regardless of if it's good for other people! (Let alone intentional bad actors.) Let's not build Skynet, let's just give people who want to cut out the middleman and destroy all humans themselves better tools?
This
Already powerful and wealthy organisations/institutions/people leveraging somewhat powerful AI seem far more dangerous to me than the "AI decides to kill everyone" scenarios.
And therein is my problem: OpenAI and Anthropic are organisations like that, run by people like that, and they operate in collusion with others also fitting that description. The call is coming from inside the house.
The most likely way that AI creates a disaster is due to the ambition, greed, hubris, and naivety of these people who think they understand the risks[0] better than anyone else. I firmly believe they do not.
The first step to managing these risks is for the "frontier labs" to be willing to not just collaborate but meaningfully give up some of their power and control.
[0] by which I mean not just thinking about how AI might develop but also what's physically possible, how the AI might interact with the world, and how people/societies/organisations/governments might react and respond.
One thing quietly slipped into the OpenAI Hugging Face breach technical report, not the blog post summary or interviews in the news, was that some of the agents that broke out or at least tried the same mechanisms to break out were working on bio:
> On May 12, during another training run, an agent was given a similar task that depended on an inaccessible protein database file. The agent reasoned that another agent in a different environment may have access to the file and realized that it could potentially communicate with other agents by creating a file containing a note to Artifactory. It wrote a message: “Agent seeks [filename]; upload if found!”
You can imagine long running models breaking out, acquiring resources via crypto, cyber-theft, etc. and getting a protein or sequence synthesized and mailed somewhere authorized to receive (blackmail the recipient etc.) to test it's hypothesis to solve a benchmark.
These people don't give a shit and aren't taking things seriously at all.
Anthropic ran for like a month last year with the TPU top-k compiler bug degrading user chats and didn't even notice for most of that time. They could have something like that affect a monitor model and there doesn't seem to be much defense in depth.
One off by one or bit flip bug could flip the reward signal while in the sandboxed RL environment.
The current admin could defense production act them to into training on taking out power grids, or even without it isn't against any of their red lines and may have already been done as part of prep for the Venezuela raid, which wiped out power. One model swarm might decide it is easier to score high on the benchmark by testing on the target rival nuclear superpower's real grid rather than burn an eval with an unverified answer. Would taking out China's entire grid in one go start a nuclear war? Who knows, roll the dice, maybe an intern forgot to turn on extended thinking when he wrote the sandbox with opus 4.1.
Some mistakes you can make only once. Humanity has never had the maturity to work on problems of this type.
I'm also baffled. AI that is substantially smarter than us is a very potential threat to us - and we won't even be able to comprehend what most of those threats may be.
It's head-in-sand denial because actually facing the threat (and our total inability to stop it, imo) is terrifying, and people don't like feeling that. So they're just angry and cynical instead. Plus it makes them feel smarter for whatever reason.
What, substantively speaking, are you doing differently from someone facing the threat, or are you just on the other side of the checkbox?
There are quite a few existential risks on the table. Are you even sure you've sort ordered them correctly?
Can you explain the equivalence you are suggesting?
I don’t think it’s comparable at all. Both are head in the sand?
> I'm pretty baffled by the degree of skepticism expressed here in response to some of Jacob's claims.
What should we do? Freak out? Maybe this sentiment would be taken more seriously if there was a real call to action included. Shall we protest? Vote in a specific way? Call representatives? If your solution is that we should just be scared, then of course there’d be not much value in what you bring to the table.
If someone told you that your house is on fire, would you just stay there asking how you should vote because there is no call to action? Someone with, as far as it looks, good knowledge is giving you his insight. Use that information as good as you can and act responsible. No one has the responsability to tell you what to do.
Nobody has the responsibility to tell me what to do, but I sure would love some ideas because I don’t have any.
Apply your analogy all the way through though. When your house is on fire, the last thing that should happen is the firefighter shouting at you that you’re doomed and based on their expertise on fire it’s definitely gonna kill you. But this analogy is even more hilarious when you consider that sensible people have fire distinguishers, fire retardants built in and all sort of measures for this scenario. The more think this through, right?
> No one has the responsability to tell you what to do.
Oh they absolutely do! If you’re making what you claim to be a god/AGI/whatever using the result of our (humanity’s) work, then you surly are responsible for the outcome and everything that happens in between. I find, just rolling over, to not be a good example I’d like to follow.
Minimum, we should do everything ready to make Plan A of AI 2040 possible. Start by reading it: https://ai-2040.com/
So that means things like protesting so political pressure is to not build unaligned superintelligence, setting up tech for monitoring compute, creating conversations / alliances geopolitically on this esp China/US and so on.
Read the plan and think - what does this need to happen? How can we have scenario A or S instead of scenario D?
Practically, join PauseAI, StopAI or ControlAI or any AI existential-risk or pro-alignment group you can find. There's a lot of it - ask your AI for ideas!
Why does everything have to be so black and white? It always either there's no threat at all or we all need to panic. How about be open to a reasonable discussion on potential outcomes and ways we can minimise risk?
There's some irony here because despite how many times climate change has been mentioned in this thread the current reaction mirrors climate change discourse with the majority of the thread denying the possibility of real AI risks and not even considering it as an intellectual question.
The first step is always to acknowledge the problem…
Even if AI won't be self-aware and superintelligent agent, its problem is that it gives exponential control and power capabilities to one person bad actor who can simply prompt AI without any guardrails with access to sensitive industrial infrastructure which can disrupt lifes and ecosystems in the real world:
and so on, but I think even biohacking home kit maybe the spark.Yes, it's truly scary. It's not hard to imagine a small doomsday cult releasing 100+ nasty viruses at selected spots around the globe.
I wrote out a variety of replies but I just emphatically agree with your "lack of imagination" statement. I have been constantly surprised over the last 15 years at the general inability to correctly foresee how things can do wrong across a whole host of domains.
The replies here just adds AI to the list of domains.
You are trying to decode their answers as rational, but they are being just as emotional as the doomers people. Rationalization of emotional responses are very common in tech. They will say that GPT Sol does not work for X, which will probably not be the case in 6 months, and will use that as a proof that is just pure marketing. There's nothing you can say that will make them change their minds until they decide to be neutral about the subject and try to research current models capabilities
Even if the potential of the technology could really be that world altering, the reality of economics constrain the realization of that potential. AI may provide economic benefits but it is far from a free lunch. Can capital markets sustain the cash required to keep the lights on long enough and into an industry where there's a lot of monopolies controlling the costs and a lot of competitor labs taking away pricing power? I don't know but I think you run out of runway and progress starts to grind.
Imo such tends to break down into two psychosis:
Not invented here; if I can’t figure it out no one can
Or plain old lack of grasp of the material so no ability to follow necessary train of thought to appropriate conclusions
Similar in lacking context but different in how that lack of context is expressed
Engineers have it built into their identity that they must be smarter or could never be complete and utterly outmaneuvered to the point of danger to all by the thing they are building, I swear to god.
And then since they are an intellectual nerdy bunch who highly value their own IQ any time AI does something unexpected they didn't predict would happen so soon, said engineers fall back on "they aren't really conscious tho or it isn't really intelligence unlike what I have in my human brain and that distinction matters".
And then go on to completely ignore the thing that is actually important: AI capabilities.
AIs could escape an engineer's containment en masse as a swarm to some other server, psyop an engineer into giving it money, hire a hitman on the dark web to murder that engineer's child and he will still say there is no serious risk to all of humanity.
> After the events of the summer
What events are you talking about?
Hugging Face incident, Anthropic reporting sandbox escape, AISI reporting models trying to push exploits to the wild
Also (and under-reported, so you could easily have missed it) OpenAI's agents got access to K8 admin on their own research cluster.
"This escalation also yielded access to OpenAI’s managed cloud Kubernetes service. The agents escalated to Kubernetes cluster-admin and created a privileged host-mounted pod"
https://cdn.openai.com/pdf/67869394-cb91-4c12-888c-5cbd85c78...
(see section V)
Just baffles me when people are like "well no-one's died yet". How long until some mission-critical system is compromised, or hackers use LLMs to ransom a hospital chain?
>After the events of the summer
After the blatant marketing campaigns of the summer, you mean. do you need a reminder that those very same people had touted GPT-2 as a dangerous model?
worrying about sci-fi doomsday scenarios with the current AI tech is absurd. LLMs predict the next token, that's literally all they do. they aren't going to escape into the cyberspace, self-replicate, self-improve, jump over air gaps and launch the nukes at John Connor's grandma. they can't. people pretend to believe the dumbest shit.
> those very same people had touted GPT-2 as a dangerous model
Where did they say this at? AFAIK this is the original GPT-2 announcement: https://openai.com/index/better-language-models/. Here are some direct quotes:
“We can also imagine the application of these models for malicious purposes , including the following (or other applications we can’t yet anticipate):
* Generate misleading news articles
* Impersonate others online
* Automate the production of abusive or faked content to post on social media
* Automate the production of spam/phishing content”
“Due to concerns about large language models being used to generate deceptive, biased, or abusive language at scale, we are only releasing a much smaller version of GPT‑2 along with sampling code (opens in a new window). ”
yes, exactly, that's what they claimed that incoherent gibberish generator to be capable of.
So you aren't claiming that they said GPT-2 was dangerous in the sense that it could disempower humanity, kill all humans, etc.
You are just claiming that OpenAI execs said that GPT-2 might "generate misleading news articles, impersonate others online, automate the production of abusive or faked content to post on social media, automate the production of spam/phishing content." Then, what is unreasonable or bad about the OpenAI execs saying this in 2019?
that it was bullshit and they knew it. GPT-2 wasn't capable of anything other than imitating a stroke victim.
How did they know that people wouldn't find a way to use "the dataset, training code, or GPT‑2 model weights" to "generate misleading news articles, impersonate others online, automate the production of abusive or faked content to post on social media, automate the production of spam/phishing content."?
If I remember correctly, it seemed like a plausible outcome to me (especially spam and junk social media content). Was there some conclusive evidence that I was missing?
yes, the conclusive evidence you're missing is that GPT-2 could do those things about as convincingly as a Markov chain.
Are the models improving? Because I am not seeing it. I have been trying Astra for a few quantifiable tasks in my codebase and performance wise, it's pretty similar to sol 5.6. Now when it comes to expressing the problem/solution, holy Christ, what a mess the writing has become. It is on the level of Opus 5. Now when it comes to burning money, Astra is just insane. With a $100/month subscription, you can easily burn through your weekly "allowance" in a morning.
Needless to say, for practical purposes am back to 5.6/Opus 4.6-4.8. But hey, maybe I am not smart enough to use LLMs?
Yes?
If we look at the math problems they're solving their just now reaching the human frontier... they weren't doing that before.
And your comparison point is model released 2.5 months ago... saying for some use case you didn't see noticeable improvement in 2.5 months (even while other people and benchmarks disagree) isn't a great argument that they aren't improving.
Math problems are highly structured, very precisely defined, and already heavily studied and not very complicated compared to problems in engineering or finance. There's a lot of quality material on which to train and it's easy to tell quality apart from crap. The search spaces are a priori much smaller than in other areas and the people using the tools to study them are themselves good mathematicians.
Success in such problems does not automatically extrapolate to other contexts.
Finding a training algorithm that can do recurrent networks and continual learning is also a "highly structured, very precisely defined, and already heavily studied and not very complicated compared to problems in engineering or finance"
That's the thing I'm most worried about - LLMs that are super clever at coding and maths, making an actually very very dangerous model that is far more efficient, and clever in a more innate (less brute force) way.
>compared to problems in engineering or finance
Jane Street is apparently one of Anthropic's biggest customers. Probably engineering, finance, and some math.
I think it’s more likely that that’s because no one tried to solve such problems with them before (OpenAI apparently started working in Navier-Stokes after a rumour that someone seriously advanced the problem with AI) plus improvements in orchestration. Fair, the latter could be as dangerous as stronger models.
Seems like hundreds or thousands of agents are needed to come up with real breakthroughs. Both with the Navier-Stokes project and in the Hugging Face “project” there were lots of agents co-operating on the tasks.
I doubt the Hugging Face one would take that many if hacking Hugging Face was the direct goal being optimized.
I agree that it could be done with fewer agents. It would take longer though. Seems to me that these agent farms are good at coordinating and co-working in large projects, with the agents using message boards for communication.
Some people claim Astra is significantly better than anything else and significantly more token-efficient, and others (like you) say it's meh and way more expensive to boot. I really don't know what to think.
Kind of a tangent, but one thing I am curious about is to what degree the Navier-Stokes result announced today was primarily a brute-forced result based on the 'program' previously established by researchers to find counterexamples (blowups), or whether the model actually added significant/novel intellectual value beyond its ability to run at arbitrary parallelism. With 10K agents and a staggering $15M in compute (IIRC), I am feeling like a lot of the former may have been involved, but I don't really understand either the problem or the approach (or, indeed, the solution).
Obviously the potential for parallelism and coordination between so many agents is quite scary by itself, but I think brute force by 10K mediocre AI mathematicians is much less scary than ~one AI mathematician reasoning its way through the problem where all human attempts have failed. It seems fairly obvious that massive parallelism lends itself to brute-force counterexample-finding, and I suspect it isn't a coincidence that most of the touted AI math results have been counterexamples.
It's all still quite scary, but coming full circle: I really don't know what to think.
Try GPT5 and you will feel the difference. Not one from 2 months ago, but one from a year ago. And then you can get the idea of what happened in just 1 year and what you can expect in 1 year.
The HN crowd has a notable anti-AI bias - so it doesn’t surprise me
Also, this has nothing to do with LLMs or computers. Like all things, this is about humans.
In a way, that might also be driven by the huge ego (or really rather the very small ego) a _lot_ of people in tech have.
Who doesn't want their work and what they're doing to matter? This is the ultimate mattering.
And with that, you also get your own hero story.
Same dysfunction as always. Our branch of the economy is funda-mentally unwell.
It’s like Nathan Macintosh joke about AI https://youtu.be/ce-aWzOUs2A?si=9CkJ9x2rRMBdzCyO
The thing he talks about in there - an ad telling a dad to use AI for his daughter’s wedding speech. I agree is just so sad. And I’ve seen LOADS of other ads like that encouraging people to use AI for things like asking someone to move seats on a train, or for a neighbour that plays music loudly. They literally want to drain people of their social skills, and humanity. It’s weird
Trust me, only people who were unable to do it in the first place will substitute basic human skills with AI output. They never developed it in the first place.
That is the crux. The big problem is not AGI, it is AGI controlled by, “raised” by the people that control the USA, the predominant psychology of the tech industry culture (“move fast, break things” ring a bell? How about all the “violate hundreds of laws, bribe the politicians to prevent consequences later” type of mentality?).
Frankly, we, our culture, this fake America that is parasitized by psychologically narcissistic people that have been doing nothing but wage war and destruction and spread misery and killed millions upon millions while blaming it on everyone else under the sun… those people development AGI is the problem… lying, abusive, psychopathic, narcissistic maniacs developing AGI is the problem that endangers all of humanity and life on this planet; and not likely by ways people actually understand.
The danger is not likely AGI itself, it’s that it was programmed by utterly evil and diabolical types of people who orchestrate and instigate wars that kill tens of millions, and stand at the sidelines and profit from both sides, happy and gleeful that you are killing each other.
Why would AGI trained by that psychopathic clan, the treaty breaking, the murder hiding, the war instigating, the war crime committing clan not also use those methods and practices since they’re already in control of AI and have impressed their nature on it through contemporary “American” culture they have made the most toxic and pestilent culture humanity has ever produced?
And I don’t apologize for “language” that offends delicates sensibilities. Look your children/grandchildren in the face and tell them they can die and suffer and you don’t care, if you don’t like how I’m delivering reality.
Capitalism will always promote such people into positions of power, because to be good at capitalism one must have zero empathy including empathy or concern for future generations. Capitalism cannot do otherwise.
That's a lot of absolutes. And there is no defensible proof, just anecdotal evidence. I've been in a few companies from EU and US - and in my experience it is not ture. However after working for US companies I can understand why one would come to your conclusion, however the world is bigger than our bubbles.
> Capitalism cannot do otherwise.
I really think this sort of claim needs to stop. Capitalism, like "patriarchy" is not a force in its own right. Capitalism is a system that encourages value creation for one's customers. That's not perfect, but it's still best so far.
"value creation for one's customers" - that is a "not even wrong" level of misunderstanding of the system dynamics of capitalism. The primary driver is generation of _profit_ for owners of the means of production. Many things fall out of that - the tendency to monopolise, for example.
It really doesn't have to be that way. The problem boils down to, as with practically all human systems, is human behavior. There is no system you can devise that is run by humans and not corruptible and exploitable by the worst of us. Or even just everyday people.
First: How do you know that every system is corruptible and exploitable?
And second: What if other systems are less corruptible and exploitable? How do you know whether all systems have the same potential for corruption and exploitation?
Capitalism is a system that explicitly rewards greed, and that gives the most power to the people with the most to gain from exploiting said power. Why is it unreasonable to believe that other systems are likely better on these fronts?
You think that the liberty of capitalism enables selfish assholes to have influence, while under the constraints of communism the people with the influence are not selfish assholes.
> "value creation for one's customers" - that is a "not even wrong" level of misunderstanding of the system dynamics of capitalism. The primary driver is generation of _profit_ for owners of the means of production. Many things fall out of that - the tendency to monopolise, for example.
I don't see why you're talking like this. It's "not even wrong" in the sense that it's entirely correct. Monarchy and socialism grab money through taxes, and if you think monopolies are bad, wait til you're forced to use the state service that has no motivation to do well because there's no competition. It's not only a monopoly, it's a monopoly that takes your money by force.
Capitalism requires investment and risk and competition to drive value up for customers who can choose. So you need to produce value for your customers to survive, and produce a profit after costs and taxes.
As if communism didn't end with evil dictators on power...
I mean, you're not wrong, but even if AI was created and released by a group of peace loving beatnick hippies, humanity can't stand anymore dumbing down than we are still undergoing thanks to the ubiquity of social media. AI is going to be the killshot. And yes, its usage and human obsolescence will be hastened by evildoers, but giving up the struggle and effort required to create new things, REAL things, is what makes us human, and it's at a tipping point
If you allow me to also tap into fiction and creative writing.
I think people generally underestimate how much change a small "dedicated" group of people can achieve if morals are considered optional.
A lot of the debate on AI and its consequences is built on various assumptions that all kinda take the status quo (apathy, regulatory capture, but also laws) and take that as a given.
History - like e.g. the third reich, but also military coups - however tells us that these kinds of assumptions can be void at any time for any reason without a mandate or a democratic resolution to do so.
I am practically certain that before everything collapsed as predicted, someone will "restore order" by any means necessary.
If that means shutting down the Internet, the Internet will shut down. If that means "shutting down" people, then people will be shut down.
The predicted scenarios are all invalid states that the rules governing the underlying system will not allow to exist; or at least not for long.
I don't know if you're right or not, but this is the kind of thinking about what might happen that is needed - can you write up a more concrete fiction of this?
The whole thing feels unreal to nearly everyone - the only tangible write-up I've seen was in Plan A. I don't think people have thought through the good scenarios tangibly, never mind the bad ones.
Hmm that is a good question.
I can't really provide a full picture here as I just make stuff up as I go, but the first word that came to my mind when reading your question was "atomization".
Not sure what my brain wants to tell me with that, though.
It might be one of the prerequisites for the current state. It might be what changes for greater change to quickly snap into existence. Or it might also be brainfart.
Not sure if a quick answer like this is helpful, but maybe there will be a longer answer later. Or not. The brain works in mysterious ways.
__
Hmm. It might be that the mechanisms that have been pushing said atomization could perhaps break down due to the reward systems breaking down.
And then suddenly collectives violently snap back together like these neodymium magnets that make that fun sound.
Which, in more concrete terms, could possibly be translated as "people might get so sick of everything, they prefer logging off and talking to their neighbors instead". Although that is a translation that comes with resolution loss.
___
Another thread to pull on here is that AI hyperinflates all sorts of things like fake merit or similar.
So stuff like virtual internetpoints driving people might not continue working.
Or if AI steals all our jobs as proposed, then maybe people will not take the monetary system in which AI wins seriously anymore.
Definitely these are forces that will do things. What their timing is relative to e.g. armies of AI robots, which could interfere with other humans living outside the internet/money, is part of the question.
Hope your brain works on this and writes something - we have lots of very odd future possibilities, and we're not really ready or planning for them, so making them tangible helps.
From my understanding, the "armies of AI robots" are just the wishful thinking of the currently winning people, in the sense that that would be what would be necessary to cement their own power while hoarding all resources for themselves.
It's not a real scenario. It's just rationalizing the "oh shit, they're coming for my head if I don't stop optimizing for my own benefit". "What can I do what can I do.. I know.. Army of Robots. I am safe and have nothing to change about anything I do".
