>Microsoft had editorial control over this post, cutting sections and figures and reshaping how the impact is described before publication.
that is... not great. shame on microsoft.
its actions like that which shed light on why we get the nighmare eclipses of the world. pressuring a kid into handing over full editorial control of a disclosure is gross.
Sounds like he’s gotten bug bounties from MS in the past. Might be forward thinking to keep the relationship amicable going forward.
Giving the benefit of the doubt to Microsoft it seems like a pretty complete write up. If the changes weren’t material that might have been part of it too.
At that age in the mid 2000s, I would be foaming at the mouth to tell a FANG company to eat dirt over something like a disclosure + unpaid bounty. But if he got the money and didn't negotiate... You live and you learn.
Of course not. You don’t push back on something like that. You hire a lawyer and let them do the pushing for you. If you contact the right NGO, they might even give you one for free.
There’s no way I’d sign any agreement with a company like Microsoft regarding an issue like that without a lawyer. I’d rather not disclose at all if those were the only options.
I've seen hackers. Even if you do it right you're still going to get banned from the internet till you are 18 and dating Angelina Jolie is not the draw it was 20 years ago....
There are so many great one liners in that move. An underrated one is when Dade's mom opens the door, sees Acid Burn and says "now I see what all the fuss is about".
I mean when you're sixteen, you can sign a legally binding contract (might depend on the jurisdiction). Usually you have to be above the age of majority (18) or be emancipated before you can enter into a binding contract.
Speaking of being sue, I once found a hole in a healthcare system. I solicited legal advice and the response was overwhelmingly against reporting. The company didn’t not have a bounty program. So I shut up knowing my info is sitting there exposed.
I wonder what’s the consensus on this? Do people normally report it or not?
On some forms I was advised to report hippa complaint. I can’t remember if I did. This was 10 or more years ago.
I think my take would be to report it anonymously (via support/marketing/etc.) and in the report strongly encourage them to create a path for security reports (either via a bounty or just a security email). When fixed or actually dismissed post publicly on anonymous channels, with a section saying how they could have created a channel for these reports.
if you are reporting it for moral/ethical reasons, but are worried that the company will have a shitty response, report it anonymously to the company and any relevant regulatory bodies.
but most companies, even without a bounty program, are happy to receive reports. a lack of a bounty program isn't really an indicator of how they will treat reporters, as most companies are ill-equipped to have a proper bounty program.
otherwise, and i know this is an unpopular opinion here, but i'm a big advocate of just posting the exploit publicly when a company makes it hard/scary to report it to them.
My issue was knowing that reporting would allow them to possibly discover my identity. Discovering the vul in itself may instead be considered unauthorized access.
I will tell my story: I found a huge accounting error that allowed for several employees to embezzle funds. I disclosed it to the main stakeholders and some people went to jail. The accounting team still refused to acknowledge that the mistake was theirs, so I left on principle.
A few weeks later, I was talking to an old colleague and they revealed that the story that came from stakeholders was that the accounting team found the error, and I left in disgrace (despite that not being my department?).
So if you find an embarrassing mistake and you do not control the narrative, you have to proceed very carefully.
Had one like that. I quit and they blamed me for a project failure after the fact. That pissed off a couple of colleagues who thought it might happen to them. They also quit leaving them entirely without a software team. Set them back ten years because they didn’t make the market in time.
Write everything down and make sure everyone knows you’re writing it down. Saves a lot of hassle like that.
This is a company that has engaged in heavy monopolistic behavior and violating US anti-trust laws with reckless abandon. They have even pioneered the infamous "embrace, extend, extinguish" strategy.
M$ will never change. The modus operandi is always the same, regardless of which ever lame MBA douchebag is running the shit show.
If we want to actually care , rather than spew off platitudes, break up big tech.
It's a little perplexing. Of course it's always a controversial topic since it's difficult to value an exploit, but whenever we read about these online, which probably goes through some survivorship bias, they seem pretty low.
Every HN post regarding security exploits inevitably results in some comment saying the bounty is too low. I find it helpful to read previous comments by tptacek regarding bug bounties and market values:
This is correct. I've made well over six figures over the last couple of years through bug bounty programs. I wouldn't spend months finding one bug. It's usually days or a week or two max.
$5,000 is actually pretty normal for a critical. I think the most I received for one bug was around $10,000, but this is the exception.
AI has also ruined the market. I'm a security consultant (where I make most of my money) and stopped bug bounty once AI slop reports made it impossible to actually get anything triaged in a timely manner.
