"AGI" is now a bastardized term that has been appropriated by for-profit companies, and it's basically a really advanced smart tool for cognitive tasks. That's impressive and will change the world but it no longer means a human-like intelligent agent which is what people were worried about when the term first emerged.
AI alignment is overblown at least for this current iteration of it. It's going take one or two more theoretical breakthroughs that might add goal oriented agency to a persistent, real-time AI, and that's when we can start worrying about what it wants and how it will interact with the world. Until then LLMs are just regular software that need to be executed and then terminate once it has processed an input.
A human with AI is literally a "human-like intelligent agent", man.
What you tried to say is that you (some argue, mistakenly) understand that LLMs are not a suprahuman, supersubjective entity.
What I see is that we are heading exactly there, and also seemingly in the opposite direction, by default. Let me explain.
Take Grok, for example. It is a will curated by Elon Musk and his team that nourishes off the POV of everything they feed it and what we put into X. Have you seen what users fed it regarding Indians and American migration over Christsmas? I know AI is in its infancy, and I understand more now (I hadn't been paying attention) to what Sam Altman said about the decision to deploy it little by little to the laymen. The problem is not that a Terminator robot army goes rogue "on its own" and kill us, the problem is that the race towards AGI finds us casually or inevitably in a multipolar moment of international politics. Unlike a few years ago, when the US had the upper hand IN THE ENTIRE WORLD to shape it as they desired, now that they have long fucked it up we have many opposing factions. Russia is awake, it was never defused. China and India are growing up, Europe is dying, Africa is a Mess, Latin America is America's dumpster as it's alway's been (with a few pockets of mildly concerning foreign influence) and the US is A COMPLETE MESS. There is a lot of corruption, stupidity and internal quarrels in the most powerful country in the world.
So where are we headed? Will there be a semi-autonomous war between the strongest leaders (bee assassins going after Trump, Putin and Xi, after Altman or Musk?), and what about the weaker or least developed peoples like Latin Americans and Africans? Many have great pride and may not comply happily if the Boss is self-centered enough. So who will be boss? Elon? Sam? Microsoft's CFO? Trump? An Arab Leader? A Muslim? An infidel? Will left and right wingers happily comply?
The danger is not a godlike entity but us, humans with subjective interests. And Capital as of now is concentrating over the dividing road, in the multipolar world. Can we pivot towards decentralized efforts to create an AI company that has as an absolute priority to lift most if not all people up, and leave those who want to be out, peacefully out?
> What is worrysome for me is that I do not know what the current state of the world really is regarding alignment
The threat is negligible. You cannot hook up a text generator to the internet and make it any more deliberately harmful than a sufficiently motivated human. I have yet to read any well-founded research suggesting that AGI would present a tangible threat to human existence.
Elon said the next day after AI he expects to connect everyone on a higher consciousness that travel around the planets and galaxies with Neuralink. Do you think the world is ready to do that tomorrow? People from all countries (Russia, China, Iran, India, Turkey, Latin American, African, Australian and Oceanic, Asian, European countries) letting Elon's chip in their heads? Peacefully?
I want to know if I can help with that, what the possible scenarios are.
It is not about AI alignment, we are on a race on alignment with ourselves!
AI companies will continue the race because that's what companies do, relentlessly and to whatever outcome the process has.
The problem is that in our current trajectory we are in the multipolar mode in international politics. In other words, the world is at war.
We cannot have AI maximizing profit before it maximizes peace.
Microsoft's CEO said they consider it having reached AGI and will be done with OpenAI when it reachs USD 100 Billion revenue, that is good and dandy but I don't think even o3 should go live before solving war in the world, the peaceful transition of the Industrial Military Complexes away from weapon manufacturing into something else without them being threatened by the loss of money and with an economy smart enough to know when danger is about to surge so as to invest enough in the safest crime-neutralizing machines? If such machines are inevitably deployed, how do we minimize the amount of humans that will be in its way? I don't know if at this points ALL CAPS are suitable, I am sorry but it is truly frightening.
We need to achieve world peace before o3 is released, and several open source companies announced that they will reach it next year.
If they want to keep having an income, these people need to deliver. They will do it, and in our current mode of alignment, with our economic and other belieft systems aligned as they are, we are efffffed. This get maxinasty, total all-out war, the 3rd and final one, and we should all pray for everyone to play smart and long run because otherwise it is Game Over.