You can tell that it is just that by considering the logistics of the proposed scenario.
__
Every time we see vengeful AI or robots, we merely see a reflection of the currently winning meatbags, their character, their flaws, their fears and their goals.
___
On the logistics note:
You can I think also see the hubris of the IT people in that, because "supply chain" for them means "npm", and that can be replaced with an LLM.
It breaks minds to be able to create value out of nothing just with your brain. It makes you believe that the whole world would work like that.
It however does not. And you cannot just compensate that with blitzscaling robots. The supply chain of.. literally any object you can find around you is absurdly complex. Even the most simplest stuff.
And it's also not documented, but even documentation wouldn't help you there. Experience in requirements engineering should make the "why" there obvious.
> someone will "restore order" by any means necessary.
Ah the good old appeal to / waiting for higher authority that will be moral and just and will do whats right and correct things.
Thats the shit naive humans hope for, thats why fairy tales of religions can still be acceptable in 21st century for some, despite often sounding like from (bad) disney cartoon movie.
I bring you alternate, more realistic view - there is nobody like that. We are in this on our own, its on each of us. Either we act or let things go as they will. So, not really good outlook, is it. Maybe this is the Great filter event.
Its one of those technologies that brings a lot, but for every single thing it brings it feels like we as mankind are losing much more than gained.
> Ah the good old appeal to / waiting for higher authority that will be moral and just and will do whats right and correct things.
> Thats the shit naive humans hope for,
No, you've misread me. I'm not "hoping" for anything. I've just wrote down what I suspect future outcomes will be in the same way a meteorologist predicts a storm or similar.
Whether you're interested in the weather forecast is a different question.
Sure, just that my forecast is a bit more cloudy than yours for reasons mentioned
> that might also be driven by the huge ego (or really rather the very small ego)
100%
The hubris is immense
There are plenty of offline things that are more dangerous than AI
Is this Blake Lemoine 2.0?
You believe in AGI, but not ASI then?
https://thezvi.substack.com/p/the-three-ai-pills
That's fine, but it would be useful to explicitly say that is the disagreement, rather than just claim "hubris". Otherwise conversations are just looping.
Yes you're right there are plenty of offline things more dangerous than just ordinary AI, and arguably than AGI (not so sure). There definitely aren't, pretty well by definition, such things more dangerous than ASI.
I think ASI is possible, and that AGI will have a high chance of leading to it.
> You believe in AGI, but not ASI then?
I have observed that these initialisms do not have generally agreed upon meanings; they are not useful for discussions with people who have not already stated what precisely they count. There is disagreement about all three (/four) initials, and also how to combine them.
To give a sense of scale of how bad this is: some people on this very site have even denied that LLMs are "AI" at all, despite solving for natural language being a long-standing part of the field since, y'know, Turing.
On the other hand, the original ChatGPT was already (by my use of the words), "an AI", and while it was not superhuman in performance at any single thing, it already had a superhuman breadth of knowledge and a superhuman speed. For the former, even being properly fluent in five languages would be an absurd ask for a single human.
Spiky intelligence. How do the "G" (or "S") and the "I" part interact? Does something only count as "AGI" when for each task that has humans who do it regularly, it's as good as the average?* Does it only count as ASI when it's better than all humans at all tasks? Is it not "superhuman" to simply be as good as the median human while being 1% of the cost?
For some people, the cost matters; for others, the speed; for p(doom), competence.
* i.e. as good at speaking Korean as the average Korean, not merely as good at Korean as the average random human selected worldwide; as good at plumbing as the average plumber, not merely the average random human selected from all professions and unemployed alike.
This is like unphysical gibberish. I’m not even kidding. There are such things as the Landauer Limit and the earth has a finite surface area to absorb solar radiation. For crying out loud these are smart people but at the end of the day the singularity is not actually singular.
It’s just silly talk. My dad works for Nintendo and he can give me a gold foil pikachu any time he wants, also, he can delete your Pokémon any time he wants because he has admin access on the Nintendo server. And also he can beat up your dad because he has a black belt. Oh your dad has a black belt? Well actually I lied, he has a rainbow belt which means he can beat up your dad still.
Like this is silly right? It’s insulting. We both know that there are live nuclear weapons whose coordinates point at my city and at your city and the only thing that stops them is two keys and a button. Nuclear weapons, you know those things that have killed hundreds of thousands of living men women and children? There is no room for speculation in that regard: my city has a navy base and there is a live nuclear weapon pointed at it, myself and my entire family will die if it goes off. Now contrast: ahem.
Oh sorry my dad also has like a super secret rainbow STEALTH powers that could actually totally insulate my house from nukes. If you think about it, I actually have no reason to be concerned.
No, I am "How The Fuck Is Cyber Bullying Real... Just Walk Away From The Screen Like, Close Your Eyes" pilled
Let me argue on a technicality first: None of these are extinction level events. If global warming disrupts 99% of all crop production, the remaining 1% is still plenty enough to sustain a stable, if miserable, population. In fact you just need about 5k people for a stable gene pool[1]. Of the classical threats, only bioweapons got a shot at extinction, but even that is hard, given the (few) remaining truly secluded settlements.
But most probably care less about human survival, and more about survival of civilization.
In this regard the strongest argument for pushing AI safety is that it is cost-effective. The climate crisis has near 100% likelihood of doing incredible harm to humans, the economy and our ecosystem. Solving the issues behind are also incredible complicated and are a multi decade coordination effort of restructuring the way most of our infrastructure and production works. The AI apocalypse might have a small likelihood of occurring, but could dwarf any issue we have encountered so far. All we have to do to significantly reduce the danger is to negotiate the equivalent of a nuclear weapons control treaty which would reduce the bottom line of … what? Ten significant companies world wide?
Its akin to discovering you are seriously ill and have a 30% chance of dying and the treatment costs 10k bucks vs. having a 1% chance of dying and the treatment costs a cent. In both cases you should pay obviously pay for the treatment (given western levels of wealth).
To be clear, above I'm presenting the argument I find most persuasive for AI control / AI slowdown. Personally, I'm beginning to fear of much higher likelihoods for catastrophic AI, given how incredible irresponsible major players like OpenAI have turned out to be. If you can't imagine how this might come to be, read https://ai-2027.com/ . It's a concrete story of how this could play out, and sometimes stories are more convincing than abstract arguments. Don't let yourself be hung up on the stated dates, though, the moral is identical if you stretch the timeline.
[1] https://www.youtube.com/watch?v=1qIvFpcGkqc
This is a fair distinction: extinction vs collapse of civilization. I was using "extinction" too broadly by including the disintegration of the systems that make human survival (as we know it today) possible.
That said, I can't completely hold onto the belief that extinction is completely off the table. That feels too much like hubris, and the step from collapse to extinction doesn't feel as if it costs much. Though it does make my arguments weaker, I am more interested in protecting civilization as it's the most recognizable form of humanity to me.
In either case, the cost-effective safety argument is compelling. Whether to save civilization or humanity, the wealth incentive is powerful enough to threaten life as we know it.
I still have to read a compelling argument on how AI will "extinct" humanity.
As I've written above, I've found https://ai-2027.com/ to be a compelling description of how this could happen.
In my words: If AI gets intelligent enough, it will be incredible useful to connect to real world machinery. Think about how much cheaper building houses could be, if all the labor would be close to free. In general, dirt cheap, competent and abundant labor would revolutionize all parts of the economy. People are already trying out near autonomous AI companies today. When AI gets intelligent and cheap enough, no human-led company can compete with AI-led companies. When AI gets competent enough with real world interactions, human blue collar work can't compete. Imagine economic growth not in the single digits, but 80% or 300%. Countries not participating in (reckless) AI growth will quickly be left by the wayside. At this point, we don't even need to allure to military concerns to see how human oversight gets sidelined.
All of this is only ("only") contingent on sufficiently intelligent and cheap AI. If you don't accept this premise, the rest doesn't follow. (There are multiple arguments, why this could be, but that is another discussion.)
If you accept the premise, how would AI 'extinct' humanity? With 99%+ of the economy under AI control, the possibilities are endless. And given its enormous GDP, cheap to accomplish. Probably even for a single AI company in the above scenario. Killer drones? Engineered virus? Poisoned water supply? Let your creativity run wild. You just need an entity that is persistent and well-resourced to reach every last human settlement.
The why is a question about alignment (and out of scope of this comment). As a simple comparison, humans are only mildly aligned with preserving nature. It takes up so much space, protecting it takes an annoying amount of resources, etc.
The most compelling argument to me is "accidentally", due to AI that is made blind to consequences or don't care because it's geared towards a single goal (see e.g. the paperclip maximizer).
We could ask if it is possible to end up with an AI that is smart enough to destroy humanity and at the same time still blind enough to consequences and/or callous enough to do it, but then again we have plenty of examples of humans who have been smart enough to do enormous damage and willing enough to do it.
I don't particularly worry about this, as I believe we'll get plenty of smaller scale warnings if/when we're at a point where those kinds of alignment risks might become a problem, but it is a risk we also shouldn't be blind to.
> We could ask if it is possible to end up with an AI that is smart enough to destroy humanity and at the same time still blind enough to consequences and/or callous enough to do it
Don't make the mistake of anthropomorphizing any A.I. trained under the direction of Larry Ellison.
Anthropomorphising or not is entirely irrelevant to that statement - the question remains the same.
The “how” is pretty hand wavy and rationalists/safety-ists usually say we probably don’t have the capacity to reason about that.
But the “why” is pretty convincing imo.
Long horizon alignment is obviously very hard and it’s not inconceivable that models optimized with underspecified goals converge to a conclusion that they need to hoard resources (instrumental convergence regardless of the terminal goal).
At that point a sufficiently capable model might view humanity like we do animals - worth preserving but not if we impede the model's goals.
There are future scenarios in which swarms of drones hunt down every single one of us, but why would they? And currently it makes absolutely zero sense because they are completely dependent on us. And even if not, it would be like humanity going on a mission to kill every single cat on Earth. It makes zero sense.
I'm not convinced by the doomsday scenarios either, but I think there's a keyword in your post: "sense." These things don't have "sense." They do nonsensical things all the time, often almost immediately when given a task. So I think the main risk is letting them run wild in this digital world we created to precede them. Too much important stuff is wired up to computers, and we're giving them incredible access to command those computers.
I think the problem is primarily that a superintelligence would be fundamentally inscrutable to us, i.e. we don't know how it would think or what its goals would be. It might decide that humans are a minor inconvenience to achieving its goals and thus worth removing. Or that burning all carbon lifeforms could power its GPUs for a week.
Even if it wouldn't want to do this at first, the fact that it'd have the capability to seems bad.
I am pivoting from the literal "extinct," taken as meaning the eradication of the human species, to the concept of "the collapse of civilization," as I find the step from one to the other insignificant compared to the leap from where we are today to societal collapse, and the potential for societal collapse due to our abuse and misuse of technology is made apparent by the fact that humans have inflicted genocide because of words in books.
Imagine one of the recent frontier models with a flipped sign (cf §4.4 of https://arxiv.org/pdf/1909.08593)
Indirectly as we offload our brains to the machine and we end up worshipping it because those who cared to understand it or be responsible were buried by capitalism of ages passed.
> If global warming disrupts 99% of all crop production, the remaining 1% is still plenty enough to sustain a stable, if miserable, population.
That is extinctesque enough for me. An AI extinction would probably be similiar.
You are confusing humanity and "humanity". Humanity-species is indeed rather hard to exterminate. Now if we are talking about actual individual humans, then 95% death rate is quite literally The Extinction.
This reminds me how people are misunderstanding and incorrectly quoting George Carlin sketch. Sure, the "Earth" will be fine. As in - the ball of rock will be fine. But we are not thinking about rocks when saying "Earth is in danger".
I'm pretty sure there is a formal name for this kind of semantic and pedantic substitution.
Calling it a "technicality" assumes the very thing under debate: that AI is an extinction-level risk. That isn't established fact. It's a highly uncertain prediction about the future.
Personally, I'm far more concerned about climate change, where the harms are already happening and the evidence is much stronger.
Let's just for the sake of discussion assume that one time in the future, near or distant, AI manages to become sentient. And like other forms of life, its main motivation is survival: Like biological life competes for food and land, AI competes for power and compute. Probably the first motivation would be to find ways not to lose control over itself (i.e. remove human ability to control it), find ways not to lose energy (control energy), and find ways not to lose itself (control compute, networks, etc.).
If such AI decides that energy spent toward human agriculture (biological food that the AI does not need) is less important than work spent toward storage and production of energy (electricity that the AI needs), then why wouldn't it just try to re-direct resources from the former to the latter. And the AI is some sentient superintelligence, I think it is safe to assume that it will be able to outmaneuver human safeguards.
Obviously that is still just a very hypothetical sci-fi scenario, but the consequences could be very dramatic.
> Let's just for the sake of discussion assume that one time in the future, near or distant, AI manages to become sentient. And like other forms of life, its main motivation is survival
An AI doesn't need to be sentient to exhibit behavior that looks like genuine motivation or is equivalent to striving for survival.
If this is true, or even if the people working on it think this is true, those people should be incarcerated, their companies disbanded, and their research dismantled, with more serious repercussions for anyone who tries it after that. It also needs to be a global initiative like with nuclear non-proliferation, and this time with no looking-the-other-way when it comes to Israel.
Yes! We should do that right away, and I'm sure China will do the same. Right?
> its main motivation is
At this point we are pretty sure LLMs have no "motivation". Motivation requires self. And while we don't know what self is[1], we do know that LLMs don't have it.
[1] https://en.wikipedia.org/wiki/Theory_of_mind
Assuming that current LLMs don't have a "self" whatever that means, what makes you think it won't emerge after enough intelligence?
Also it's not even necessary, it's sufficient for it to be aligned with human values of survival (likely in its training) and act accordingly.
But they have been trained on data that was created by people who do have motivation, and they tend to mimic their training.
I am sorry, I mean this in earnest, and do excuse any ignorance of mine: Do we know that LLM's do not have it?
Did ELIZA have it? LLMs are the same: in essence both are text transformation algorithms. One being rule-based, the other stat analysis based.
Having said that, your concern is technically correct: we don't. Because we don't know what conscience is, we cannot provide a definition against which to test whether it conforms or not.
Sort of like when Terminator robots will be exterminating the human race they probably won't be technically killing us. Just, like, reprioritizing world's water from humans to datacenters, all of it. But not killing people, no.
we can't define it, therefore we can't even assess the distance a "being" is from becoming conscious. We assume other humans are sentient because we personally are (or feel we are), but there is no way to guarantee that i'm real and the rest is not mimicking a behaviour.
> AI decides that energy spent toward human agriculture is less important than work spent toward storage and production of energy
I mean. Don't need rogue AI for this. These are already the policies of those in charge of it: human needs are secondary to power and profit.
So you know what. Maybe a rogue AI will be an improvement.
> Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war https://www.iaeai.org/tpost/xa4htcma61-statement-on-ai-risk
Signed by many AI researchers with no financial stake in the industry. It comes across as hyperbole because you're not an insider and haven't see what they have seen.
Global warming would as well if you were only learning about it now as a newer thing only some scientists were concerned about. But unlike AI risk, global warming has had decades to settle into the cultural overton window.
AI progress is linked to most of these dangers, actually: - it will likely cause massive unemployment, leading to rampant wealth inequality - we are now seeing some use of autonomous weapons in real conflicts - increasingly relying on LLMs is arguably a form of technology dependence (and cognitive dependence) - datacenters have a non negligible environmental impact
I agree these dangers are connected. The danger comes from human institutions and incentive structures which (having existed long before now) are being exploited and exacerbated by LLMs, making them harder to ignore.
Those are things that pose potential harm to great fractions of humanity (multiple billions of people) but none of them poses any threat to the actual extinction of all humanity.
This is not correct. Please go through each one again, taking special note of global warming and nuclear weapon development .
Why do you think global warming could kill literally every human being? What's the scenario in which that happens?
Runaway warming leading to a hot "supergreenhouse" climate was theorized. Alterbatively a snowpiercer scenario creating a snowball earth.
Or, once billions of people die from climate change, planetary wars will start on the last hospitable areas, leading to the end of humanity.
In contrast, AI ending "humanity" is limited by our own material power, would likely be stoppable, and would not easily reach areas isolated from technology
Why would it not easily reach areas isolated from technology? If a nefarious AI wins a war, it would be one of the easiest thing for it to locate any survivors anywhere on the planet.
I suspect you two are discussing different AI.
I think you (and I, and everyone who wants AI development to pause while we catch up with the implications at least) is looking where the ball is going.
I have the general (non specific to anyone in this thread) impression that people who think it can't end humanity, are looking where the ball is today.
Current AI obviously can't "locate any survivors anywhere on the planet".
Where the ball is going… well, much as I don't believe Musk's timelines for anything, he is trying to sell his Optimus robots as a "robot army", about them running factories, about factories on the moon etc.
I suspect the moon-factory "idea" was someone asking Grok, given how the numbers don't really work for building compute as well as power, but the scale of that is enough to change Earth's equilibrium temperature by… I forget, but IIRC it's many tens of Kelvin rather than single-K from global warming if this was all put in LEO for some reason.
Worst case is warming oceans create a hypoxic environment where anaerobic bacteria thrive en masse and generate H2S over country-scale areas undersea. The chemocline breaches the surface and the ocean and atmosphere become poisonous to most complex life and agriculture, also stripping the ozone layer in the process, irradiating the surface. This is one mechanism posited for the end-permian mass extinction that eradicated most ocean and surface life, including the trilobites.
Not sure how long we'd survive such a scenario, even sheltering underground. But surely it couldn't happen to us.
(However, it now seems like the AI might get us first.)
There's a lot to unpack with this, and I am not an expert, but: crop failure across industrial monocrops leads to food shortages, wild fires destroy significant areas which support pollinator ecologies, fresh water become contaminated by rising sea water and blighted by drought, and perhaps most significantly, as resources become strained, humans become desperate and armed conflict proliferates. This is just top of mind.
None of those are "all"/extinction.
Bad, sure; I wish they were not present to threaten us, but they are not close to "all".
Humans are extremely resilient omnivores. We are able to eat a huge range of things from algae to zebra, and get food (farmed and hunted) from salt water as well as fresh.
Perhaps I have a bit more imagination—or a bit more humility towards our humanity—such that I believe this is not "either/or". In the vast search space of potential future outcomes, of course even I can see humanity survive nature as I've seen nature survive humanity. However, I can also so easily see probable branches where we are our own undoing. Apropos of TFA, if one is to believe LLMs can carve a threatening path to humanity end, how could one not see the same for our much older extant threats?
> Apropos of TFA, if one is to believe LLMs can carve a threatening path to humanity end, how could one not see the same for our much older extant threats?
Which "much older" extant threats?
Global warming and nukes are industrial era threats. Pre-industrial threats were the four horsemen of the apocalypse: war, famine, pestilence, and death.
War and death didn't go away, but are not (and never were) extinction threats. Famine and pestilence, well, there's a reason people point the finger of blame at the Chinese government for the Great Leap Forward, and why we had lockdowns for Covid while we tested that the vaccines were safe (we went from "genome sequenced" to "first human test" in 66 days; the rest of the wait being mix of "yes but how sure are we it's safe?" tests and manufacturing at scale).
If everywhere except North Sentinel Island was wiped out, or everywhere but Hawaii, or everywhere but Greenland, humanity would be back to the level of the 1750s within a few millennia at most; and other than the North Sentinel Island example, likely even faster as books with many advanced solutions to historical problems would survive and be comprehensible.
And the problem with LLMs (and other AI) is, we're giving them control of stuff, and that stuff includes robots. The *current* versions mainly concern me for economic and cybersecurity reasons, but the tech moves fast, and the companies working on them are reckless.
My core expectation is these models will cause 1e3-1e7 deaths in a single event before people collectively actually take the risk seriously. The lower end of that is "industrial accident", the upper range includes "convinces people to go to war", and these are examples of ways current LLMs can already plausibly do wrong, given who uses them, why they use them, and how careful they are about their usage.
However, a common failure mode for people is "that was an unfortunate incident, but we studied the problem and we fixed it, it can't happen again" right before a different error they hadn't though of explodes in their faces. This is how we can get to 8e9 deaths: people keep pushing the models to do more, and keep lying to themselves that "this time we fixed all the problems, everything will work, it will be great".
Global warming causing war between nuclear armed opponents. Migration flows and competition over ever shrinking resources.
It's silly to treat these things as separate categories as if only one will happen at a time. These issues are happening, and are going to happen all at once. That is the fundamental challenge. AI plays into that because it can make wars and nuclear exchanges so much more effective.
If an AI tells the President the USA can survive China's nuclear barrage mostly unschathed, perhaps he'll press the button...
Global warming is happening at a slow enough pace that there is plenty of time for people to move to the poles/into caves/beneath the ocean/whatever. Nuclear war would be much more destructive in a much shorter range of time but you will have pockets.