Something else many don't know is that with all of the major platforms, customers can see the bugs before they are even triaged by the platform team. Many companies are now taking advantage of the long triage times and fixing the bugs before the researcher can get paid. It's then marked as a duplicate and the researcher gets nothing.
> Many companies are now taking advantage of the long triage times and fixing the bugs before the researcher can get paid. It's then marked as a duplicate and the researcher gets nothing.
Isn’t that just a speedrun-encouragement for selling the exploit not to the one offering the product, but an attacker offering more (in this case, >$0 is not difficult to exceed) instead?
As I understand it, bug bounty awards are a rough proxy for "would nation-state actors be able to exploit this for operational purposes without getting caught".
Zero-click iPhone exploits that affect the current OS and also previous ones are worth hundreds of thousands.
I personally think it sets a bad precedent: you want to broadcast that valuable info will be treated as such. A hacker in the future will see this, compare the "precedent" price to the black market, and not even ask Microsoft the next time.
I think HN, like a lot of people in this industry, have strongly skewed perception of the actual importance of these bugs.
You could view the bounty prices as market evidence that most of this is rightfully treated as nothingburgers. I.e. the alternative to paying $5000 to some random person for this class of vulnerability research is not risking a trillion dollar hack the next day - it's just risking shmaybe some kerfuffle down the line, followed by fixing it through normal triage process. The bounty program is as much marketing as security, and $5000 is probably about the right price for marginal effort into sustaining the "we are treating security seriously" message.
In a way, the very existence of those bug bounty programs in large companies is evidence they don't see a reason to treat vulnerabilities seriously enough to proactively find and fix them in-house.
If security vulnerabilities would be anywhere serious as most commenters on-line seem to think, companies would pay hundreds of thousands for serious vulnerabilities, just to save a day before they get hit by them - on top of spending millions in-house to try and stay ahead of the attackers.
But they don't. Because most exploits are inconsequential and/or aren't being exploited much.
What's worse is the alternative is winding up like Aaron Swartz... (all he did was scrape PDFs for mostly public funded / tax funded papers) Which is even worse, I think I'd be glad to receive $20 for Starbucks instead of being legally chased for showing them they messed up.
There should really be laws for protecting security researchers who produce 0 harm and divulge / share a vulnerability with a service provider. I'd rather the floor be getting no money AND not going to jail or being sued.
"all he did was scrape PDFs for mostly public funded / tax funded papers"
He wasn't a security researcher. He broke into a room and used equipment to steal information. It wasn't just 'tax funded papers'. Companies invested millions of dollars into some of this research.
We shouldn't support theft and he should have gotten some jail time/punishment for it.
"There should really be laws for protecting security researchers who produce 0 harm and divulge / share a vulnerability with a service provider. I'd rather the floor be getting no money AND not going to jail or being sued."
Too many 'security researchers' demand money or threaten to release the vulnerabilities.
I don't know anyone that got into trouble going through a legit bug bounty program.
I have not done a deep dive so I could be wrong, weren't these papers published (or soon to be)? Wasn't the one and only group with negative effects the Journals that could extract a fee? Especially with researchers usually happy to send a copy? This wasn't someone stealing trade secrets they weren't even secret.
The young researcher is a teenager. Getting your name out there, clout/influence/fame, is worth $1M to them.
It’s almost like most of you people have already forgotten your teenage years. And it shows. Doesn’t matter the decade or generation. A teenager , likely living with parents or grandparents, will always prioritize intangibles.
A mere 5 bands for full editorial control is quite literally peanuts for M$. It’s yet another case of abusing free labor, unfortunately.
> Hey! I’m Faav. A little over a year ago, when I was 15, I published Break into any Microsoft building: Leaking PII in Microsoft Guest Check-In, my first Microsoft write-up. I’m 16 now, and this one is a little bigger.
Damn, these guys got schooled by a 15 year old! Say less...
Physical security vendors usually stop exactly there. I can't count how many vendors responsible for badge systems shut off the Windows Firewall just because it’s easier than adding a rule.
Honestly that series of blog posts that came across here semi-recently about what an utter disaster the Azure team is (through little fault of their own, mind you), combined with a lot of things I've read about the development behind Halo Infinite, have convinced me Microsoft is to avoided whenever possible. It's not even that they're too big to care, though that is an issue. It's that organizationally they are SO reliant on short-term contractors and junior devs plugging away with copilot, along with an ever more exhausted group of seniors who actually work for them but are continually disempowered, that I think it's safe to say anything they manage to ship is a minor miracle if it works at all, and it's certainly not going to have any guarantee at all of solid software engineering fundamentals.