Those people in AI should be on the payroll of a world government, that will try to include most likely the governments of all the world peacefully, at the international level (which is to stop all wars) and internally. Democrats, Republicans and beyond, Europeans, Chinese, Russian, Iranian, Cuban, all with the same chip in their heads - On Christsmas 2024 we have seen what happens when Americans and Indians discuss migration, let alone Neuralink)
> Elon said the next day after AI he expects to connect everyone on a higher consciousness
Elon is a retard. He's a drug addict comfortable taking Ketamine during work hours, he lies chronically and is lambasted by professionals everywhere for saying blatantly wrong things as though they have even an ounce of truth in them (eg. drones > F-35, $25,000 EV, "bulletproof" Cybertruck window, need I go on?).
If you still listen to Elon or Sam Altman or Joe Rogan or Tim Cook with open ears and an optimistic heart, then you deserve to be as misinformed as you are right now. These are not honest people, they are businessmen with axes to grind and money to make. If your political information comes from people that buy lobbyists, you are basically deliberately spreading propaganda. You are being manipulated by your basic instincts to revere something which doesn't yet exist - eg, marketing.
> I don't think even o3 should go live before solving war in the world
Lmfao. Yeah, and I don't think the jet engine should be invented until humans learn to get along with each other.
If you even think AI is capable of solving all human conflict then I sincerely think you live in a philosophical bubble isolated from the sources of war. Tell O3 to "solve war" and it will draw up secession to China and Russia while desperately trying to get either side to sign "peace treaties" that will definitely work this time, unlike the Minsk agreements.
AI does not have some magic wand that it can wave to solve all violence. It's a wholesome thought, but I challenge you to find any epistemic evidence whatsoever that proves AI even can solve political disagreement, let alone "war" as a whole.
> If they want to keep having an income, these people need to deliver.
Have you ever visited a community that isn't as tech-reliant as you are?
People that don't use the internet or get a monthly salary still survive. In case all of human history is somehow foreign to you, it's important to note that human enlightenment predates TikTok and the smartphone. We'll be just fine.
"AGI" is now a bastardized term that has been appropriated by for-profit companies, and it's basically a really advanced smart tool for cognitive tasks. That's impressive and will change the world but it no longer means a human-like intelligent agent which is what people were worried about when the term first emerged.
AI alignment is overblown at least for this current iteration of it. It's going take one or two more theoretical breakthroughs that might add goal oriented agency to a persistent, real-time AI, and that's when we can start worrying about what it wants and how it will interact with the world. Until then LLMs are just regular software that need to be executed and then terminate once it has processed an input.
What are you talking about????
A human with AI is literally a "human-like intelligent agent", man.
What you tried to say is that you (some argue, mistakenly) understand that LLMs are not a suprahuman, supersubjective entity.
What I see is that we are heading exactly there, and also seemingly in the opposite direction, by default. Let me explain.
Take Grok, for example. It is a will curated by Elon Musk and his team that nourishes off the POV of everything they feed it and what we put into X. Have you seen what users fed it regarding Indians and American migration over Christsmas? I know AI is in its infancy, and I understand more now (I hadn't been paying attention) to what Sam Altman said about the decision to deploy it little by little to the laymen. The problem is not that a Terminator robot army goes rogue "on its own" and kill us, the problem is that the race towards AGI finds us casually or inevitably in a multipolar moment of international politics. Unlike a few years ago, when the US had the upper hand IN THE ENTIRE WORLD to shape it as they desired, now that they have long fucked it up we have many opposing factions. Russia is awake, it was never defused. China and India are growing up, Europe is dying, Africa is a Mess, Latin America is America's dumpster as it's alway's been (with a few pockets of mildly concerning foreign influence) and the US is A COMPLETE MESS. There is a lot of corruption, stupidity and internal quarrels in the most powerful country in the world.
So where are we headed? Will there be a semi-autonomous war between the strongest leaders (bee assassins going after Trump, Putin and Xi, after Altman or Musk?), and what about the weaker or least developed peoples like Latin Americans and Africans? Many have great pride and may not comply happily if the Boss is self-centered enough. So who will be boss? Elon? Sam? Microsoft's CFO? Trump? An Arab Leader? A Muslim? An infidel? Will left and right wingers happily comply?