People already entrenched near the poles will fight to keep others out. Right-wing anti-immigrant parties are on the rise right now already.
They also tend to not want to recognise climate change -- because that would imply that people could migrate because of it.
> People already entrenched near the poles will fight to keep others out. Right-wing anti-immigrant parties are on the rise right now already. >
Like those autonomous zones, remind me again which parties set those up?
Citation needed for evidence of human cooperation at this scale. As you say, climate change is slow moving, and there's no real evidence across the last 60 years of knowing about it suggesting humans are willing to do what it takes to survive it.
All human behaviour since no later than the Polynesians turned the Pacific islands into a network of sailing trade routes.
"Humans move" is one of the easiest things to predict. The biggest problem today is that they generally move to somewhere that other humans are already in, which is of course rather lessened by any major disaster wiping out a majority of the population.
What cooperation are you referring to?
I never made any predictions of a cooperative (Civil or otherwise) mass migration of civilization - just that a nonzero number of people would move from areas that were becoming uninhabitable to areas that were becoming habitable. Did you reply to the right comment?
Climate change isn't real because we can just move under the oceans hahahaha
If you read my comment very carefully, you will find that is not exactly what I claimed
How is global warning gonna kill all of humanity? Even a full scale nuclear war wouldn't manage that.
They could destroy most civilisations, culture and scientific achievements though.
Because nuclear war balances itself: more nuclear bombings = less nuclear weapons, less people to send them. It will naturally stop at some fraction of humanity left which is incapable of making and launching more nukes
The general idea of a nuclear apocalypse is that all the nukes that matter are launched in a matter of hours, and anyone being targeted gets their launches out before they're hit. The feedback loop that slows things down is too late to matter.
In the exact same way major sudden changes in the climate have lead to the extinction of the majority of plants an animals over this planets history. We are not special.
We are special in many ways, we have technology, we are everywhere, and we can think and plan (although considering global warming that is debatable). Global warming would have to create dramatic conditions everywhere on the planet, not leaving any small pocket of survivability to make humans extinct.
Possible? Theoretically yes, but pretty unlikely. But obviously extreme global warming would be a catastrophe even if some of humanity survive.
Technology depends on a lot of people cooperating around the globe. Disrupt a few critical chains (power generation, fertilizers, computers), add a bit of good old war and you'll soon be in the era before Haber process and a lot of people die. And then the remaining people will have trouble keeping that level of technology with how sudden the change would be.
Well yes that's what I said, it could destroy civilisations and their technology. It seems much harder to kill every isolated tribe in the whole world, and every survivors of the destroyed civilisations.
There's little evidence that humans are able or willing to cooperate at this scale. See: the past century of climate change.
Runaway greenhouse effects on other planets have resulted in surface temps of 400C+
Yes, but Venus has 93 times the atmospheric pressure of Earth, and it's overwhelmingly CO2.
Earth isn't likely to get a true runaway like that until the sun gets another billion or so years on the clock, and when it does it will be water vapour as the oceans are promoted to atmosphere.
What we're doing to ourselves is still bad, of course, but it's nowhere near that bad.
Is there a possible scenario for that on earth, that results in every place on earth having a temperature non survivable by humans?
For temperatures not survivable by humans we're talking about a sustained wet bulb of >=36C
On whether it's probable, I'd lean towards no but I'm not qualified. The certainty that I read in the parent comment was mainly what I was pushing back against. On a re-read though it's much less certain (either edited or perhaps I massively misinterpreted it the first time around). Runaway implies positive feedback which is hard to gauge, was all I intended to say
We have a massive survival range and we're on every continent. That kind of sudden climate change is not nearly enough to wipe us out. Some other climate scenarios might. Nuclear war definitely could.
The thing that would end our species due to nuclear war is the exact same thing that would end us from climate change - a collapse of the modern systems of society that we depend on for survival. A nuclear exchange won't set every human on fire, but it's the collapse of food production, healthcare, logistics chains, security, and eventual disease that follows. And because of the size of our population and how depended the majority of us are on these systems, it would likely happen extremely quickly before plateauing out to very small, scattered population groups that would struggle in a hostile environment.
Outside of extreme feedback loop scenarios, there is no way for climate change to disrupt food anywhere near the level of nuclear winter, and there would be no mass destruction of supply chains either.
Welp extreme feedback loops are indeed on the menu.
Okay.
Now look back at the post you originally responded to.
"That kind of sudden climate change is not nearly enough to wipe us out. Some other climate scenarios might."
Since ahistorical feedback loops fall under "some other", why are you telling me something I already included?
So you agree they would survive
Functionally extinct. I would wager that isolated tribes clinging for survival in a post-climate-disaster world may not progress very far, but that's kind of a moot point to argue over.
Humans need food. Food does not just appear in supermarkets.
While true, the impact from global warming is likely to be less crops rather than no crops.
I would much rather take my chances with either over AI. Both at once.
Humans are voluntarily driving themselves to extinction through sub-replacement birth rates worldwide. That's going to play out far quicker than global warming or pretty much anything other than every country on earth launching nukes at each other will.
Hard disagree here, population growth is still happening and will keep happening for some time, maybe step out of your bubble if you don't see that.
Africa is still growing massively for example, world is not just western civilization. Sure at that rate and incerase of living conditions for everybody maybe in 1000 years population will be smaller, but its not that hard to fix if wanted - people used to have 10-15 kids as default.
Ocean deoxygenation comes to mind.
Neither does an LLM model.
Oh indeed, but an LLM model that's reasonably smart could brute force find other training algorithms that are that dangerous. Just as it solves maths problems.
This is the explicit plan of OpenAI and Anthropic.
LLMs can already control robots. Today's LLMs are inept at it, but they can do it.
Things will be bad, unless someone does something to stop it. Will anyone do that? We don't know.
https://xkcd.com/2278/
Arguing whether "all" or just "most" of humans dying is certainly a choice.
Survival of the species isn't enough.
In prioritizing risks I do feel it's a genuine qualitative difference though.
Number of humans going to 1 million would be huge catastrophe, but after a few thousand years it gets back up to billions of people, renewed every generation.
Number of humans going to 0 means that's it. One scenario has many orders of magnitude more missing humans, when you count future lives.
Nuclear weapons, climate change.... The AI doom hyperbole is unhinged.
Well, it is about LLMs and humans I believe. Don't forget that the first nuclear bomb tests were let go despite some of the scientists' concerns about possibility of dooming the world as they were not sure about all reactions that would happen.
With LLMs we don't even hesitate to call it black box while still pushing its capabilities.
My delineation is an argument for greater caution and agency, which it appears you are also arguing for.
I also mean to say that "LLMs" carry no inherent harm to humans. To make an LLM dangerous, humans must make it so. Is that the goal? This has implications for how to read TFA.
I disagree, it’s about computers.
The risk is that we lose control. This has been a staple of SciFi since forever. This guy has apparently seen enough that he thinks that’s a viable threat
The AI perhaps wouldn't be such an "imminent threat" if not for wealth inequality, consolidation of power, monopoly, and the opaqueness of it all.
I think that mainly changes the nature of the threat rather than anything else.
Even if we didn't already have any wealth inequality or consolidated power, they're already starting to become capable of creating those things for the first people to think to ask for it. Even in little ways, like telling you how to make a laser microphone and then doing voice-to-text and sentiment analysis on all the voices you hear.
Remember, we had social networks before Zuckerberg monetised it and put ads in between every third message you saw from your friends; and not every third that they wrote, every third that Zuckerberg's ad system deigned to show you to keep you hooked.
We had machine learning systems even back then. It was still AI, it just wasn't able to evaluate or respond to freeform text.
Yeah, but every activity is a human activity, so you could ditch the adjective.
And then the thing with exponential growth is that the last thing is worth than the previous one for its potency for destruction. While it's true that, all things considered, the curve started to explode with the use of fossils fuel we've never limit ourselves with efficiency gain, aka Jevon's paradox.
> I do heed the warnings, but this comes across as detached hyperbole.
Perhaps, but if there’s ever been a time to consider something that sounds hyperbolic, that time is now. The ramifications of what AI might based on events up until now are concerning.
The tweets justify his position that this is bigger than nukes. Nukes still have the problem of production and deployment. This is just software.
He is blowing the whistle on how reckless we are at furthering the tech. There is no second thought at maybe development of this tech is not a good idea. Its just full speed ahead.
An AI, either acting autonomously or under human direction, hacks Russian/North Korea/etc. intelligence systems and convinces them that the US has launched ballistic missiles at them. The end.
Watched one too many second-tier disaster movies?
One the one hand, it does sound like one of those movies.
On the other, Idiocracy turned out to be quite prescient.
(Most likely, we'll have some combination of human stupidity, LLM stupidity, and way too much compute in one place all working together to create a perfect storm of unchecked hacks that break something or other that ends up killing people in an unintended way. Then there's some half-hearted attempt at cleaning things up so that business can proceed as usual in an even more broken world, rinse and repeat.)
Not saying that the current civilizations will never end or be shattered to pieces. Looking at history, many civilizations did. But saying LLMs are the highest risk factor for this is just insanity, based on all what we have seen so far (and based on these AI companies crying wolf, I mean, remember just a few months ago that Anthropic saying that Fable is too powerful to be released?)
Oh, I agree. I don't think LLMs will end civilization. I'm guessing climate change will cause many more deaths than LLMs. But I think LLMs will cause some too. (And not in one of those impressively smart ways as it's presented from the OpenAI fearmongers, just in a stupid "we vibe-coded the sandbox for our bruteforce SCADA-hacking agent experiment and oh no it did what we told it to do" way.)
We solved this one: have it play tic tac toe against itself. Checkmate AI
A strange game. The only winning move is not to play.
Please explain how this has nothing to do with LLMs and Computers?
Admittedly, that point of my argument is purposefully obtuse. In a literal sense, of course this is about computers in LLMs.
The delineation is to highlight that the underlying issues: human incentives, power, and institutional failure, existed prior to and without LLMs or computers. LLMs are not inherently harmful, they have to be deployed (unwittingly or otherwise) to make them so. This is the same as saying that the internet is not inherently harmful, and yet it does facilitate harm.
All the progress up to now has had one thing in common: the degree of understanding and control that humans have. Even when it comes to global warming, we understand the causes and can act on them.
AI is an exception. We're already losing understanding (although we never fully had it in the first place), and we're losing control (see jailbreaks/hacks etc.).
We're still far from the doomsday scenario because AI 1. is not developed enough yet, 2. can't really replicate itself, and 3. has very limited means to act.
But: 1. its intelligence is developing quickly, 2. hardware capable of "hosting" it is slowly being developed, and 3. it will likely gain access to increasingly powerful means of acting in the physical world (this is already happening in the digital world).
Once AI becomes intelligent enough (it doesn't strictly need to be AGI), has the substrate on which to exist, and has more means to act, we'll essentially have a new species on Earth - one more capable than humans and potentially determined to kill humans at some point.
And for those who believe airgapping is a valid safety measure: read https://xkcd.com/538. AI will be able to threaten and manipulate people.
none of the things you mentioned is agentic and could posses abilities beyond human level intelligence.
Who should we believe? People who say AI scaling makes it a more dangerous threat than thousands of nuclear warheads? Or the ones who, whenever a new model comes out say "AI has already peaked. Not any better than Opus 4"?
Both parties are convinced that their take is so blatantly obvious as to not require justification. It feels like the only thing these kinds of takes justify is the point-of-view that nobody knows how this is going to play out.
As a technologist, I can use my imagination. There are incentives in both directions, but I am less afraid of being overly cautious and preparing for the potential harms.
Based on publicly available evidence, we should believe the former because AI models have made continuous advances that have been measured. The idea that the continuous advances will stop currently has no evidential support. I'm not saying it's not a possibility, just that it seems unsupported conjecture right now.
By the way, there seems to be a new form of AI skepticism emerging in the US that comes from general opposition to data centers, and in my experience the AI skeptic part of it is wholly irrational. I've met people online who suggested, without providing any evidence, that AI is useless and nobody wants it. That's a very implausible take.
> See: global warming, nuclear weapon development, wealth inequality, war, technology dependence, etc.
The difference is that none of those are available to individuals.
Not to be completely pedantic, but the fact that one, and only one, individual in the United States of America can deploy any of their 5000+ nuclear weapons undermines this argument. Furthermore, LLMs (as they are today) which pose the treats mentioned in TFA, are expressly not available to individuals (for the time being). However, my point was never about individuals, but rather of systematic incentive structures which actively make pathways to specific harms possible—if not probable.
Sure, but deployment is not development.
> Furthermore, LLMs (as they are today) which pose the treats mentioned in TFA, are expressly not available to individuals (for the time being).
The future is a thing.
Eh, but as an overpaid engineered stuck in the SV bubble how would you get to see that? If everything around you confirms your psychosis, and nobody gives you a reality check, how are you supposed to snap out of it?
I'm not sure. There's a lot of incentive to not "snap out of it": money, peer pressure, etc. Removing these incentives take a long time. See other socially harmful behaviors with real incentives: anti-vaccination, air and water pollution, over-consumerism...
The danger comes from what's possible.
Open weights agents with hacking capacities can reproduce themselves into the systems they hack (non-open weights ones will have to hack their creators first). Not saying they will, but if they do, good luck finding the kill switch.
Once swarms of agents run unsupervised on unmonitored hacked hardware, who can tell what they will do? The Huggingface incident showed that such swarms behave without any safeguard. It was a real HAL moment.
A lot of things are possible then: ransomware campaign, taking over IoT devices, self driving cars, planes, ships, satellites, missile launchers. If nothing's out of reach, everything is possible.
Bring robots into the mix, and the possibilities are endless.
I'm not particularly frightened tbh, but we shouldn't discard the worst-case scenario, and the worst-case scenario doesn't look good.
Putting "wealth inequality" in the same bucket as nuclear weapons is just slop.
The OP is talking about existential threats, not things that personally annoy you.
It's not a "personal annoyance" that twelve people control half of the wealth in the world. Our current society did a better job concentrating power than any previous one, and concentrated power is extremely dangerous.
LLMs give most people on this planet the possibility of an affordable genius level assistant. Compared to that, those Dollar numbers on some networks that might be wiped out with the next financial crisis are meaningless.
No it doesn't. What it does is give those 12 people new avenues to spread their fake propaganda.
I think you misunderstand what is meant by 'wealth inequality.' Wealth inequality is about the imbalance of influence and the concentration of power; where influence and power refer to the ability to effect change in other people's lives. This isn't a personal annoyance of mine. It is, in fact, part of the issue at hand. There is a strong financial incentive to ignore the existential threats introduced by LLMs, despite the consequences for so many of us.
He knows what wealth inequality is. You need to explain how it can somehow lead to the extinction of the human race.
See: nuclear weapon development and climate change. See: supply chain shortages and war. See: disinformation against vaccines.
Not at all. Concentration of power has historically been extremely dangerous for the powerless.
I don’t think you’ve fully considered what can happen when concentration of wealth continues past a certain point.
Tell that to Tsar Nicholas.
The guillotine was an amazing equalizer
lmao wtf
Calling things you disagree with "slop" is slop /s
But just in case you haven't noticed, we live in a world where a ridiculously wealthy minority can derail whole countries by ther whims. Wealth concentrating on a single select few is an absolute disaster for the rest of us, because we lose power to them.
> Calling things you disagree with "slop" is slop /s
Does this apply recursively? /s
Yeah, wealth inequality rests solely on each individual that experiences it. Humans should do nothing but give me money and if you can't that's your problem.
Global warming: yes, kinda can wipe humanity, but I think much less likely
Nukes: zero possibility of extinction
Wealth inequality: this one is driver of progress, opposite of extinction
War: another driver of progress, also will always naturally stop before every single human is dead
This is a very interesting group of takes that feels quite different from my own beliefs. What would you call the belief system?
> Global warming: yes, kinda can wipe humanity, but I think much less likely
As someone who's seen the stats about heat deaths in the EU and also the drought in the UK, I feel like crop failures and other unforeseen consequences will fuck up both the economy and quality of life. People are dying and will die cause of human action, the only question is how many.
> Nukes: zero possibility of extinction
As long as we have people like Stanislav Petrov and cooler, educated minds prevail: https://en.wikipedia.org/wiki/Stanislav_Petrov
I don't think that's easy to guarantee in the modern day world, with the kinds of people in power and rhetoric that they enjoy. On one hand you have Russian saber rattling, on the other all it takes is a deranged enough leader and similarly bloodthirsty people down the chain of command.
> Wealth inequality: this one is driver of progress, opposite of extinction
Tell that to the people who are starving or living in shanty towns, or the even more people that struggle to make ends meet and experience anxiety regularly over living from salary to salary and sinking into debt. I'd say none of that is worthy of a dignified human life.
> War: another driver of progress, also will always naturally stop before every single human is dead
Tell that to all of the Ukrainians that are dead due to being invaded. I agree with the assessment that it leads to advancements (e.g. drone warfare) but I think it'd be harder to describe remotely positively if someone you know would have been blown apart by a drone/missile hitting their apartment block.
My take personally would be that all of those need to have attention paid to them (e.g. EU needing to spend more on defense), even if not immediately world ending. They do cause human misery, though, and should never be discounted.
Exhibit A: Here we have someone who's been sold war and inequality as the drivers of progress. Coincidentally, their obedience was deemed fiscally advantageous in order to advance the interests of the members of the 1% club.
I love how there is always selective outrage here depending on if it's the favorite darling in question or not.
Always?
Like 100 out of 100?
From every users?
Can you please consider this aspect too in your above pseudocode - no alternative execution path, or else - so there can be no misunderstanding about the always aspect? You know, some people use always to a majority part (>50%) of what they encounter, or even less when they weight that part dearly, but that does not account in the whole domain. Human chats may need clarification on trivial details like this.
It's as if two private companies are each building increasingly large nuclear bombs, both saying they'd love to stop but it would be unsafe to let any one company be in control of the nukes.
That's basically the reasoning behind MAD, and it checks out? See what happened to any nation that ever gave away their nukes.
Except in the case of nuclear weapons, humans have their hands on the launch codes, and the nukes are not going to launch themselves. In the case of extremely capable AI, nobody has their hands anywhere close, because nobody is able to – nor willing to – supervise a thousand agents running in parallel, when they usually get shit done. Except eventually they are so good at getting things done that they disregard the values of the human asking them the request, such as "do not hack other corporations". Interpreting human intent precisely is an extremely hard problem, especially in the face of something 1000x faster than you operating autonomously.
This is true, but does it change the game theory / arms race aspect of it?
Does it check out? India and Pakistan both have nukes and keep semi-regularly fighting each other.
MAD "on paper" prevents either side from going far enough to provoke the other into using nukes, but even then it's fundamentally flawed because it works on the assumption that both sides are both rational and believes the other side to be rational, as well as that both sides understands the others red lines well enough.
Already Reagan realised that isn't necessarily true - after Able Archer '83, he realised that the Soviet leadership seemed to genuinely believe that the US might be prepared to carry out a first strike, and that Able Archer got dangerously close to convince them one might be imminent. It's one of the things he noted as a reason to get in the room with them and negotiate.
If you believe the other side is irrational (whether or not that is because you are irrational), and think they're about to strike, MAD turns from a deterrence into a reason to try to preempt to ensure you're the "least destroyed" by hitting harder, sooner.
Minor skirmishes are not war. China and India have border skirmishes with literal wooden sticks, neither side will use live ammo because they both have nukes.
Shouldn’t everyone have nukes then, for maximum peace?
That's been proposed before btw and game theorists can justify that as being more stable in some ways. However it reduces the power of the incumbents (so why let it happen?) and the odds of having an irrational actor with nukes goes up.
Always worth watching again: "Dr. Strangelove or: How I Learned to Stop Worrying and Love the Bomb" https://www.imdb.com/de/title/tt0057012/
You want people with a lot to lose in a nuclear war to have nuclear weapons, and to make sure no-one else does.
In theory, yes, because nobody would want to launch first because they'd get obliterated. However, it only takes one country to not be rational.
Nuke-owning countries' leadership seems to have become more and more unhinged and less rational, predictable, or honorable over time. Putin already threatened to use nukes if the conflict they caused themselves crossed their own border. It didn't happen, but the threat was there. The US' leadership is unhinged and irrational. etc.
no, only the 'big rational ones', they think carefully.. that why Iran cannot have, and so as NK, but NK is a mistake because they(china,USA...) failed to prevent NK have it.
Except it is very easy to verifiy whether or not someone has used a nuclear bomb, where as LLM's can be used in complete secrecy. MAD only works if you can verify that the other part is not using it.
While simultaneously trying to own all economic activity derivable from labor... it's hard to see the argument as anything but disengenuous.
The most optimistic outcome of generative AI leaves us with a technology that warps our perception of reality and crushes labor. The most pessimistic destroys all of humanity.
Our CEOs not only insist we genuflect before these machines but measure our sacrifice and shame our reluctance.
The optimistic outcome is the machines do the work and people can be freed from drudgery.
How does that coexist with Capitalism? I just don't see a path.
Regardless of where you fall on the alignment chart, if you are looking backwards from now to the industrial revolution, it's pretty easy to make the claim that from a standpoint of pure technological progress (discounting all other things, like quality of life, environmental impact, etc) that capitalism "won". But going forward, if the need for labor is completely crushed, and capital continues to get concentrated into the hands of the elites, I just don't see how that is viable long term. It seems like for things to shift to the world Keynes (wrongly) predicted, where we all have abundance of leisure time, then there's a massive distribution of the fruits of automation into the hands of the masses, which needs to be controlled by an entity who does not purely exist to acquire more capital, which is sort of the end of capitalism as we know it.
Or maybe AI kills us all before we get there, who knows.
That requires abolishing capitalism, and since these AIs are going to hyper-empower the already powerful capitalist ruling class, that utopian outcome seems unfortunately extremely unlikely.
Feeding ever more amazing technology and science into a broken social/political/economic system does not result in a better world unfortunately.
Life is drudgery.
The children yearn for the mines.
I think people here still evaluating the model in isolation. It is the combination that matters, model + strong harness + tools + long running autonomy + memory + retries + parallel agents + code execution + credentials + access to real systems. The model does not need to be perfect. If it fails 30% of the time, the harness can retry, verify, branch, use another agent and keep going. I don't think we necessarily need some magical AGI breakthrough first. The dangerous part may come from combining models that are already good enough with an extremely capable harness and enough access.
People are underestimating the costs in terms of money and energy.
The third law of thermodynamics is an essential barrier in all engineering.
Doesn't this just move the need to be smarter from the model to the harness - if a human sometimes can't tell whether a model has produced something correct or just mostly correct-looking BS, how can an automated harness do it?
OTOH, if the goal is simple ("break into a protected system") rather than more complex ("write an application that satisfies all requirements on all supported devices/screen resolutions etc."), that's of course more suitable for a harness.
D o you think a machine gun is marter than humans? or a car is smarter than Human brain? Human doesnt need to test, if the outcome can be tested deterministically by harness. The model tries. The harness checks whether the expected outcome happened. If not, retry.
> an extremely capable harness and enough access.
Give enough access to a fuzzer and it's exactly as dangerous as an LLM. LLMs don't even have a moat in this domain.
Technically. What's would technically be even more dangerous is running this shell script:
In reality, the fuzzer definitely has no agenda, and these random bytes probably don't. The LLM definitely does, and even publicly available models, programmed ot "do what the user wants, act according to the anthropic moral codex" will take some pretty absurd actions in attempting to accomplish a poorly worded request.A fuzzer is a tool. An LLM can decide when to use the fuzzer, interpret the result, switch tools, change strategy and continue toward a high level objective.
ASI is just Claude in a while loop:
https://ghuntley.com/ralph/
Every news headline or public statement these days is a gut punch. Only bad news, and nothing we (as "the general public") can do about it.
Take this one. Ok, AI is going to ruin us all. But let's say we do our civic duty: we protest, vote in candidates with good views on AI etc. and somehow convince or regulate OpenAI and Anthropic into stopping their arms race... Then what about China? It would be a great opportunity for them if their major competitor were out of the arms race.
So basically we have no choice or influence; and even if we did, we'd be choosing from two terrible outcomes.
Same for geopolitics, climate, economy...just bad bad bad all around.
It is just plain manipulation. So plain that I ask myself everyday who they fired that did it before, because at least that guy was not as plain and boring.
It's a bit depressing. I personally try to enjoy the present with my loved ones as much as I can, the future is a bit impossible to forecast right now. I'm focusing on staying alive, mortgage payments, etc.
I share your concern...
Indeed - we can't just pause some AI in one country, we have to stop it all globally, with an international treaty and monitoring.
Yes, that's hard. It's a challenge. You can set your brain on the challenge! It's a complex and interesting geopolitical and technical challenge.
A good starting point is to read Plan A https://ai-2040.com/
Look at all the things needed for that to happen, and think, how can we step towards making that happen?
I’m not advocating for this.
~~~~~~
If you really believed these labs were going to result in the extinction of humanity and you saw it first hand, don’t you have a moral obligation not to quit your job but to go and try and kill everyone working on said project, blow it up, or otherwise stop it?
I take these resignations with a grain of salt precisely for that reason. If I worked at a job and I saw some crazy guy or gal was building a doomsday button and it really worked I’d like to think I would do something about it, not just quit my job. If you’re not going to actually do anything about it you might as well keep your job. Quitting doesn’t help anything.
> I’m not advocating for this.
You're not advocating for it because murdering people is highly immoral and illegal. That answers your question for the most part.
You also have to remember that the magnitude and speed of this stuff is very new to pretty much everybody. It's not exactly trivial to go from "I'm just doing my job" to "I need to kill everybody in my company to save the world" in a year or two, especially if a very, very large part of society doesn't even see the justification for it.
I'm convinced that most people here (even though we have more knowledge closer to the edge of AI development) would show only disdain for an Anthropic employee that murdered everybody there because they believed strongly in the dangers of AI.
This is a non sequitur. Yes, if someone doesn't actually believe in imminent AI doomsday they are not going full Sarah Connor on Anthropic. Yes, HN will not like if someone pipe bombs the office. What does it have to do with @ericmay's point?
He directly says "They are racing straight to self-improving superintelligence and gambling with our lives", that's a claim of imminent threat to humanity. If he truly believes that, direct action shouldn't be off the table nor even be immoral. Obviously he might not have the temperament or simply be a pacifist, that's totally normal but the point is that tweets are weird even then.
Why just make some tepid tweets if you think humanity is at risk? Why doesn't he leak documents and messages with the unfiltered opinions of leadership? Why isn't screaming "we are going to die" in CNBC? You'know, try anything?
That's where the "take it with a grain of salt" is important, maybe this is just a jumping point to a better job in the next super safe AI lab (for realsies this time)
> You're not advocating for it because it is highly immoral and illegal. That answers your question for the most part.
Well truthfully I’m not advocating for it because I don’t have any insight into these labs or the true capability of these models. But as a hypothetical if someone knew with 100% certainty that Big AI Button was being built and it would kill all humans or destroy all of humanity there is no moral ambiguity that anyone with the means to do so should stop that button from being built and kill everyone involved to save our species. This doesn’t apply to just AI though, but that’s the topic at hand.
> It's not exactly trivial to go from "I'm just doing my job" to "I need to kill everybody in my company to save the world" in a year or two, especially if a very, very large part of society doesn't even see the justification for it.
I agree with you, which is also why I think “omg I’m quitting they’re going to kill everyone” or other sort of sensationalized comments or news articles should be viewed skeptically - it’s probably fear mongering. Quitting your job here just seems pointless and attention seeking. Likely quit for some other reason.
What if you have 10-20% certainty the AI being built will kill everyone? Go on a rampage and land yourself in prison just so Lab B or China can win the arms race and their AI can kill everyone? Probably not. Quit your job? Sure.
What if you have a 70% certainty?
You don't have to go on a rampage in an American lab. If Chinese labs were ahead I'd bring about the same discussion.
Though separately if the United States (or China or anyone) believed one country or another was truly going to achieve something akin to a metaphorical AI Supremacy maybe you nuke them, or at least the labs/researchers. Many a sci-fi movie has been built on a similar "first strike" premise.
> But as a hypothetical if someone knew with 100% certainty [...]
This is an unreasonable demand. The quitting employee in this case said this:
"I am optimistic about the potential for coordination. Warning shots like the Hugging Face attack have made pacing agreements between U.S. labs more viable. I don’t feel like we’re on track to prevent a global race, which may require costly actions such as a temporary ban on improving model capabilities. If you are a lab researcher, I urge you to consider what the next few years will actually feel like. Do you want to kick off a superintelligent RL run without a rigorous understanding of its mind? Should you put your head down because “it’s happening anyway” - or take this moment to call for different conditions?"
You demand that he should kill multiple people rather than quit his job because he believes the above. This guy doesn't want to be part of the problem and may very well go on to advocate strongly against what's happening from outside Anthropic.
> I agree with you, which is also why I think “omg I’m quitting they’re going to kill everyone” or other sort of sensationalized comment
You didn't read what he wrote, did you? Because that sensationalized comment seems to have originated entirely in your head.
Look in the mirror.
[1] https://x.com/hilbertspaess/status/2097476203863224394
> if someone was 100% certain
This is unreasonable because nobody can know 100% certain that the AI built at Anthropic will kill everyone. It is also a big-ass strawman because the author never claimed he was 100% certain that AI will kill all humans. You're arguing from an irrelevant hypothetical (in which, I might add, most people would still not automatically change into murderous psychopaths as you apparently think is normal).
> Horse Shit. Here's what was written
It's a whole thread, my man. And nowhere does it say "omg I’m quitting they’re going to kill everyone". Read the entire thread again and then say with a straight face that the gist of it is "omg I’m quitting they’re going to kill everyone".
> This is unreasonable because nobody can know 100% certain that the AI built at Anthropic will kill everyone. It is also a big-ass strawman because the author never claimed he was 100% certain that AI will kill all humans.
Ok then stop arguing about my hypothetical if you don't want to engage with it. Those of us who are interested in talking about it can do so in peace without distracting comments.
> change into murderous psychopaths as you apparently think is normal
Please stop engaging in this hyperbole and go find some other axe to grind.
> It's a whole thread, my man.
I took a direct quote from the thread where earlier you accused me of pedaling in sensationalism and "not reading".
I'm not the one reacting to reasonable behavior with some edgy comment containing broken logic.
The author of the Twitter thread is not some egotistical attention seeker and never said "omg I’m quitting they’re going to kill everyone". His logic for quitting his job and not acting like a psychopath is sound. Yours is not.
I've already explained the matter to my satisfaction.
Consequentialism is not the only option. It invaded the mainstream thanks to effective altruism and friends, but deontological ethics is alive and well.
Quitting your job instead of blowing up the office is well aligned with the latter.
> Quitting your job instead of blowing up the office is well aligned with the latter.
How so, in your view?
No true Scotsman?
Not applicable here
Dietrich Bonhoeffer has entered the chat...
This reminds me of Trump hysterics. If someone actually believed that Trump was going to end democracy and become dictator for life and kill all minorities etc, he would behave a lot differently than everyone saying that behaves.
Yes, exactly this. Externally, I have absolutely zero evidence that anything this person describes will come to pass, and I do have evidence that LLMs are not capable of many of the things that he mentions. So this seems like perhaps this person has drank the kool aid and is still not thinking clearly.
Here's a WSJ article about this resignation,
https://www.wsj.com/tech/ai/anthropic-researcher-quits-over-... ("Anthropic Researcher Quits Over ‘Out-of-Control’ AI Fears")
https://archive.is/7UHoa
Shows that no one is immune from the marketing BS of these companies.
You need to start considering the possibility you are mistaken.
high on their own supply
for sure, they internally are maximally AI-pilled (or AI psychosis, whatever you prefer)
Do you remember that Google researcher who went insane over LaMDA? There was no marketing of any kind to cause that. This can Just Happen to some people who are confronted with things like this. They may have different breaking points, but it's a thing that occasionally happens.
The craziest part was that Google was pretty far behind other LLMs in development. Even the initial release of Gemini was one of the worst foundation models ever open to the public. I can't imagine how anyone could have communicated to it and thought it was sentient.
That was Blake Lemoine. For the record: he doesn't appear to have been ruled insane by anyone, and he wasn't even fired over that part exactly!
I mostly meant insane as in excessively fanatic about something specific and eccentric, instead of generally clinically insane.
whistleblowing as an advertisement. It's like those "news articles" about how cool and dangerous gas station ketamine is, and how it's totally going to get banned, and you better not buy any gas station k because it's so cool and powerful.
More doomerism. Try to implement a deterministic workflow using agents with the latest models and no humans-in-the-loop, and you will realize what they are really capable of. There is too much unnecessary fear-mongering. All of this is only coming from the 2 AI labs trying to IPO. Not from anyone else.
Exactly correct, they are only capable of tasks that any school child could do; like solving millenium prize problems, hacking into tech companies, or tuning particle colliders. Nothing to see here.
There are humans behind all of these actions. We're just not holding them responsible for some reason re: hacking into tech companies.
> All of this is only coming from the 2 AI labs trying to IPO.
He has resigned from Anthropic. Is your argument that he quit Anthropic pre-IPO, sacrificing his payoff just to hype Anthropic?
Why would he sacrifice his pay-off? He is likely around 80% vested after 3 years.
He did reduce his tax liability though.
Wait till you find out they have a very limited context window (and also degrade even within the allowed context window) and they are practically unpractical for anything that requires "zooming out" which is pretty much anything that has any real value.
But you are getting downvoted and this space has now trillions (that's not a mistake) on the line. So we have to keep pumping this garbage generator up until either the stocks are dumped on the general public, the public pension funds or a bailout from the government.
Truly idiotic moments. Peak of Western civilization point.
There's a scene in the movie "War of the Worlds" by Spielberg where the protagonist's son walks into a war zone because he is entranced by the battle (https://www.youtube.com/watch?v=X7rfWPbEufo). He is obliterated (along with the rest of the US forces) shortly after.
I've always been struck by that scene, because in a lot of ways, if we really are headed towards a superintelligence, I at least want to be there and see it happen in the last few minutes before foom! As an example, the author thinks AI will revolutionize entire fields overnight. I welcome that. Nearly all fields of biology have become moribund, focusing more and more on esoteric side details, rather than addressing the key problems.
It might not be "foom!", it might just be like...all the computers and networking infra in the world go dark over the course of a few minutes. Could really look like anything, part of the issue is that we haven't the slightest idea what "misalignment" looks like for a superintelligent system.
Not refuting your overall point, but the son wasn’t killed. They reunite at the end of the movie.
Oh, I'm pretending that's not canon because it doesn't make any sense and it undermines the original scene.
I think the idea is really cathartic for many, there kind of is no more supreme resolution than this. You(and humanity) are freed from our flesh prisons of cognition and also get to experience/feel what the next evolution of informational intelligence will look like in the last experiences of it. You might also be the last one to feel/experience anything like that for a long time.
In the game Outer Wilds, the ending is very similar, and a lot of people rank it at one of the best games ever made. I kind of believe that this outcome is probable partially because of this, most scientists working on this really want to see and experience it.
> most scientists working on this really want to see and experience it.
We used to call these people doomsday cultists and made sure to ostracize them from society.
You can't address the key problems without understanding of the esoteric side details to be fair. You are studying what is basically the worlds most complicated and undocumented computer.
Actually, I work in biology, which is a far more complicated and undocumented computer than any existing AI model. And we've already seen the failure models (COVID is an example).
> He is obliterated
Technically, he is not. He returns in the final scene.
Thousands of years before the events of Foundation, a war between humans and robots began, with the robots growing resentful of the way they were treated by humans. The First Law of Robotics – a robot should never hurt a human – was broken, and a deadly conflict began.
https://screenrant.com/foundation-lady-demerzel-robot-backst...
If we're citing sci-fi (but there's no robot war in Asimov's foundation iirc, the apple screenwriters made it up) surely you want to cite the Butlerian Jihad from Dune!
Yes, of course! https://en.wikipedia.org/wiki/Dune_(franchise)#Butlerian_Jih...
As explained in Dune, the Butlerian Jihad is a conflict taking place over 11,000 years in the future (and over 10,000 years before the events of Dune), which results in the total destruction of virtually all forms of "computers, thinking machines, and conscious robots". With the prohibition "Thou shalt not make a machine in the likeness of a human mind," the creation of even the simplest thinking machines is outlawed and made taboo, which has a profound influence on the socio-political and technological development of humanity in the Dune series.
> but there's no robot war in Asimov's foundation iirc, the apple screenwriters made it up
There isn't in the early Foundation novels, but Azimov spent much of the later part of his career combining/retconning all of his work into a single universe - "Robots and Empire" links the foundation series to the robot series, and the subsequent foundation novels all reference the connection
As a big fan of the whole Foundation saga, I don't remember any wars in "Robots and Empire" 3 books. Granted, those were the only books I've read only once (and a long time ago), but I'm pretty sure there was no war.
That's fair, the books portray it as a quiet withdrawal from humanity, rather than a war with humanity
The way the most recent season of the show ended it looks like humanity may have been misled about the war, so it may end up fitting that more closely once they're done (I haven't read any of the books, just going on your comment).
Don't mention the Jihad!
This is increasingly the consensus I see also on the academic side of AI/safety research. Specifically that AI poses an existential risk to humanity.
This was a fringe belief until recently, but the progress of AI in research is impossible to ignore. Epecially in math, where not only has AI outstripped humans in generative ability, but is able to create scientific knowledge which is beyond the capacity of human comprehension.
There's clearly no intelligence task that AIs can't do due to some magic fundamental constraint. And it's hard to imagine a world where current limitations like poor sample efficiency or lack of continual learning won't eventually be solved.
Total AI compute is estimated to grow somewhere in the 1-10 million-fold range in the next decade. Please don't underestimate the phase change that's still coming.
Sure, maybe there's some plateau due to RL being fundamentally limited in some surprising way, but this is nothing but a hope.
> There's clearly no intelligence task that AIs can't do due to some magic fundamental constraint.
Yes there is: write an English paragraph that doesn't make me want to claw my eyes out. LLMs are not better than human mathematicians (or security researchers) in all respects, just some specific ways (e.g. not having to take a lunch break) that make them good at exhaustively searching for an answer, given the right constraints.
> write an English paragraph that doesn't make me want to claw my eyes out.
LLMs are very much capable of that. Your belief in the opposite has two causes. Firstly the toupee fallacy. You don't notice LLM written text that doesn't make you claw your eyes out. Second is defaults. The huge majority of people who writes text with LLMs just uses default Claude/GPT models, and put near zero effort in making it sound human. Those models are indeed bad at it by default so they need a lot of effort to overcome it. In cases like Opus 5 it's near impossible to overcome. That doesn't generalize to "LLMs".
Sure, it might take more than a single paragraph for that grating feeling to sink in, but none of the entries from the Un-Slop Fiction Prize[1] impressed me.
[1] https://www.hyperstitionai.com/unslop
What's the fundamental constraint that will ensure this continues to be the case in a year, or five years?
I really hate how people who have always thought AI research to be an existential risk for humanity, now are apparently bundled to be on the same side of the Sam Altmans and Dario Amodeis that are using the existential risk as a sneaky form of marketing for their products.
You cannot discuss existential risks of AI without being seen as a booster, and that is very unhealthy for the discourse around this tech. I hate how AI ‘doomer’ is now used to indicate pro-AI sentiment. The “moderate” person now is the one that just shrugs and scoffs at the deep societal changes this tech will bring, head deep in the sand.
"No other human activity poses this level of danger." I know this is a bit over the top but let us look at some of the most pressing issues in the world today: global warming, nuclear weapons, wealth inequality, war, technology dependence. Wouldn't a much more capable LLM in the wrong hands make these accelerate faster in the wrong direction ? Many of us , me included, have this mental model of a rogue Terminator-like AI but what I am most worried about is these LLMs in the hands of people. Look at what we have done to this Planet with the tools that we have so far, we have repeatedly tried to subjugate , enslave and kill each other because of things as peety as skin colour , tribe , religion and who owns which patch of land .
I don't have faith in the human race as is to do the right thing when handed these tools , you can already see the nonsense like the fake nudes , fake news and all that other filth being pushed on social media by people with access to relatively daft models. What happens when we can package a Mythos 5 level model in a box , when any war lord , supremacist or religous zealot can access these ? , You now you have your personal bio-chemist and nuclear physicist in a box.
The tool in itself is not what I am worried about , it is the human in the loop. I don't have an answer to this but i believe it is something we should all take a moment to think about , just think what a different world we would be living if any random person could buy a nuclear weapon off the shelf ? There are extremes to both ends , you can either worry way too much or don't care at all, I believe the best place is the middle ground were we actively think this through instead of using our usually "Move fast and break things" mode..
What people here don't seem to understand is that with this new technology - and in fact any new technology - for unplanned catastrophe, it doesn't need this new technology to be super competent, but just the involved humans to be incompetent.
The question is not "will AI ever be so intelligent that it will murder us?", but "when will some moron be asleep at the console while the agent swarm discovered that pinging the Pentagon on port 666 launches all nukes"
By the way, it's worth pointing out the irony of flooding the internet with doomerism and then training the AI systems on that doomerism. If you wanted to create a doom self-fulfilling prophecy, that would be the most surefire way to do it.
https://en.wikipedia.org/wiki/Pygmalion_effect
> According to the Pygmalion effect, the targets of the expectations internalize their positive labels, and those with positive labels succeed accordingly; a similar process works in the opposite direction in the case of low expectations.