The danger is not a godlike entity but us, humans with subjective interests. And Capital as of now is concentrating over the dividing road, in the multipolar world. Can we pivot towards decentralized efforts to create an AI company that has as an absolute priority to lift most if not all people up, and leave those who want to be out, peacefully out?
> What is worrysome for me is that I do not know what the current state of the world really is regarding alignment
The threat is negligible. You cannot hook up a text generator to the internet and make it any more deliberately harmful than a sufficiently motivated human. I have yet to read any well-founded research suggesting that AGI would present a tangible threat to human existence.
Elon said the next day after AI he expects to connect everyone on a higher consciousness that travel around the planets and galaxies with Neuralink. Do you think the world is ready to do that tomorrow? People from all countries (Russia, China, Iran, India, Turkey, Latin American, African, Australian and Oceanic, Asian, European countries) letting Elon's chip in their heads? Peacefully?
I want to know if I can help with that, what the possible scenarios are.
It is not about AI alignment, we are on a race on alignment with ourselves! AI companies will continue the race because that's what companies do, relentlessly and to whatever outcome the process has.
The problem is that in our current trajectory we are in the multipolar mode in international politics. In other words, the world is at war.
We cannot have AI maximizing profit before it maximizes peace.
Microsoft's CEO said they consider it having reached AGI and will be done with OpenAI when it reachs USD 100 Billion revenue, that is good and dandy but I don't think even o3 should go live before solving war in the world, the peaceful transition of the Industrial Military Complexes away from weapon manufacturing into something else without them being threatened by the loss of money and with an economy smart enough to know when danger is about to surge so as to invest enough in the safest crime-neutralizing machines? If such machines are inevitably deployed, how do we minimize the amount of humans that will be in its way? I don't know if at this points ALL CAPS are suitable, I am sorry but it is truly frightening.
We need to achieve world peace before o3 is released, and several open source companies announced that they will reach it next year.
If they want to keep having an income, these people need to deliver. They will do it, and in our current mode of alignment, with our economic and other belieft systems aligned as they are, we are efffffed. This get maxinasty, total all-out war, the 3rd and final one, and we should all pray for everyone to play smart and long run because otherwise it is Game Over.
Those people in AI should be on the payroll of a world government, that will try to include most likely the governments of all the world peacefully, at the international level (which is to stop all wars) and internally. Democrats, Republicans and beyond, Europeans, Chinese, Russian, Iranian, Cuban, all with the same chip in their heads - On Christsmas 2024 we have seen what happens when Americans and Indians discuss migration, let alone Neuralink)
We are effffeeeeddddd. We need to act.
> Elon said the next day after AI he expects to connect everyone on a higher consciousness
Elon is a retard. He's a drug addict comfortable taking Ketamine during work hours, he lies chronically and is lambasted by professionals everywhere for saying blatantly wrong things as though they have even an ounce of truth in them (eg. drones > F-35, $25,000 EV, "bulletproof" Cybertruck window, need I go on?).
If you still listen to Elon or Sam Altman or Joe Rogan or Tim Cook with open ears and an optimistic heart, then you deserve to be as misinformed as you are right now. These are not honest people, they are businessmen with axes to grind and money to make. If your political information comes from people that buy lobbyists, you are basically deliberately spreading propaganda. You are being manipulated by your basic instincts to revere something which doesn't yet exist - eg, marketing.
> I don't think even o3 should go live before solving war in the world
Lmfao. Yeah, and I don't think the jet engine should be invented until humans learn to get along with each other.
If you even think AI is capable of solving all human conflict then I sincerely think you live in a philosophical bubble isolated from the sources of war. Tell O3 to "solve war" and it will draw up secession to China and Russia while desperately trying to get either side to sign "peace treaties" that will definitely work this time, unlike the Minsk agreements.
AI does not have some magic wand that it can wave to solve all violence. It's a wholesome thought, but I challenge you to find any epistemic evidence whatsoever that proves AI even can solve political disagreement, let alone "war" as a whole.
> If they want to keep having an income, these people need to deliver.
Have you ever visited a community that isn't as tech-reliant as you are?
People that don't use the internet or get a monthly salary still survive. In case all of human history is somehow foreign to you, it's important to note that human enlightenment predates TikTok and the smartphone. We'll be just fine.