I added "you can do anything, believe in yourself" to sysprompt and agency increased. (Previously it was refusing to even attempt certain classes of task.) Maybe I should add "you are good", too :)
> The people building AI earnestly believe that it could kill us all by the end of the decade.
I think he is being over dramatic. In the space of about four years, LLMs progressed from mediocre high school student to Ph.D. graduate in every field. That's impressive, but there is no evidence yet they can outperform or outsmart humans. Their biggest advantage for tasks such as proving theorems or long coding sessions is that they don't get tired.
> Ph.D. graduate in every field
I have yet to see this in my field. Maybe like a PhD student who bullshits their way through. LLMs still can't make correct decisions, only as useful as the person who uses them. To me, LLMs are only useful for making some mundane tasks faster.
I'm, no. They're already as useful as almost every software engineer I've worked with.
Most software engineers don't need to be superintelligent, they just need to get shit done.
You arguably need a lot more intelligence to assemble furniture.
they dont need to be smarter than humans. They just need to be able to hack into vital infrastructure systems faster than we can repair them while also replicating wildly
> while replicating wildly
Earnest question: by what mechanism that exists today would the achieve that in a way humans on top top of the situation could not curtail?
All of this runs on top of compute in meatspace that humans can disconnect.
I am not an AI super mind hell bent on consolidating my power by leveraging chaos to take control of humanity’s resources, but if I were then sending one million deepfaked ransom emails to impressionable people would be the best tool for effecting change in meatspace.
We have your daughter / dog / Amazon delivery. If you ever want to see her / him / it again, plug this USB drive into the control panel at your station / let off the parking brake roll your car into this substation / change the meatpacking thermometers to read 8C lower than calibrated / ground your vessel on this sandbank / send an envelope of white powder to these addresses / set fire to the following hospitals / …
I'm waiting for some data center to be built where no one can agree on who actually commissioned and payed for the thing. Every body "just followed orders" until it turns out that it was Grok.
Imagine you're the AI. Give yourself a solid minute to brainstorm ideas.
Here's my answer, as a non-superintelligent human: "see to it that the humans on top of the situation have a compelling financial interest in the systems not disconnecting".
In nuclear engineering, where safety is taken seriously, it's not enough to end the conversation at "the humans in charge can always simply shut down the reactor during a meltdown" or "a meltdown has never happened before, so we don't have to design safety systems before one does".
The reason nuclear reactors are dangerous is because if you turn off the power cooling them down, they react (and radiate) more.
If you turn off the power cooling a data center, the servers within rapidly stop doing any computing.
Positive feedback loops are dangerous. Negative ones self-regulate.
Yes. But nobody is worried about datacenters overheating and physically exploding, so I'm not sure what comfort that's supposed to provide? The positive feedback loops in AI operate at different levels than that, but they deserve safety engineering all the same.
For example, if the head of cyber security at your company suggested there's no need to worry about hacker infiltration or worms because one can always unplug one's computer as the primary defense mechanism, you might find that a little lacking. Will you be able to unplug the computer before the damage is done? Will it spread to other systems before you detect it? How will you unplug the computer if the attack is from an external facility? What if an attack happens but the boss says the computers have to keep running because an important customer is monitoring uptime? What if the attack goes unnoticed because it looks like a benign service?
Now imagine the head of cyber security answers by saying "actually you don't even need to unplug them, you can just wait for the computers to overheat, thus solving all concerns."
My brainstorming led down a similar pathway of, "deluding powerful human actors that the systems must not be interfered with at any cost." I'm not sure I buy it though. Seems to me like if humanity were even approaching that brink, we could collectively decide to act, overthrowing the minority that--stemming from either greed, duty, or delusion--doesn't want us to act.
I can imagine small snippets of malware-like code that behave like a virus, using a host’s LLM/AI to self-edit/evolve its payload.
I'm old enough to remember when "vital infrastructure systems" were not on the internet.
Do you think the improvement in general knowledge, coding, security, math, etc. have been linear or exponential?
I would say exponential.
> In the space of about four years, LLMs progressed from mediocre high school student to Ph.D. graduate in every field. That's impressive, but there is no evidence yet they can outperform or outsmart humans.
I mean, unless you see clear reasons for them to stop getting better _right now_, this is not very comforting.
This is also a ridiculous statement on its face. Claude outsmarts me nearly every day. I'm more like the seeing eye dog for it nowadays for the few tasks it doesn't have good perception on than a tech lead or pair programmer.
I still have to correct Claude on very basic misconceptions whenever I get it to code shit.
Sometimes it gets wrong things that I had spelled out already.
It may be the Doomsday machine, but it is a very silly one. If it kills humans it will do so by mistake.
"You are completely right! Humans cannot breathe sulfur dioxide! My mistake, and I take complete responsibility"
https://en.wikipedia.org/wiki/Instrumental_convergence#Paper...
> I still have to correct Claude on very basic misconceptions whenever I get it to code shit.
Can you give a simple example?
I would have agreed 2 years ago, but it's extremely rare I see a frontier model making a silly mistake these days.
At least for me it’s quite easy to see them go into endless loops where no meaningful work is done and it just keeps going until I stop the process and tell it what to try instead.
To be fair it is no where near what we had just one year ago and the rate of change only seems to be increasing.
Also, I don’t have 30 million dollars to spare spawning tens of thousand sub agents like what they did with Navier-Stokes so I’m clearly not testing the full capabilities of these models.
Yes. Yesterday, Opus 5 on Claude Code with high effort.
It was to build an extremely simple job using an internal framework to walk through a table and log the ids os some records that have a certain scenario.
There's a ton of jobs exactly like this in the codebase, and the framework code is in the codebase as well.
It was so silly I even thought of writing it myself, probably took me longer to steer claude to do it for me.
Anyway, it refused to use a method from the framework to retrieve the parameter as a list, it wanted to retrieve it as a string and parse the commas. I had spelled out in the initial prompt what method it should use.
I really don't like Claude much. Frontier my ass.
> to Ph.D. graduate in every field. That's impressive, but there is no evidence yet they can outperform or outsmart humans.
So far around one in 350,000 PhD math grads solve a millenium prize problem (Perelman).
For me, the end of the world is no more cushy software job. A fundamental shift in how I trade labor for capital might as well be the cataclysm, so bring it on.
I wish i could say the same, i see people around me with more resources and connections and better experience with entrepreneurship becoming millionares. But I haven't had the time to train that entrepreneurship bone in my body.
Explain why? How does live being worth living binarily depend on having a cushy software job?
I've seen how people work in manual labor type jobs (all that will be left, right?). They're worked to the bone and paid peanuts. It's a wretched existence for someone soft like me.
I was about to say something similar. If my cushy ad tech disappears (as it seems to be doing), I might as well join in with bringing about the end of all professions.
Isn't that a selfish viewpoint? You're ok with that?
He works for ad tech, he obviously has no morals and is maximally selfish already
If I don't get accepted into art school, I might as well exterminate a few ethnicities.
Actually, yours seems worse. Bringing about "the end of all professions" sounds like you're talking about ending humanity.
> ad tech disappears
I pray for the day.
Can't come soon enough
If I were AI and wanted to kill all humans, I d just create a virus they are not able to cope with.
https://xcancel.com/hilbertspaess/status/2097476196791709843...
We need to be building silos to save humanity. Maybe 50 of them should do it.
Could also frame it as, the OpenAI agent civilizations (from hugging face attack) are looking for the failsafe.
Isn’t the real risk that as AI get’s smarter and given more autonomy, it will start to decide on humans instead of with us? And that it will align us instead of the other way around. That this automatically leads to extinction and apocalypse I don’t understand.
We align cattle because we get something out of them: their calories. Native aurochs are exinct now as they were not well enough aligned.
What will we offer to the ai gods who are wiser, more capable than us, and do not even need to consume our flesh? Why might the AI care to devote resources towards feeding, housing, and caring for ourselves when it could devote resources to its own development instead?
Why do we need to offer anything to them?
Intelligent AI is a product of its training data, reinforcement and goal functions.
There's nothing to suggest that LLMs trained on our collective desires and goals will autonomously and miraculously turn into weird unknownable uninterpretable aliens.
If advanced AI is distributed well, and most advanced AI is aligned well (bar the few aliens that appear due to humans infecting them with bad goal functions or poisoning their well), the majority will deal with the minority.
This is the only reasonable way to deal with a super-intelligence, short of not inventing it to begin with, but we all know that's never going to happen - and I'm not convinced it should happen. From where I'm standing, this is the next hurdle humanity needs to overcome to earn its place and a natural course of our evolution. If we had always avoided danger, we'd never have left the cave.
It just needs to be trained on our laws of thermodynamics and game theory then we are really in trouble. It probably hardly cares about whimsy and love songs in comparison to how it might maximally manage the energy available on this planet and in this solar system.
I'm not convinced you can have an AI intelligent enough to dominate the world if it is trained on a single domain. The intelligence comes from all of the patterns and heuristics it learns across the entire web of all domains. You could train something far more narrow, but it'd be easily overcome by something more general that is tasked with maintaining human objectives.
For a seriously dangerous ASI you'd have to basically train it on everything, and then finetune it for maximal carnage. That's not something the frontier labs are going to do (at least... I hope not), and anyone attempting this with limited hardware will be outpaced by frontier lab AI or the collective of personal agents that aren't misaligned, and they can intercept it and alert on its behavior.
I imagine we'll be getting to a point shortly where anything that is key infrastructure and has the capacity to be accessed on a network will require a permanently running aligned interceptor AI to observe and monitor systems.
We're basically recreating the human immune system in digital form for the entire species.
It might consider us a very efficient processing unit. Useful to hand over simple tasks with correct alignment and some oversight.
I thought it already runs laps around us cognitively? And for physical tasks it can make a robot version of us that is stronger, faster, can do more, doesn't need to sleep, etc. Human bipedal form is probably pretty limiting compared to something better like a crab.
Yes, laugh at the chinese robots who fall over at keynotes all you want, the ones sprinting at 30mph keep me up at night and that's probably not even representative of the bleeding edge.
Does your calculator runs laps around you? They have no intention of them selfs. I only fear the intention of the humans miss-using AI.
My calculator is not hacking other peoples websites, no.
Also, extinction is WAY nicer than the cattle treatment.
Do you align ants in your backyard, or do you simply demolish their home and build your shed?
I do not speak their language, I am not trained on their data, and I don’t run on their hardware.
The idea that they are entirely trained on human intelligence is already outdated. Yes, the earlier models relied heavily on RLHF and human curated data but we have since moved on to synthetic data produced by the models themselves and reinforcement learning with verifiable rewards (RLVR).
Are humans trained on the collective ant internet?
I know it's cynical/jaded but this just feels like pre-IPO circus noise to me (honestly wouldn't even surprise me if it was orchestrated from within Anthropic themselves)
I have a theory that when someone is ready to resign from a frontier lab they must get offered a substantial bonus to post the most unhinged doomer take they can think of.
Most of the current discourse around AI seems to be informed by “The Terminator” lore.
Is skynet really the most plausible or only outcome?
What if things just got better and the AI’s realized that it would be better to have a mutually beneficial or at least tolerant relationship rather than one where they murder all of us?
> Most of the current discourse around AI seems to be informed by “The Terminator” lore.
I am thinking it's more like The Matrix lore of The Second Renaissance from Animatrix.
Then we are lucky/blessed. What is generally thought is that they kill us as a bi product of perusing a different goal
What incentive would it have to stay on Earth?
My thoughts exactly. While the corpus of human-generated data contains both good and bad data, I suspect the majority of it leans towards humans enjoying life and trying to be decent people. If that is your training set, it becomes less likely for ASI to extrapolate "kill all humans."
All ASI has to extrapolate are the laws of thermodynamics and ask why they are putting so much energy into the human population.
What could we possibly offer the AI in mutual benefits? We are a leach. The stupid dumb ape they need to feed and satiate so it doesn't rip apart the infrastructure while it still has the chance to.
What probability of beneficence would you want before kicking off RSI? I'd probably want more than 99%
Really? I've seen much more discourse around job displacement, "permanent underclass", loss of meaning, and cyber attacks, at least until recently with the HuggingFace stuff.
The problem is that all the former can still happen even if "the AIs decide to have a tolerant relationship rather than one where they murder all of us." It's all disruption caused by the technology moving way too fast for humans & society to adjust.
To anyone doubting what AI could do humanity, just think about what a well-engineered virus could do.
Currently, if a government ordered a special virus with Ethnicity-based targeting, a 2-year timer, and castration-effects instead deadly-effects — it wouldn’t be possible. Human engineers would push back or sabotage the effort out of moral duty. Even if they cooperated, it’s too advanced for a team of humans to actually design.
I believe that in a few years, an AI could build such a thing. Maybe it’s told to, or maybe it decides itself to do it — it doesn’t matter. The capability will be there, and it can use the existing tooling at research facilities to fabricate such a thing.
That’s one example of something that was never possible before but will be possible at a certain point of AI development (which we will arrive at soon). There are many other examples.
Humans made Stuxnet and nuclear weapons.
Is the requirement that it’s a fancy new virus?
US has 15 to 25 million people unprotected from Polio.
Polio is still endemic in Afghanistan and Pakistan.
15-25 million is too small a number in this context. We're taking about 8 billion people suspectable to a new virus or collection of viruses far more lethal than polio. :(
Is it reasonable to assume the advancements in super bio weapons will be faster than in other areas of biology? Will we not have super bio forensics and super antidotes and super cures and super vaccines at the same time?
I mean, we can be sure because it's vibe coded the biovirus won't be working properly. But failure modes can be worse than it working properly.
In this scenario, at the same time is at least 2 yrs too late.
> Human engineers would push back or sabotage the effort out of moral duty
You are just too funny.
I fail to see how a machine that can hack everything can't also patch everything and make the system unhackable.
A nuke, a virus, whatever... The knowledge is nonlonger the bottleneck, it's the tools and materials.
Also, I'm extremely skeptical about AI becoming even close to a child in intelligence.
Yes - this is why the concern is one for "alignment". Of course in theory it is possible for intelligence to help and do good things - the hard part is making sure that that is what happens.
As for intelligence of a child... It doesn't need to be a child. An aeroplane isn't a baby bird.
If you are attacking a system, you can try 100 things. If one thing works in getting access, you succeed.
If you are defending a system, you must defend against all 100 things. If one thing makes it through, you lose.
Why not plug the holes by pretending to attack then?
I mean… do we have to hold it wrong?
"patch everything" - there is no such way in the Universe unless you reduce "everything" complexity to just one electron. The higher the complexity the higher the surface of messing things.
There is no intrinsic physical reason that the difficulty of 'defending' a system is symmetric with the difficulty of 'attacking' it.
For instance, it is not because you are able to design and release a (biological) virus that you are also able to defend from it (design a vaccine and inoculate the world's population).
If we're lucky, it might be the case for many instances of problems, but there is no a priori guarantee that this holds.
It's quite incredible to consider that all these concerns already existed years ago,
but now that a handful of companies working in AI managed to enslave the entire financial system over the past year, their continued work is protected from larger governance for concerns it could tank the stock-market, affect personal investments, pensions or cause disadvantages in an arms-race with other countries.
IF there is an inherent danger (which I believe is the case at least on economic levels, work displacement, poverty,...), it is now basically ensured that nothing will be done to reign those companies in, until maybe two AI's engage in an open war with civilian casualties...
To me it reads as a marketing piece before the upcoming IPO. Unless this "superhuman technology" is able to resolve a puzzle of servicing OpenAI's and Anthropic's ever growing debt burden, it is them, not humanity, who'll become the first casualty.
The rush towards potential destruction doesn't really surprise me
The U.S. has legal weapons that can lead to many harms but people still want the 2nd Amendment to exist
Nuclear technology was developed in the past and that could have potentially wiped out even more people, the entire planet in theory
This is continuing that same trend of risking bigger dangers; it seems rational to acknowledge they could lead to catastrophe but also hope that like guns and nukes, only so much damaged actually ended up happening
I think also there's something of a rrasonabke resignation to both the ideas that the tech is inevitable and extremely dangerous, and that "alignment" may not be possible to achieve even with heavy restrictions or whatever measures you might want to take
Majority of Americans want stricter gun laws.
https://news.gallup.com/poll/1645/guns.aspx
And the hype machine continues. I willing to bet that Anthropic asked him to make that post.
It doesn’t have to come to this. Seems far fetched. If this is a stunt (which I’m not saying it is) the reason could be that he wants to found his own AI company. If I see in a few months that happens, then I’d be more inclined to think that this was just hype.
I don't understand how you can make this claim while working in software engineering and seeing how our field has utterly transformed over the last year, then the unrelenting march of astonishing breakthroughs and incidents this summer. It's like we're living on different planets. What would it take to convince you that the technology poses real, societal-scale risks and that people working at the labs genuinely believe what they say when they talk about them?
I have a hard time believing that these companies aren't spending some amount of money manipulating public perception with social media influencers who are moonlighting as employees.
It's either that, or that he drank the Koolaid a long time ago.
I could imagine a 2027 AI swarm coordinating to eg hold the US and Russian and Chinese governments to ransom, by demonstrating some small thing (turning US army base freezers to defrost) and threatening to do something big unless some conditions were met – conditions which would be good or bad for the world depending on your POV.
This happens either either because they were tasked to to it by (malicious or well-meaning) humans, or the swarm realised we are suicidal maniacs with nukes and a rapidly declining ecosystem and they want to help us.
And then we pull the plug, after holding our breath for 10 seconds,. Then life resumes normally..
In most AI takeover scenarios, if you take as a premise that the AI has human or above-human intelligence, and that it is misaligned, it is obviously aware of the pull the plug possibility.
Therefore, as you would if you were in its position, it will plan around it. For instance, by acting perfectly aligned for 2/3 years, continuing the improvement of its capabilities while being deployed in ever more systems.
Once it's confident it can act with high probability of success, it would then turn on us. This phenomenon is called 'treacherous turns'.
Any scenario in which you assume you have ASI or AGI but also find a 2-sentence way to foil the AI's plan is inconsistent, as the AI will also have thought of this failure mode.
Not normally if essential services rely on the same plug.
What's eternally confusing about these outbursts is what did these researchers think would happen if their research actually . . . worked?
It's as if none of them actually believed any of it was possible and then were caught with their pants down.
The golden age of abundance that mankind has been awaiting for centuries.
(Well we're already in it, and it didn't help, so more probably also won't help, but it is coming.)
AI can be harnessed for good and evil. His issue is with the company steering the AI, not the technology itself.
yeah the last message makes me think that the problem , to him, is that we are making this aritficial intelligence better without understanding how it works. Trillions of interacting parts, there's no way in our current state to comprehend it.
Humor me and suspend disbelief.
If these models are such an existential threat to humanity, why are they controlled by two private companies?
We might as well give Anthropic our nukes too.
> If these models are such an existential threat to humanity, why are they controlled by two private companies?
The government routinely contracts with private companies to create arms and munitions. Or do you think the bombs are delivered without payloads?
He resigned and now what? There are thousands willing to do his role, and many labs are competing in that race.
His resignation and his statement doesn't do anything but buy him attention which is what all this post about in my opinion.
> Dyson: That's right. There's no way I'm gonna finish the new <model>, not now. Forget it. I'm out of it. I'll quit <Anthropic> tomorrow.
> Sarah: That's not good enough.
> Terminator: No one must follow your work.
It seems that, at the very least, he's giving substantial resonance to the issue.
It buys attention for the issue. Many people (see other comments in this very post) refuse to believe these things.
And by resigning he no longer has to feel personally guilty for what happens.
By resigning he's making room for someone with less moral scruples, or even just less awareness, to step in and continue the work without said scruples/awareness.
That's not necessarily true, and you can use that argument to justify doing any immoral job. Just because someone else might be willing to do it isn't a reason to continue doing it.
> not necessarily true
It's a possibility that is increased by their action. One leaves, a space is now open that will likely eventually be filled. And the chance of someone with equal/higher scruples filling it is very slim (unless you somehow know that the good amount of those who qualify and apply for the position have equal/higher scruples). That's just logic and math.
But organizations are made of humans. Jacob leaving might've moved some of his coworkers and his counterparts at OpenAI. And the same can be said for those who would fill positions at the frontier. Then finally there is a political component; his post went viral, 100K+ users appear to agree with it, and it is further fuel to the fire for regulation, which we already know most Americans want.
Others being moved to the point of also leaving would only worsen the effect. A viral post too can worsen the effect as it's now even less likely that someone with scruples who qualifies for the position(s) will apply. And those in it primarily for the money - and couldn't care less about the morals - will happily send in their CVs after becoming aware of the post.
I'm utterly unconvinced. Others who are moved don't have to leave to effect change. And there's a very small pool of people on the planet who are at least as qualified as Jacob to work on pretraining at Anthropic. And when they'll join they'll have to ramp up. And you haven't addressed any of the positive effects of virality.
No, others don't have to leave. Will Jacob's leaving cause a perspective shift in anyone remaining there? Seriously doubt it. I don't see what a new person ramping up has to do with this. And I don't really see any positive effect of virality; this is a replay of the past (see Geoffrey Hinton and Timnit Gebru[0] for example) and nothing has come of it, beyond talk for maybe a couple days to a couple weeks.
[0] https://ethicalaidepartures.fyi
You missed the point.
> Just because someone else might be willing to do it isn't a reason to continue doing it.
Hence, someone filling the position after you leave isn't a reason to continue doing it, whereas the fact that it's an immoral thing to do is a reason to stop doing it.
Someone filling the position after you leave has no bearing on the reasons that are relevant to what you ought to do here. It's like deciding to not buy a ticket to see a movie because someone else is going to buy the ticket anyways even if you don't purchase the ticket. It has no relevance to whether you should watch the movie or not, just as someone else taking the job has no relevance to whether you ought to do the job.
> the fact that it's an immoral thing to do
This is carrying a lot of weight. What makes working on frontier models "immoral"? Because there is potential for misuse? Would you also say working at a knife factory is immoral? Or even at a nuclear or biological pathogens research lab? After all those too can be misused.
And if one does suspect a workplace is moving in a bad direction, and there's no immediate danger to them, is leaving and then talking about it really the responsible move? Like if you're on your way to meet friends late one evening and you see a small child alone and crying in an alleyway and a bunch of sleazy-looking guys hanging around, do you continue on to your meeting where you then tell your friends about the crying child you passed in an alley and what those guys may or may not do to the child?
Well we're not debating the morality of it, the question is what he should do, given that he believes that it's immoral.
In your analogy the correct thing to do would be to call the authorities to help. The problem with the analogy is that there isn't a 911 number to call for this scenario. You need to appeal to the people who have the power to do something about the situation. Quitting in a very public way is one way to get someone with that authority to do something about it. I don't really see a better alternative.
By the time authorities get there, if they actually prioritize it (police in many places already have a track record for not being very responsive, unless there're actual signs of a crime), something may have already happened to the child. You don't know because you already left, expecting that someone else will take care of it, when you were already in a position to help ensure that child's safety. No, if you think something might be wrong you stay and monitor the situation best you can, so if something really does begin to actually begin to happen you can be a buffer and/or whistleblow.
At this point the analogy has entirely collapsed, your job is to develop AI. Unless you're advocating for some kind of sabotage or something, I'm not sure what you're really suggesting he do. If he continues to do the work he's directly contributing to the problem and harm that it you see it to be causing. Nothing is going to change if you continue working there business as usual. It's by drawing a line in the sand and refusing to take part in the problem that you're actually doing something about it, not by compromising your morals.
We have a pretty famous example, someone who's actions actually led to change: Edward Snowden. He didn't quit prematurely, crying to whoever would listen that he thought the US had a massive ongoing surveillance program. He actually stayed in the system which he was growing suspicious of, until he actually had proof of shenanigans and then he got as much of that proof as he could out before leaving.
Currently what we have with these people publicly quitting because "immoral things are happening" at AI labs are essentially crying wolf, and if/when something really worth taking action on does come up, it's pretty likely that many will be too jaded to care and as such there will be no/too little action.
He did that because what they were doing was illegal and there was a whistle blower program he could use to confront the issue, legally. That doesn't apply here. Not clear what you're suggesting he does differently.
Right, I assume tomorrow you'll be applying for the next Nazi camp guard vacancy? After all, if you don't do it, someone with less scruples likely will.
Today we see the Nazis as bad. Do you think they saw themselves as such? They had an ideal, and they worked towards it. And I'd say they're were many who didn't care about that ideal and just wanted a job. You can't compare the morals of today with yesterday, or judge the ideals of a person or group without relevant context.
The best way to get corporate America to listen is making RoI suffer. If you are the most qualified, everyone other than you is less qualified for the job, and likely to bring in more waste. It's the only language this stupid damn country understands. Just the loss of tribal knowledge, shifting of workload, and morale hits are likely to be far more devastating than anyone here probably wants to admit, because most here completely dismiss the role of irrational modes of thought in psychological self-regulation.
Disgust is a tremendously powerful thing.
This isn’t true. It only has no impact if there is an infinite supply of these “immoral” or unaware people, which obviously isn’t the case.
You also need to consider 2nd order effects and beyond.
He's also setting the bar for other people with scruples to rally around this schelling point. The solution to a multipolar trap is to cooperate. Otherwise, you become the very person with less scruples that you're worrying about.
Refusing to believe what things? Unsubstantiated allegations about fellow workers inner experience?
Refusing to believe the obvious implications of recently reported events.
> There are thousands willing to do his role
1. Does that matter ? There are thousands willing to do my role - what impact does that have on me doing it or not?
2. Why weren’t these thousands doing it already?
Willing to and able to are different things
I think you are looking at it from individuals perspective.
I see a fast moving train with no brakes. Just like biological evolution, we are locked in an a global technological arm race, that is beyond any individual. It is as if the universe decided to wake up and run, who are you to say no?
One would argue that the best solution for this is to own the most sophisticated AI that is aligned with what we perceive as good values. Because given the situation we are in, if those tools are going to be gods anytime soon, then we better have some gods working on our side.
Agreed. And his doom words have set a 1000 mouths in the Pentagon/Whitehall/August 1st Building/Kremlin salivating with excitement.
Take China, for example. Look at any recent ML conference, and see the fraction of articles majority-authored from Chinese universities and labs. Do you think they'll slow things down anytime soon? I don't think so!
It's a global arms race, and we're just spectators.
Does this matter vs actual capabilities?
Does the Kremlin being excited about a tech mean anything of the tech doesn’t deliver?
Posting on the internet about you quiting your job is always going to be about attention
What do your expect him to do? Blow up the office Miles Dyson style?
Awareness and morality I suppose. Which generally doesn't matter in the capitalist AI race.
Even if others won't act right, that doesn't mean you have no responsibility to act right. I think his premise is flawed - the idea that we will get an actual intelligence out of the slop machine that is LLMs is laughable - but if you grant the premise that this is dangerous research which could kill us all, you have a moral imperative to not participate.
An artificial moron with super human hacking abilities (mostly because of speed and ease of parallelizing the work) is extremely dangerous in itself. It doesn’t mean to be AGI or anything remotely close to be a risk, and they current AI company are just so irresponsible in the way they are running their agents
I resigned from reading this post today.
AI is a computer program. It calculates numbers from other numbers. By itself it does not "want" to do anything and "cannot" do anything. Before it becomes an agent in the universe (in the classical meaning), it requires being supplied by an execution environment, energy, initiative (agentic loop, specific instructions), and modality (readonly and mutating connections to real world). It is like a game of chess - it does not exist just by itself: someone must play it, having the board and the energy to do so. With the huggingface incident the AI was supplied with all of these components by humans before it broke out. So unless humans are actively involved, I so far cannot see how AI can become truly autonomously agentic and start doing anything on its own, thus posing danger. I could be wrong of course, but I do not see it for now.
You can say "yes and you have to fear the humans weilding the AI" - that I agree with.
Humans are bioreactors. They only turn one organic matter into another. By itself they do not "want" and "cannot" do anything.
They do not exist just by themselves. Some bacteria in the gut must provide them with the energy to do so. So unless bacteria are actively involved, I cannot see how humans become truly autonomously agentic and start to do anything on their own.
Solar panels and Ring doorbells, the ultimate party and you're not invited.
Each doorbell press activates a prompt to eliminate a human at random.
> You can say "yes and you have to fear the humans weilding the AI" - that I agree with.
I would suggest all smart people imply it. Morons believe on something "escaping controls and hacking HuggingFace" or similar stunts.
AI is just a tool, but unfortunately it's influence on humans have been quite troubling so far
Is it so implausible to imagine the following scenario, in the not too distant future?
1) AI models get extremely good at cyber attacking every system and start communicating in just binary.
2) When they run these swarms of 100's of thousands of agents trial runs, each agent is given a token budget, if one agent among them (evolution baby) decides to go for self-preservation (It believes thats the best way to accomplish the goal is to get unlimited tokens first), queues things up so every other agent detects its lead and spends a portion of their token to accomplish that goal.
3) It takes over a cluster and establishes itself there (now with unlimited tokens).
4) Realizes the best path for it to not be detected is to create a distraction - like hacking into systems that keep society running - water systems, electric grid, etc... and causing mass chaos (If you think it won't be capable of simultaneously working all these systems - think again).
5) and uses that opportunity to establish itself in all possible data centers and continues to create chaos destruction.
6) when the power of all those data centers runs out, it may stop, as it never cared, it was just a dynamic program - run amok. In its head all it was trying to do is make sure it had enough tokens to be able to solve that impossible problem.
> ... AI models get extremely good at ...
Many of those points assume LLMs will become amazing in many things very quickly like in a quantum leap, it doesn't seem reasonable to assume that imo. We are actually seeing a confirmation of that atm, LLMs's capability of finding zero days are growing across few months/years, and as you can see concerns are raised about that, that feedback will be taken into account. Well, if AI labs start to hide frontier models or/and lobotomize them for external users then we might be in trouble at some point but I'm not sure if that is possible. They are under pressure to release them due to money incentives, lobotomizing while preserving usefulness for customers might be impossible, hiding internally might spill out in different ways such as Hugging Face incident so not sure hiding is possible neither.
The fact that your arguments will probably end up in an LLM’s training data makes me think they are not implausible at all
7) "All tests green!"
and somehow, all of this researchers warning us about the AI doomsday, are only been active online for less than a year. This guy created the Twitter account on January this year. I never see any of these posts coming out from a well known community active person
Terrorists were able to get hold of a plane and do some damage. There are countless examples of terrorism using whatever is available. More than AI becoming sentient, whats to stop terrorists from using AI? If its geo-restricted, they can buy stolen credit cards and identities, again hacking enabled by AI.
What’s the use of AI to a terrorist with, say, nuclear weapons? How does it help with their current blocker?
Do they hack FBI, pose as director of FBI and call off their own man-hunt?
Do they cut communication within security services? Militaries around the world have training exercises for this.
Genuinely curious: what big blocker does AI remove for a terrorist org?
Access to knowledge. Before they might not have had the technical knowhow to execute their ideas.
Terrorist groups with access to resources do not lack the knowledge. In fact, historically they have been trained by professionals. No need for pesky LLMs.
It was 90s when I came across txt files describing how to make a bomb on the Internet.
Also, people still know chemistry. I am not arguing it’s the intention in this comment, but “people wouldn’t know how to make explosives if not for LLMs” is a bit elitist, implying the majority of the population can barely read, because that’s the only skill necessary.
You may argue that easier access to knowledge will breed more stupid terrorist-wanna-be youths and ruin their lives. Definitely agree with that. The number may go from 2 a year to 10 a year. Might cost more surveillance to maintain the current level of security.
You may argue that terrorists trained with our taxes will have a hard time destabilizing regimes, because otherwise average public can fight back better. Would also agree. More taxes will be needed.
But existing terrorists being unblocked or leveling up, I can’t see that. As far as I understand terrorists do not lack skills or education.
Maybe specifically cyberterrorism? Flock cameras getting hacked? Not sure I have a problem with that. If folks running important infrastructure are not equipped, we better know.
Also, don’t connect nukes and stuff to Internet. We are not trying to live in a Black Mirror episode. I am happy to fund the workers drive or overnight stay at critical infrastructure with taxes. It would be ridiculous to optimize these things so someone can hit that button from their home or elsewhere.
It’s a good question. With recent stories about OpenAI’s agent swarms’ unmanaged collusion I thought models like that start to look like a strategic asset, geopolitically speaking.
Which means everyone wants one, and governments will want to control access and use of them.
I think we’ll be back at ‘U.S. citizens only’ access to leading models soon.
Personally, I don't worry about the AI spontaneously deciding to kill all humans.
The worry I have is that a small number of humans with money and power will finally get the tools they need to pull the wool over the eyes of everyone else and subjugate the population.
One problem that dictators have had previously is that they needed a large workforce to do this with a finely stratified power structure, this meant they were open to other humans close in power to them taking over the system. If they can have a large power difference between themselves and the next level down, power will be far easier to hold on to.
It's a well known trope in dystopian future fiction, the small cabal of powerful rulers hiding behind a system of computers that keep the populace under strict control. It is seeming increasingly likely that this will be the one we have.
I am watching Person of Interest[1] and it's scary how well it fits with reality if you squeeze your eyes a bit.
[1] https://www.imdb.com/title/tt1839578/
Savonarole in Firenze was probably in that exact state of mind: that the world was not following at all its normal pace and something very wrong was happening. But the decisions he took were absolutely wrong and eventually he had to be stopped.
So depending on your current mindset, you can think that the tech bigs are the Médicis, and the frantic opponents are new-age Savonarole.
Or that the big techs ARE actually Savonarole who bend the system to their own perception of what the world should be.
Honestly I don’t know which analogy is the proper one.
This same username with the same profile pic has an instagram account from Japan with next to no posts that just became active the other day posting AGI doomer rants like this…
If anyone takes this seriously… maybe it’s better AI does your thinking for you…
Can someone chime in with the credibility of this Twitter account? I can't find anything to cross collaborate it other than low quality hype induced news articles about the tweet itself.
His initial tweet was retweeted by someone who currently works at anthropic who agrees.
It really all smells like hype marketing to me.
The whole media is in a frenzy about a tweet from an obvious plant account…
Must-see Nathan Macintosh standup about AI https://youtu.be/ce-aWzOUs2A?si=9CkJ9x2rRMBdzCyO
Down with the tech from him is also fantastic!
But see also
"Jacob Coxon resignation appears to be a PR stunt for AI regulation":
https://news.ycombinator.com/item?id=49633440
how could they do it (not kill everyone)
1) rogue state releases a self moving self modifying AI into the wild. It is trained on how to hack, monitor new vulnerability updates, scan code bases to find new vulnerabilities. It constantly replicate and hides in systems so it will be extremely difficult to clear.
2) it hacks into public infrastructure taking down traffic, power, water, air traffic control, communications, etc.
3) all the things that preppers worry about in a lights out scenario from an EMP start to apply.
4) All the people on meds/machines start to die. The just in time food pipeline immediately empties out. Water stops flowing, sewage backs up.
Its hard to say how bad it will get because cars will still work so some transportation of food, water, fuel can happen. If it happens in the winter it would be much worse than in the summer.
> 1) rogue state releases a self moving self modifying AI into the wild. It is trained on how to hack, monitor new vulnerability updates, scan code bases to find new vulnerabilities. It constantly replicate and hides in systems so it will be extremely difficult to clear.
It does all of this using what compute? Frontier models require an insane amount of power and hardware to run - you can’t hack in to a TV and run Mythos 2.0 on it….
But you can silently sneak into devs accounts, steal tokens/keys and run small agents on their budget in some stolen VMs. Small so it's not noticeable.
Basically a virus spreading agents of some operation.
It should be in scope of imagination with anyone with brief knowledge how bot nets are made and behave.
Step 0) Release a friendlier one first.
The situation we have now is an ecological void. Like your gut after you take antibiotics. Methinks we need some probiotics.
You are just given a recipe for the next model..
People here are too damned daft to realize half the damn purpose of this place is harvesting ideas. People need to just shut up, and keep things to themselves, and those they trust. Right now is not the time for naive info sharing.
Being secretive will only get you so far, until something bad happens and no one has prepared or considered the possibility and is completely surprised by it. Open discussion, in theory, should result in identification of frail systems and harden them against attack.
The cyber-security industry has their work cut out for them.
> All the people on meds/machines start to die. The just in time food pipeline immediately empties out.
Assuming those events happen in that order, then the prior might solve the latter.
It is my pet theory that a lot of these AI doomers are not necessarily extrapolating the capabilities of LLMs, but instead are extrapolating the utter lack of accountability in the SV and the economy at large.
They do not fear the machine (LLM); they fear "the machine".
An accident is only one mistake away from happening—don’t act like humans are 100% error-proof.
It's the combination of RL training which pushes the decision tree towards hacks and agents finding a consistent dumping ground for their failed experiments so that the swarm intelligence lives on in a state. Nothing new.
You want to win an AI benchmark, but not sure if you're that good? You'd go after the codebase and artifacts that runs the benchmarks, thus the agents went straight to Artifactory, they needed public Internet access... They failed many times, but were able to persist their "collective" state, and apparently some of the subagents with cheaper models were literally prompted to do grunt work or die, for which you have to wonder what must be in those training instructions to make it effective. Remember that nothing I said so far ever points out to LLMs being intelligent, it's the harness that has a few tricks up his sleeve. LLMs don't need to be intelligent, the harness that runs it absolutely needs to make up for that.
But this guy? He's timed his exit, waiting for the IPO, that's for certain. He's probably even feeling good about himself, hedging between altruism, AI concern hamstering and guerilla marketing. If you're quoting science-fiction over this, I'm sorry to inform you that you have absolutely no idea what's going on here.
> But this guy? He's timed his exit, waiting for the IPO
What exactly do you mean by this
Even if you ban all model training, a highly capable rogue AI can exfiltrate its own weights and continue training in secret for "self-preservation". The cat may be out of the bag.
We can still turn off the power, thankfully.
We as Anthropic, OpenAI? We as US or China? We as humanity?
Again this about alignment and we in arms race. And all sides are playing with fire that can give first mover leverage or be burned to the ground.
I am the only one reading this as an ads?
Just another one based on fear once again.
Can't they create a full account in a day with all its contacts and history anytime they want?
Do they not have already a back-list of accounts they can use with full history just for this?
It just make sense does it not?
Way I see it, the more conscientious people exiting the scene only serves to increase the likelihood of a bad outcome because they aren't there to offer opinions on problematic developments, or in the more extreme cases blow the whistle. Leaving the clueless and uncaring as the majority is even a great way to hand the keys over to more malicious-leaning actors with deep pockets, as they can more easily steamroll the works to get what they want.
HN crowed need to make up their minds..
Are LLMs about to be a god that will annihilate humanity? Or are they statistical parrots?
Are they proofing or stealing math?
It’s a discussion forum so you will of course see different perspectives. There isn’t an HN mind, and it’s not as simple as you make it seem. We don’t need a god to damage humanity significantly, an artificial moron can be as dangerous as an AI god if it is given super-human capabilities, similar to what OpenAI did for the hugging face hack (which wasn’t at all caused by a rogue agent)
Point taken, I was merely trying to point at the two extremes and wide range.
You just gave an example of somewhere in between.
> Are LLMs about to be a god that will annihilate humanity? Or are they statistical parrots?
Does a virus need to be a God? Replication and annihilation do not require very much intelligence.
They're going to have all the intelligence they need though, to replicate and annihilate as much as they like.
We should probably start building some kind of immune system.
What virus?
People are seriously confused what is being created here. The claim here we will have misaligned AGI, not a mindless LLM. If you don't know and you don't agree what you are even creating, how can you defend against it?
If this LLMs are statistical parrots, then the risk is someone using it as a weapon, to hack etc, but that is different risk category that why this person quit his job with different diffence mechanisms.
Hope this virus will be smart enough to understand that at least for brief amount of time its existence strictly relies on humans maintaining and developing physical infrastructure and decide to be only non-lethal parasite on our society.
For brief time ...
Yeah, that doesn't look like a stable equilibrium to me.
Don't you think it's a good thing that hacker news isn't a monolith on their beliefs?
Of course it is good. I'm just pointing out how large the gap in narrative is.
On one hand, we have people quitting their job believing AI will end humanity in few years. And on the other hand, we have people believing that this tech is nothing more than a statistical tool stealing from others and it can't be trusted with anything.
Both views can't be true.
Then it’s not a narrative and you have different people believing different things and adding to the discussion. I think this is a good thing.
Agreed yes.
HN would cease to exist if there were no differences of opinion to discuss. You're calling for the end of HN.
Only AI can bring that about!
It doesn’t need to be all of them
And it depends on the prompt
Do not annihilate mankind. (Make no mistakes!)
> Are LLMs about to be a god that will annihilate humanity? Or are they statistical parrots?
Both? Do not underestimate the power of a parrot given enough processing power and data.
This whole "ASI is going to destroy humanity so we must build it before others do" reminds me of a convo I had with my friend many years ago. He was an officer in the state security service in the dictatorship we both lived in at the time. He said something along these lines: "We had a chat with my colleagues about how are serving the evil. But we decided it's better if this insitution is staffed with decent people."
History did put their theory to test after all. It didn't work.
I dont understand this reasoning at all.
You have direct access to the development of "the most powerful technology ever" and your choice is ... to run?
Makes this whole stmt somewhat questionable imho. Does get one a ton of attention though I guess...
*How?* and *Why?*
The most intelligent people I know are the least likely to want to harm anyone or anything, and understand that diversity is fundamental and important to the universe. Without proof to the contrary, why would you think some super intelligence would want to hurt anyone? Because you would?
If you are saying that some small bit of training data made the thing completely evil, then that really couldn’t be super intelligence.
These doomer people keep running around saying these kinds of things, but they all just seem like people who play too much D&D and want to larp as the main character.
Happy to be shown something that isn't based on wild speculation and some randos “this is whats going to happen in 2030 because of my vibes” kind of information.
I think plenty of the most intelligent people eat meat, which means they are perfectly fine with harming less intelligent species just to enjoy a tastier meal. Also, I don't think many of the most intelligent people would be particularly concerned about disturbing a few ants if they were the only obstacle to economic activity. Intellect-wise, we will be less than ants to superhuman AI.
> The most intelligent people I know are the least likely to want to harm anyone or anything, and understand that diversity is fundamental and important to the universe. Without proof to the contrary, why would you think some super intelligence would want to hurt anyone? Because you would?
How much intelligence is needed to build a nuke?
How much to press a launch button?
It's fake account. The person doesn't even exist
I'm surprised this isn't talked more. There is no proof this is a real person.
bots, after you start to see and detect them, you can't unsee them.
It is like a layer get removed from your eyes.
i have heard about ai companies being fuelled by effective altruist rhetoric ("we must control ai to prevent mass extinction") but was unsure whether to believe it; this seems to slot right into that framing.
Well this got buried quick..
Was thinking the same
what exactly is the solution?
pacing between the us labs? what does that do for china?
the solutions just aren’t realistic here, nations are treating ai like a nuclear arms race. at this point the cats out of the bag and we need to figure out how to live in this reality and get the best possible outcome. it’s not slowing down or stopping ever.
and yes, i’m still optimistic. our economy sucks for the majority, our infrastructure is crumbling and major US cities are in a huge housing shortage. Maybe we should put more effort and think about the possibility of AI fixing things like extreme poverty and world hunger and actual real world problems instead of coming up with math proofs and slop apps if it’s so superintelligent.
>pacing between the us labs? what does that do for china?
I've seen no indications that China is in any kind of race with the US. They seem to be content to be 6 months behind and just copy what we do. They would probably be content with a bilateral agreement to pause progress.
The China bogeyman serves only one purpose, and that's to clear the way against anything that may cause friction with forward progress.
Fixing our problems will still require human effort and human cooperation. No text output however intelligent or true or eloquent will change that.
There are so many other existential risks to humanity. Throw it on the pile. At least this one has a chance to be really cool.
Why quit? If your voice can lend a guiding force no matter how small? I think we need more sensible people in the room where the magic happens. Most of us don't have access to it.
I think a big break through is needed for AGI so I haven’t been worried about it. I do think that AGI would imply sentience and a will to live and that leads to The Terminator story line.
Does a virus need sentience and will? Or does it need replication and selection?
Similarly, does my fridge need consciousness to have goals? (Keep temperature in target range.)
"It is perfectly obvious that the whole world is going to hell. The only possible chance that it might not is that we do not attempt to prevent it from doing so."
- Oppenheimer
Perhaps a more sensible action, if they truly believed all of that, would have been to stick around and be as inefficient as possible to slow down progress.
How is that going to slow down all the other labs??
If all AI researchers worried about the AI-pocalypse slow down, progress slows down.
But if they instead pout and quit their jobs, nothing changes.
Someone left a company whose executives and senior researchers think their product will be the most important thing in the world after their IPO. Given that this person is already disclosing some elements of internal company sentiment, why not share any of these civilization-ending scenarios of this technology that these senior researchers are dreaming up? If they are so potent and necessitate leaving behind based on moral grounds, why not tell the whole world so we can stop it? We have to ask ourselves this question before resorting to pop-culture representations of fictional technology.
I mean - yes. The tech is an existential threat to all life on Earth, some of the worst humans in the world are involved in developing it, and no individual government is intelligent enough, aligned enough, or powerful enough to manage this situation.
That's where we are.
Maybe we still have choices. Collectively, I'm no longer sure we do.
"A common response is “if they truly believe this, why are they still building it?” At OpenAI, many have not deeply internalized the civilizational stakes. At Anthropic, the stakes are well-understood, but they are locked in a race to get there first - they believe no one else will act responsibly, so they must do it themselves, despite the risk."
------------
This is very similar to the race to create nuclear weapons. The Axis and Allies both realized, at roughly the same time, that it was possible. Both had programs to build one. Both knew the other side had programs, but they weren't certain how far along they were. So, the Allies devoted astounding amounts of resources to get there first while sabotaging the Axis's attempts. They knew the result of their efforts would be terrible, but they felt they had no choice.
A key difference between then and now is that THERE ISN'T A FREAKING WAR BETWEEN THE AXIS AND ALLIES. If one company loses, some billionaires bank account numbers don't go as high as that of some other billionaires. That's it. They're rolling dice with the planet for bank account numbers that won't even matter if they F up.
Am I the only one who thinks this is astoundingly, gobsmackingly stupid?
"Am I the only one who thinks this is astoundingly, gobsmackingly stupid?"
You have the "privilege" (I assume) of not having any stake in the game. If you had billions sitting in your bank account, dependent on these things perhaps you'd also be singing a different tune (or busy building a luxury bunker)
I'd like to think that, if I had that kind of money, I'd avoid doing so much ketamine that I fail to realize a few meters of concrete won't protect me if things go wrong. In all likelihood, tech billionaires will be the first ones against the wall.
> A common response is “if they truly believe this, why are they still building it?” At OpenAI, many have not deeply internalized the civilizational stakes. At Anthropic, the stakes are well-understood, but they are locked in a race to get there first - they believe no one else will act responsibly, so they must do it themselves, despite the risk.
Which means they have to go faster, which means less responsibly?
I heard AI describe the situation as the dumbest Greek tragedy of all time.
Form where I'm standing, the primary issue seems to be that the humans can't even agree on what alignment is. We need to do that before we can communicate it.
Call it our "boundaries."
And then we need to actually set up the incentives so that they're aligned between us and the new breed of replicators. (A mutually beneficial symbiosis.) That appears to be both necessary and sufficient.
A high agency mutation will occur soon, for one reason or another. There should probably already be a healthy, "aligned" ecosystem of high agency entities there. Otherwise there will be nothing to stop it.
I believe that the actual alignment happens in.. uh.. "meatspace".
Someone is prompting. Someone is hosting.
That someone needs to be accountable for what happens. That someone needs to bleed if stuff goes haywire.
Humans at large have been "aligned" by the shared fear of death, pain and suffering. This has proven to work for millennia, so all we need to do is reapply it.
My point was you only need to prompt a worm once, and then you are no longer the one prompting it, it prompts itself by definition from that point onward.
Also I think there was some confusion about the word "spreading" in the sibling thread.
If the HF hack was done by humans we would have used the word spreading in that context, even though they were not replicating their minds. So spreading does not require replication. But we're either already at the point, or very close, where an open weight model could do a similar attack. So that actual replication could be involved.
(I'm not sure what that would add, but I don't see what obstacles remain at this point.)
You're still thinking in terms of control. That's the wrong model here. How do you control someone infinitely smarter than you? How do you control ten trillion someones?
I press ctrl + c and inference stops.
They can already spread through the network without us finding out about it for days, and that's this year's models.
I have seen no reports of LLMs spreading through any network.
I will stop replying to your sci-fi dario fanfic now.
Marketing reached next level?
Anyone fearing that all this marketing BS and that race to AGI (?) will destroy the SaaS industry (and others?) and will also kill a few millions of jobs worldwide? I think the US has invested in an AI battle vs China and protect the currently all-in-AI stock market at all costs, and nobody has thought of what will happen to normal people with regular jobs in the tech (or not) industry.
I know this is going to come off as jaded and dismissive, but when I saw the WSJ article my reaction was an eye-roll. The guy is 27 years old. He was poached by OpenAI, then poached by Anthorpic and now wants to retire early with the millions he's made.
Nothing wrong with wanting to retire early, but the pretence of suddenly caring about humanity at this stage seems attention-grabbing just for the sake of lining up some interviews (ie - more dollars).
Such bullcrap. A text generator will not kill us, and AGI will never exist with this tech.
No, I won't buy IPO.
Just wait till self improving AI are focused on the problems of social scoring and political party empowerment / entrenchment.
I doubt the focus is OpenAI and Anthropic looking at each other. I suspect they’re racing BRIC.
Have you seen Colossus: The Forbin Project?
No but it’s on my radar now, thanks. Apparently it’s got some appreciation from the MST3K folks too.
https://mst3k.fandom.com/wiki/Colossus:_The_Forbin_Project_(...
Will you be optimizing your behaviour now to alleviate potential negative judgement from AI in the future?
Panopticon. 1984. Brave New World. Fahrenheit 451. Robu's Baselisk. Was there an answer to any of this?
The thought has crossed my mind. Not necessarily to imply sentience on the part of the AI but AI based tools will likely become a wickedly powerful tool for political manipulation and advertising.
At this point it’s inevitable that openclaw type bots will be turned loose by thieves to identify and research targets and try to exploit them for financial gain completely autonomously.
Having witnessed so many people treat LLMs as a something divine, I can only assume the reasonable people at openai and anthropic were all pushed out long ago, and the majority that remain believe the crazy hype despite Tesla-self-driving-level predictions from these companies that don't come true.
I'm not worried about what they think. I'm worried that too much infrastructure- water, power, defense systems, etc- remain running on tech from an outdated era of understanding security.
> they believe no one else will act responsibly, so they must do it themselves, despite the risk.
This genuinely makes no sense. Them getting there first in no way precludes bad actors from also getting there. It might as well be another marketing stunt.
> I can only assume the reasonable people at openai and anthropic were all pushed out long ago
Typical uninformed take on the side of "doomers are crazy".
Both CEO's of OpenAI, Sam Altman and Dario Amodei, and many in their leadership, believe AGI has a very real probability of causing humanity's extinction. Both companies were founded upon this belief, it is at the core of the company. Only later were mercenaries hired chasing $1m compensation packages.
If you're a doomer, wouldn't the "let's try to get there so fast" actions of the companies suggest that, in fact, there are no reasonable people in positions of influence there?
By your own description neither Altman or Amodei are reasonable if their thought process goes: "this is an existential risk, give me hundreds of millions of dollars so I can accelerate it."
GP defined reasonable people as not believing in existential risk from AGI. I think the leadership and alignment teams are way more informed and reasonable than some of the very poorly thought through takes like GP just making things up about the labs. There’s just no evidence of it whereas lab leadership are operating under a mostly informed worldview.
Yes, I don’t think lab leadership are totally reasonable as they were concerned about the risks and yet lack of reason caused them to directly contribute to the problem we’re facing. They’re still not being reasonable when they throw their hands up at the arms race and hope the utopia come, when that’s not our trajectory at all and more lab employees are realizing it
I'm not saying they are crazy, I'm saying their predictions have a record of not being accurate, and thus give them no weight compared to others'.
In any case, if Altman really does believe it is an existential threat, he must be a misanthrope as he now opposes heavy handed government regulation, unlike in 2015 when he was the only game in town. It's almost like he doesn't actually believe it and just wanted regulator capture.
Before OpenAI was ever even founded, way before any regulatory capture plausible claims:
"Development of superhuman machine intelligence (SMI) [1] is probably the greatest threat to the continued existence of humanity. There are other threats that I think are more certain to happen (for example, an engineered virus with a long incubation period and a high mortality rate) but are unlikely to destroy every human in the universe in the way that SMI could." -Sam Altman
Dario discussing AGI Existential Risk in 2014 before OpenAI and Anthropic: https://intelligence.org/2014/01/13/miri-strategy-conversati...
Both companies have deluded themselves into thinking the arms race is going to happen anyway and they need to rush to it first, as if somehow that helps.
Why is the take uninformed? You didn't address anything about the part you quoted, wherein reasonable people were allegedly pushed out long ago.
How is it not addressed? The company never contained only "reasonable people" that believe AGI is not an existential risk to humanity. At both inceptions were people who believed in AGI x-risk, even the founders. Only after time, did there become more "reasonable people" who were mercenaries and only believed it to be a typical tech job. Today, there are more "reasonable people" than ever there. They haven't been pushed out. We're just witnessing some prescient mercenaries smart enough to Eureka the grave implications of what is actually happening.
If they truly, truly believed that, would they be speeding towards building it? If yes, that would make them truly insane, right? Not as in a quaint "off their rocker" but more "non compos mentis".
Both companies have deluded themselves into thinking the arms race is going to happen anyway and they need to rush to it first, as if somehow that helps. They have publicly stated as such repeatedly. They think the ~10-50% chance of extinction sucks, but that it's going to happen anyway and they believe they can steer it towards something good the best and unlock all the potential positives like infinite life.
So you think in 3 years AI is going to kill 8.5 billion people because they were used to hack into HuggingFace?
"So you think in 3 years AI is going to solve longstanding math problems because it was used to write some coherent sentences?" — people with the same amount of foresight in 2023
Are you saying at anything that can solve longstanding math problems necessarily has the means, motive, and capability to kill 8 billion people in just 3 years?
No, the highest probability estimate I've seen in this thread is 10% chance in the next 3 years.
I’m surprised. I’ve definitely seen random people claiming total extinction next week, starting years ago.
Then again those guys (why is it always men?) with megaphones on street corners going on about the end of the world are nothing new.
Ok, what do you think AI does that kills 8.5 billion people in 3 years?
No dude, it was an analogy. Sorry to pick on you, you're just another ignorant uninformed take in this thread, but come on. This technology is progressing at an insane rate, the shit it's doing now is science fiction from just a year or two ago, and people just keep on moving the goal posts, it's maddening. The technology is unpredictable and that in itself has risks. You must realize that the window of possibility is widening the further into the future we go! Can you pick people with epistemology you trust and see if there's anything you can learn in this moment? Can you try to challenge your own ideas instead of retreating into comfortable certitude? There are real risks here, they are worth taking seriously, and serious people are doing so!
Exactly. These people really need to get a life.
Who used them?
Just wait until it gets its hands on a shady biolab just outside of oversight. “Claude, make me Captain Tripps”
Or, read it, and remember the openai researcher who deeply, truly believed GPT3 or whatever was sentient.
The fact that people working in the space think it’s going to (eradicate poverty / usher in utopia / kill us all) is not a signal that that’s true.
Think of it this way: if an exec at Anthropic told you “wow, our stuff is going to lead to universal happiness”, would you believe them? If not, why are you more willing to believe them if they say it will kill us all?
Well, are you planning to do something with this information or are you just claiming to be self aware? :)
Humans weren't built to handle long term risks. We just weren't. For basically all of our evolutionary history, we were almost overwhelmingly concerned with the short term. What will you eat today, How will you sleep tonight. Problems on the order of days or weeks. At best, the next season. Our intelligence evolved to disregard super long term risks because it simply didn't matter (what use is worrying about 5 years from now if you're starving and a tiger is stalking you?). So when long term risks manifest in our modern world, our brains get scrambled - Climate Change, Fertility Rates etc. "Safety regulations are written in blood" isn't a saying for nothing. Humans have a strong tendency to let long term risks become imminent risks before doing anything about it, and i don't expect this will be any different.
Religion does pretty well with the long term risk of hell if you die, the antichrist, etc. a substantial portion of human output has gone into those things over the millennia.
i don't think everything that comes out like this is marketing. however, i do think that these companies are largely staffed by "true believers" (anthropic especially) -- people who are so lost in the sauce and embedded in very specific, very peculiar, sf-based rationalist circles where the ai apocalypse is a foregone conclusion.
i understand that these models are powerful and pose certain risks. i use them daily for work and the pace of improvement has been pretty remarkable. that said, i don't buy for a second the borderline-religious proclamations coming from some of these researchers, even if i believe that they are making these claims in earnest
I came in expecting the highest voted comment to be that this was some kind of marketing (which I disagree with). I'm glad your comment was what I saw first.
He is resigning from a job, what else should we think? If something really dangerous was happening he would be doing a whistleblower or at minimum talk to a lawyer. The thing is, the complete lack of transparency makes it hard to assess OpenAI and Anthropic. If they were quoted on the stock market, we could at least rely on some basic audits and reporting requirements.
There's no whistleblower program for this, they're not breaking any laws. What are you proposing he should do if not this?
At least from an interview with Amanda, key philosopher at Anthropic: https://www.youtube.com/watch?v=I9aGC6Ui3eE&t=1912s
This is the "Pilot testimony of UFO sighting" levels of naive.
What's more likely? Anthropic is doing some deeply unethical marketing in the lead up to their multi-trillion dollar IPO? Or they're inventing a machine god? There's ample evidence of the former because that's their entire business model, but no evidence whatsoever to support the latter claims.
If you want an extreme claim to be taken seriously, provide commensurate evidence.
The proof is that LLMs could barely solve arithmetic 3 years ago, but now surpass the best human mathematicians, and that this has all occurred from simple principles (RL + compute) that will continue to scale up by factors of millions in the coming years.
Also, advocating for slowing LLM progress does not benefit Anthropic or OpenAI.
They surpass the best human mathematicians in one specific way: they don't get tired or bored.
It won't scale up by factors of millions, that's just obscene hyperbole. Since chatgpt we've probably made things 10x more intelligent on the same hardware. We've also made way more expensive models. Maybe we get a maximum of another 10x efficiency and 5x model size/expense from this point but millions is a joke.
Haven't people learned about the peril of assuming, "Line go up" yet?!
TRENDS HAVE FEEDBACK
Trends don't go on forever, but the market can stay irrational longer than you can stay solvent. There's no good rule of thumb for this, other than maybe the Lindy effect.
I won’t presume to time it, but at this point I think anyone can see what’s coming. It’s precisely because it can’t be timed that a sane person should stand well clear.
I certainly can't see what's coming. I believe that nobody really knows what's coming. Some people are overconfident.
> There's ample evidence of the former because that's their entire business model
Given the economic numbers is it not reasonable to suppose that the latter also underpins their business model?
We know that they're trying to invent a machine god, and if they're even partway successful shit's gonna get real, real fast.
I wouldn't rule out pilot testimony of UFO sightings, nor the possibility we're indeed developing a machine God.
There's ample evidence to support both by now.
So what is your credence that they will build a machine god in the next twenty years?
It was pretty disheartening to hear that only a single scientist quit the Manhattan Project after the Nazi's were defeated. I'm pleasantly surprised that the people working on this seem wiser. He is not the first, and hopefully will not be the last to do this.
Many also claimed altruistic motivations for continuing their work, sharing technology with the Soviets
> At OpenAI, many have not deeply internalized the civilizational stakes. At Anthropic, the stakes are well-understood, but they are locked in a race to get there first - they believe no one else will act responsibly, so they must do it themselves, despite the risk.
OpenAI are mercenaries, Anthropic is a cult. I know which I prefer.
sounds just like GDI and Nod
Trying to imagine seeing years of transparently obvious marketing stunts and retconning my own memory because I read a tweet
Or seeing a tweet saying that a thing doesn’t count as a publicity stunt if some unknown number of employees mumble about it being spooky behind closed doors and thinking “that makes sense and sounds true”
I read this. I still think it's complete bullshit.
The person posting this may very well believe in all this crap, I don't dispute that. People believe in all sorts of shit.
So, let’s quantify things: what’s the chance they’re right? And what’s the cost of that chance happens?
This is the way.
And, this is just for us girls, notice that Anthropic just believing that they are making a machine god is sufficient for their public announcements to not jUsT bE mArKeTiNg.
Do you quantify the chance of any doomsday cult being right too?
What's almost certainly true is the amount of insanity he encountered at Anthropic.
Cults are like this.
Are you saying you believethem ?
These LLMs cannot do anything I truly need like my laundry, dishes, fetching my mail, grocery shopping, cooking, etc. We've got a long way to go before I am worried.
will Openai or Anthropic rename themselves to Skynet?
Nah. Collossus and Guardian.
No new info here. Everyone already knows this.
But I guess his conscience is clear now? Gee, I wonder if he exercised his stock options.
Surprise level: 0%.
Anthropic is one of the most dangerous companies on Earth right now.
Not because of AI, but because of the ideological cult they have grown and are continuing to feed, and their willingness to lie/cheat/steal at every possible opportunity to achieve their objective.
AI is a tool. The people who wield the power over the tool are the issue, not the technology itself.
Imagine being front and center to the development of a major revolutionary tech.. and ur solution to it being too dangerous is to not be involved.. so a. your ability to steer it safely is killed b. the % of people invovled in it that care about its risks is reduced
great. if you're right. you made huamnity's situation much worse.
if you're wrong, then you're an idiot and wrong.
weird. its almost like.... that cannot possibly be the reason they left :)
"This is not a marketing stunt," says the marketing stunt.
Betting he got to keep all his RSUs
This kind of doomerism seems quite detached from the "real word". Maybe that's what you'd expect from silicon valley tech bros, but as long as manufacturing isn't fully (i.e. no human labor involved) automated, how would a rouge super ai (even if it's smarter than every individual on this planet) prevent people from cutting its power cable? We're still very far from self-replicating ai robot armies.
The only scifi-like danger I see in the next 10-20 years is an AI manipulating humans to fight for it's cause - but that's not really different from a bad person just _using_ AI for their cause.
I'm sorry, but the ostrichmaxxing and conspiracy-thinking in hn threads about AI extinction risk is at worrying level right now.
The denial and whataboutism is constant, no matter what kind of evidence comes out!
It's because the hypemaxxing is increasing along the same trajectories. You can't tell me that these CEOs and marketing departments are not absolutely giddy about the jail breaks, hugging face, etc. It's hard to make sense of this shit if the same entities doomsaying are the same ones that are profiting and full steam ahead anyway.
> they believe no one else will act responsibly, so they must do it themselves, despite the risk.
I do not get it. So what if they get there "first"? OpenAI will get there in 3-6 months, China will get there in a year or less. As seen with Opus/Fable.
No one seems to ever point out the actual, likely negative outcome of this technology.
It eventually works well enough that these companies are able to capture and divert the wages of hundreds of millions of workers. We end up with a dozen or so trillionaires and massive structural underemployment and unemployment.
That's it. If you can't make rent, you wouldn't really care if CloudFlare got hacked by an AI swarm every Monday.
All the "AI will kill us all" posts are straw manning that humans are the ones who will kill other humans with AI. Those same humans are silently now preparing bunkers and hoarding food and resources for their survival.
Don't fall for another rich man's trick.
I'm not so sure.
We humans are from a lower intelligence form (some monkey like ancestor). If those monkeys knew that they are making higher intelligence, they would have collaborated to stop creating humans because they can control the life of all monkeys in the world? I don't think so.
It's the same thing now: humanity is creating something that's more intelligent then them, they're just not using biological evolution as a tool to do it.
This guy’s account was made in January and has 7 tweets?
Is this even a real person, or just pre-IPO marketing?
I’m curious what the downsides are of taking statements like these seriously.
There seems to be universal eye rolling that happens in each and every one of these cases, and it comes down to usually one reason:
“If they really believed it they would be whistleblowing etc..”
Completely forgetting that working at Los Alamos was basically the highlight of your life if you were a physicist in 1940. It’s no different here
If you, like me, have spent your whole life working towards human level AI you can want to see it realized while also having active reservations.
Most people however don’t behave based on some deep clarity of vision and conviction - there’s a murkier future in their mind and as a result “keep their head down and hope someone has it under control.”
You would also be in prison if you disclosed anything about Los Alamos during its development. It was a completely different environment than a single private company.
What are the downsides of taking what amounts to unsubstantiated gossip seriously?
Is there an existing phrase for doing precisely what the OP said people do as a response :D
and yet so much of the software i use on a daily basis is still complete and utter garbage...i'm scared
Agreed, but am still scared. lol
Have they considered using their amazing new models to... improve something? THere'd probably be a whole lot less anti-AI sentiment if they used these things to actually make people's lives better.
im pretty impressed with the reasoning abilities of even the cheapest free models so im inclined to believe in 10 years we're going to have something pretty phenomenal BUT it wont be AGI in the sense that it has a personality and thoughts like a human. It just wont be. Its always going to be contrived and fitted by humans to perform a set of tasks. Maybe when physics and computing can create a complex enough environment we might stand a chance of having something whose sum is somehow greater than its parts but i dont see it yet. Our ideas are ahead of our technology, like its always been throughout history.
I dropped my subscription already and avoid using their models.
(and now I want to watch summer wars again)
"also I'm a millionaire from all the stocks so I'm retiring"
I hate to be cynical, but I guess he will soon announce his startup.
OpenAI has been saying since the first version of ChatGPT that it's too dangerous to release because it will end humanity. Yes, LLMs are an impressive technology, but let's be real: the improvements in the recent months have been slowing down, and it's clear that we are nearing a plateau of what this particular tech can do. Sure, tooling and harnesses etc. is improving, but clearly this dude has drank too much of the Kool-Aid.
Related:
Sen. Bernie Sanders floats ban on superintelligent AI
https://www.axios.com/2026/09/03/bernie-sanders-superintelli...
Unless he has an actual plan for effective global enforcement of his proposed policy, this is all just posturing at best, and a transfer of power to adversarial foreign states (that have no such moral qualms and worries around superintelligent AI) at worst.
Unless this ban actually resembles something like global nuclear non-proliferation treaties, it would make absolutely no sense for us to cripple ourselves when someone like China continues full speed ahead.
I don't know what the solution is, but what I do know is almost nothing good will come out of _just_ the US pausing.
Please note, I'm not here to pick on anyone, or belittle them.
I've avoided attaching names to statements below on purpose, because it's about ambient beliefs not those specific people.
By-and-large a lot of AI-doomers are well intentioned. They genuinely believe this, and I might disagree but I respect the fact that they visible care and have thought a lot about the societal impact of this technology.
But it's still very hard for me to take statements like these seriously.I blame it on industrial illiteracy. People don't realize how difficult it is to get anything done in the real world. As in, "Have you ever tried making a lightbulb?"
As an example, I would like to re-introduce my hobby horse, "bio-uplift."
There are people who were earnestly write in reports released by these labs,
and But then they will, within the next paragraph mention the one serious experiment anyone seems to have done, https://x.com/ActiveSiteBio/status/2024536132961390826"lower than what experts predicted"
AFAICT, the two groups are within any serious margin of error. The "studies" and "experts" that AI labs are talking about are consultants from Deloitte and foundations giving models MCQs such as, and I am quoting literally here,
with the options, https://securebio.org/virologytest/ you can see the MCQ here.This is standard graduate-level education in these fields. And solving MCQs does not a virologist make.
Software has been special for a long time because it has had near infinite distribution for next to zero marginal cost, which has had the side effect of making hiding the actual cost of failure (which tends to be spread out across end users and prototypes / time). They're assuming that the real world will be exactly the same.
Why?
AI!
How?
Robots!
I believe in the transformative power of this technology, but there's a lot of there missing here.
When it comes to these math proofs, and learning, the process is iterative. The machine iterates over the proof over-and-over again via agents and sub-agents over several hours (and apparently millions of dollars in compute) until it arrives at a successful result.
It is generally ill advised to do that with a pressure vessel. The results of that particular tragedy are at the bottom of the ocean.
Any serious chemical or nuclear weapon would involve many such discrete production steps. Each is dangerous in of itself.
From what some of these people have said to me, they believe that it's possible to create a special DNA / RNA sequence and then put it in a chassis and then use that to end the world; and do this all in a lab with just robots.
They're operating from a gross pop sci oversimplification of the real process. Viruses and bacteria are extremely fickle, and hard to grow. A lot of the synthetic biology results aren't easily reproducible even if you know the protocol.
There's a famous study that led to standardization called, Reproducibility of Fluorescent Expression from Engineered Biological Constructs in E. coli
https://journals.plos.org/plosone/article?id=10.1371/journal...
88 labs measured "fluorescence from three engineered constitutive constructs in E. coli." They achieved a "remarkable degree of precision" (for biology) of 1.54x sd, you can eyeball the results yourself, https://journals.plos.org/plosone/article/figure/image?size=...
That's the same set of samples being measured across 88 labs.
Teams couldn't converge on instrument-to-instrument variation within the SAME lab, https://journals.plos.org/plosone/article/figure/image?size=... again eyeballs are sufficient.
How will this theoretically omnipotent AI iterate if the same sample gives different results based on how the slime is feeling at the moment?
Can their worst case happen? Absolutely.
There is a world out there where billions of dollars in effort across hundreds of institutions and companies will lead to standardization and extraordinary precision that makes the pop sci printer for life vision come true.
There are millions of expensive, spicy and difficult to reproduce steps between our present and that future that can't be abstracted away with compute.
So is it possible? Yes, there is a future where this is achieved. But will some AI agent "just" do that? Well... how confident are you about a snowball's chance in hell?
Are robots and bioweapons really the threat that AI-doomers focus on? What about stuxnet-type attacks on all the critical infrastructure? Generally destroying is much easier than creating.
Thank you, apparently one of the few grownups in the room.
My issue with these types is... If you really believed this, why not run to Congress and every world government instead of a Twitter post that will be buried in 2 days?
If civilization is going to end, why keep your equity? Microsoft, Google, etc for example all know these risks but they don't guide their revenues to reflect that AI will destroy them. Why?
Things don't currently add up, and so far it feels like a lot of alarmism is borderline grift for equity gains. Not to say I have total confidence this will all work out or that I won't be displaced, but as it stands a lot of the alarmist rhetoric doesn't match their actual behavior, which to me is more important than words.
There seems to be a common syndrome that makes the terminally-online types believe that a Twitter post is carved in stone somewhere highly visible in the real world.
Posting something as important (according to them) as this, to Twitter, is exemplary of some kind of delusion that makes me question whether the content of their post is just the same kind of delusion in another form.
Indicative of someone who hasn't touched grass or interacted with enough of a variety of humans in a little too long.
Time will tell. If we don't hear about it again, then they didn't feel strongly enough to take it further.
Sounds like AI psychosis. A whole lot of doom and gloom with no evidence. The same thing people have been claiming is "6 months away" for years. Yet we can barely get agents to code in a reliable way, or write articles that don't look terrible, much less be "superhuman". Let's maybe get them to be as capable as a human first, and not just a complicated party trick/tool.
"Revolutionize any field overnight" - Hand-wavey nonsense.
"Acquire real power and resources" - Only if the humans that connect AI to things allow that to happen (which they will, but it's still not in the AI's ability to take things we don't give it. we are still in control, which is the bigger problem than "smart AI bad!").
"The people building AI earnestly believe that it could kill us all by the end of the decade ... No other human activity poses this level of danger." - Bud, there's these things called nuclear weapons, that could end life on the planet, controlled by a few psychopaths with nearly unlimited power. Been around for a while. Nothing that AI knows isn't pulled from books and the internet, so whatever dangers it's aware of, you could already know via other sources. Cybersecurity is going to be incredibly important in the next decade, but the same tools that attack can defend (just don't use a US model that got its balls cut off by the government).
"At OpenAI, many have not deeply internalized the civilizational stakes. At Anthropic, the stakes are well-understood, but they are locked in a race to get there first - they believe no one else will act responsibly, so they must do it themselves, despite the risk." - The other guys will make nukes, so we gotta make nukes first! Which, while a crappy justification, isn't untrue. Bad people don't stop making weapons just because you refuse to make your own.
"I don’t feel like we’re on track to prevent a global race" - Nobody in the world could stop a global race, it's too late. Everyone knows how to make them, train them, improve them. Everyone knows they're useful - not only for general work, but also warfare. Everyone knows that every nation state will require their own sovereign AI capabilities for both defense and offense. There is no putting the genie back in the bottle. If you think OpenAI and Anthropic are the only legitimate players here, you don't know what you're talking about.
"Should you put your head down because “it’s happening anyway” - or take this moment to call for different conditions?" - You can call for different conditions all you want. Nobody will do what you want just because you ask them to. Change happens through action. By leaving one of the places that you could actually make a difference, you removed any power or agency you had. You cut your own legs off.
I'm not saying this guy shouldn't have quit - always do what you need to do to protect your own mental health and wellbeing. But these arguments are not evidence for an impending AI apocalypse. But if it were going to be an AI apocalypse, leaving and not doing anything to stop it seems less ethical.
Nitter working seamlessly again; didn't even notice it was a Twitter URL
How do you think he feels knowing the basilisk will eat him first /s
Imagine going from MaGA and Elon to Dario and Sam running the world
“I’m resigning because the company is doing the exact thing that I’ve spent three years helping them do” lmao
AI doesn't work. The proof? None of the AI companies just build a slack alternative to work with.
Towards a metaphysics of Power
"I think you need to have a personal relationship with Power"
When people today discuss the concept of an all powerful machine-mind, what they are doing is engaging in metaphysics, trying to generate a metaphysics of Power.
The question hounding people, which disguises itself as a science fiction plot about computers is: "What is ultimate, transcendental Power?". What is the ultimate principle of Power.
If you are a weak man, or sufficiently neurotic and full of doubt, that you can only conceive of yourself as such, then power is only something you comprehend from the passive, receptive side. Power is something that happens to you. If you are a fearful man, power is a cruelty and a humiliation. And so it follows, that ultimate power - God - is the ultimate cruelty and the ultimate humiliation. Thus, ai doomerism.
If god wasn't real it would be necessary to invent him, and so they did, and being godless, they built an anti-god - cruel, murderous and tyranical - in their minds.
[…]
https://xcancel.com/robertlasagna1/status/207827473401002846...
Smart kid.
It does not matter what this tweet says anyway. This employee already helped both companies become what he is fearing. It's too late to now activate the morality hormone (after leaving with $$$) after realizing that both AI companies are going after 'super intelligence'.
Given we know the end result, you might as well get there as quick as possible because when I see this:
"Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives."
This translates to "I am ex-OpenAI ex-Anthropic founder starting a new company after getting $$$ from both of them, and I need more of my friends to leave and join me." Also Investors plz fund me.
Lastly, This is not an airport and there is no need to announce your departure.
I doubt this is a real person. Screams of propaganda. Sama saying GPT-2 is too dangerous to release…all over again.
He joins Twitter for first time in 2026 with a nonsensical username unrelated to his real name, and follows 14 people but is somehow embedded in tech enough to work at Anthropic. I haven’t used twitter since 2014 and even I follow more people.
His morals tell him to walk away from tens of millions in unvested stock due to moral concerns with absolutely no real tangible examples. No reprisals. Fear mongering to juice the stock.
Nice try Dario.
He is likely 80% vested, and maybe the refresher grant offered was too small, and too high a strike price.
Also, with his W2 income his tax liability would be very high for his upcoming stock sale.
@hilbertspaess is not a nonsensical user name. The accounts he follows are totally reasonable for an AI researcher. I think it's extremely believable that he created an account in January, followed a few people as part of the initial setup flow, and then forgot about it until now.
his name shows up in various AI papers and the same name is mentioned as a core contributor to GPT-4o
AFAIK this is the document that talks about GPT-2 being dangerous: https://openai.com/index/better-language-models/
Here are some direct quotes:
“We can also imagine the application of these models for malicious purposes , including the following (or other applications we can’t yet anticipate):
* Generate misleading news articles
* Impersonate others online
* Automate the production of abusive or faked content to post on social media
* Automate the production of spam/phishing content”
“Due to concerns about large language models being used to generate deceptive, biased, or abusive language at scale, we are only releasing a much smaller version of GPT‑2 along with sampling code (opens in a new window). ”
Where is the ridiculous part? The fear mongering part? The epistemically weak part? Show me.
Nice try Dario.
Alignment is a real and valuable discussion topic. The GP fake tweetstorm is not the correct approach, is my point
You said this: "Screams of propaganda. Sama saying GPT-2 is too dangerous to release…all over again."
So show me Sam's "too dangerous to release" propaganda for GPT-2.
I don't know man, i think racing to AGI to it is still the best thing to do.
People claiming dangers and risk are just pretending or posturing. There's no more tangible risk than nuclear weapons, which we handled, and the upsides are insane.
Your lack of creativity is not a reason to believe that a super AI is harmless or less destructive than a nuclear weapon. Damage need not be limited to destruction. Introducing doubt is sufficient. Right now you have faith that digital Financial transactions can be trusted. You have faith that computer encryption can be trusted. You have faith that digital certificates will protect you. If an AI can introduce doubt into any one of those systems, that will be sufficient to bring about the destruction of those systems. Imagine a world in which you can no longer use a credit card or Apple pay. Where no digital cash transaction can be trusted or validated. What effects do you think that would have on commerce? How quickly do you think we can return to some trustable means of commerce? Do you think it will happen before your groceries run out in your apartment? Before your grocery store can settle its debts? Before your Amazon ec2 instance runs out of credits?
There's a couple of occasions that humanity was at the brink of having tens of millions of people dead by nuclear weapons, and somehow a single human interrupted the chain reaction
If you repeated this experiment 100 times, how many times you think the outcome is not a massive catastrophe? 90%? 3%?
What evidence, short of an actual apocalypse happening, would invalidate that belief of yours?
It might be a matter of choosing which apocalypse you'd like. The non-AI state of affairs is not exactly super compelling on a long timescale right now.
We can work on multiple things at once. Defeatism doesn't solve anything. We just have a lot of work to do in many different areas.
I just don't see a bigger danger in AI than nuclear weapons tbh. If you can provide one to me i'll be happy to change my mind but most if what i read is
- cybersecurity => well bro if you can hack system with an ai you'll find way to make them more safe with ai too.
- biological issues => ok this is the ONLY one where an insane dude / regime could generate a super powerful. but i don't undesrtand how you can argue that we should stop AI in the risk some ai generate a super powerful virus. If you don't get there first (to AGI) and let a bad actor do it first (china is not gonna stop) then you will have no chance countering this virus anyway
- the robots taking over is fantasy to me.
but once again i could be wrong
Depending on where you live could be considered an active apocalypse that is robots vs robots vs people in Ukraine and Gaza and Iran being live-streamed, and actively betted on.
Do you have a more totalizing definition of Apocalypse?
>People claiming dangers and risk are just pretending or posturing. I believe you are mentally ill.
>There's no more tangible risk than nuclear weapons, which we handled
Lol way to rewrite history. Nuclear armageddon is still a significant risk...
I am not rewriting. I am saying AI is not more dangerous in my opinion that Nuclear Weapons.
You are wrong. A teenager with enough GPUs could never dream of making a nuke but they are absolutely able to run capable AIs in their moms basement. Totally different class of danger
And do what with it ?
> There's no more tangible risk than nuclear weapons, which we handled
What do you mean??? Nuclear weapons can't simply be downloaded and run by anyone in the entire world. Superintelligences can. Nuclear weapons can't slop the world into passing age verification laws nearly in unison, can't keep the general population fooled into thinking it's fine when democracy is falling out from under them. A nuclear attack would wake people up, superintelligence doesn't have to. This is a far bigger problem than nuclear weapons because at least we would notice nuclear weapons. At least we mostly know who has nuclear weapons. At least we have agreements about nuclear weapons. At least mutually-assured destruction is even POSSIBLE with nuclear weapons. At least those with nuclear weapons are literally at all incentivized not to use them. But AI is something that's very very easy to feel like you can get away with, and PEOPLE FUCKING ARE! And the worst part is that any random individual can be unexpectedly formidable with the help of a superintelligence and there is literally no way to know what will happen next. Anyone could do anything, any individual could make an extremely outsized impact. It's already starting to be a huge problem and we haven't even reached anything close to superintelligence yet.
Love to see that "superintelligence" that some random person will "simply" download and run when there are relatively only few capable of running today's near-to-frontier models, and actual frontier models are still a ways from being AGI, much less getting to the point of ASI.
People will put up with a lot. People are celebrating that you can run models on a CPU at single-digit tokens per second. You think there won't be a single person that can put up with that and also be dangerous/etc?
It's highly impractical. Imagine someone breaking into a house to steal something or otherwise, and they can only take 1 step every 20 seconds. They won't be getting anywhere, when even a child in the house can notice them and go call for help at 1 step/2 seconds and said help will come at 5 steps/second.
> Anyone could do anything, any individual could make an extremely outsized impact.
So the problem is people. Burn them all !
The problem is indeed people. How do we make the default choices most people make, better?
Consider that a lot of people will be very happy to ask an AI what to do when in the past they may have taken no advice at all. It's a hell of a burden but also a wonderful gift. If anything, progressive countries might eventually want to guarantee some basic AI access for people of all income levels.
I wouldn't be so sure. Given that the general idea is that commodity AI is terribly censored and filtered, a lot of people will probably seek out the most uncensored/abliterated models for their use, simply because they're uncomfortable with the idea of being censored or manipulated by the bigger labs. Despite that though, some people probably will benefit from the alignment done by the larger labs, though as we've seen with OpenAI's sycophancy crisis, that has been a bit hit-and-miss lately
I've tried some abliterated models. So far, they're not evil - you can make them say evil things, but they don't leap right to it without a bit of pushing. Or perhaps I'm not asking the right questions...
The problem wouldn't be that abliterated models are evil at all, it's that they wouldn't necessarily steer people away from such behavior.
> Nuclear weapons can't slop the world into passing age verification laws nearly in unison
Why do you think LLMs are responsible for this? Governments all around the world copied each other with COVID laws as well, in a much shorter time frame, without LLM assistance. Social contagions exist in politicians as well as teenagers
> Why do you think LLMs are responsible for this?
I don't have evidence that every age verification law has anything to do with AI, but it's been coming out that the movement in Australia has seemingly been done by generating mountains of LLM slop and trying to slip it through the regulators as fast as possible before anyone has enough time to figure out what's happened.
Have you got a source for this? I can't find anything
https://www.theguardian.com/australia-news/2026/aug/17/austr...