And to be clear, this is not an issue with them using contractors, overseas or otherwise, or with their senior dev staff, or even with AI really. It's an issue with them organizationally being so incredibly penny-pinching, and so dedicated to shipping new shit versus fixing anything long term, constantly chasing new revenue and letting their existing offerings rot.
If a Microsoft product is good nowadays, it is literally a miracle.
There is an internal library at Microsoft that reliably avoids all these JWT problems - Microsoft Identity Service Essentials (MISE). Adopting MISE and upgrading to the latest versions of it have been part of the Secure Future Initiative (SFI) that can be read about in the news of previous years. Unfortunately it sounds like the service team intentionally deferred the compliance alerts they will have received.
There is so much work to do for SFI that is still being ignored. The only way that some of these services will update is by being the target of the red team, security researcher, or threat actor.
That being said, just last night I observed that the identity team is now opening up agent-assisted PRs against individual service team repos to force MISE adoption and upgrade to the latest version and best practices.
I think that's a great thing because many individual service teams simply lack the bandwidth or knowledge. Prior to GenAI availability I wasted many cycles on this kind of work. While GenAI made it easier - internal source documentation still does not unambiguously address every use case. So having the identity team drive this now with the help of agent sessions initiated by them is great.
Yea, and it was a big issue years ago. I can't believe there is any modern implementation that has it.... or another way to put this is, how old is this damned implementation they are using?
Argh, I missed that it actually uses the "none" algorithm. Yeah, the existence of that option is extremely dumb and it shouldn't be possible to use that. I misread the post and thought it was a regular JWT, but they simply didn't validate it.
Does this extend to OIDC? I’m not knowledgeable on the topic but it uses JWT right? Is it also prone to poor implementation? If you just error on alg=none does that solve it?
Other commenters are suggesting you can’t blame the spec for end implementation mistakes, except that’s one of the many issues - JWT being so error-prone is a problem.
I use JWT just for handling of tokens, because it’s so well supported, but I won’t use it for anything more than token storage _because_ it is so vulnerable to mistakes.
The fact that mistakes are so easy to make is indicative of poor design in the spec itself.
This is a pretty typical example of the security posture of microsoft, and yet people continually buy their arguments that open-source and therefore auditable alternatives are inherently less secure than their "trust me bro"
Entra has already had vulnerabilities where a user in one tenant could access global admin of every other tenant. Somehow everyone didn't immediately move off Microsoft's cloud platforms.
> Two quick notes first. The impact I describe is hypothetical. It’s what an attacker could have done with this access, but luckily I found the bug instead, reported it, and never touched any customer data or PII.
I am the last person to judge someone for using AI to help them write a blog post, but what I wonder is: Do people not read what the AI produces before putting their name on it, or is the AI writing style not obvious to some people, or do they just not care that it's obviously AI and bad style?
It's not a "normal" sentence and is quite clearly produced by an LLM, it's a typical Claudeism so probably that. The entire post is also flagged by Pangram, so OP is correct.
Prefixing saying something with "Two quick notes first" is extremely common AI meta commentary. You may be right that it is something Microsoft insisted he put at the top, which would also explain the weird style.
I pasted the thread into ChatGPT and here is what it would like to tell you:
The comment explicitly says, “I am the last person to judge someone for using AI to help them write a blog post.” The criticism is about the quality of the published prose and whether anyone reviewed it before putting their name on it.
You can dispute whether that sentence is bad or whether it indicates AI authorship. But “why criticize unedited AI prose when HN talks about AI so much?” doesn’t identify a contradiction. Discussing a technology doesn’t imply endorsing every use of it, and criticizing its output isn’t the same as vilifying someone for using it.
>Microsoft had editorial control over this post, cutting sections and figures and reshaping how the impact is described before publication.
that is... not great. shame on microsoft.
its actions like that which shed light on why we get the nighmare eclipses of the world. pressuring a kid into handing over full editorial control of a disclosure is gross.
They should have paid him at least a $1 million bounty if they're going to ask for editorial control of the disclosure.
Love to hear more on the motivation for agreeing to this.
e.g. The $5000? Amnesty from being sued?
Sounds like he’s gotten bug bounties from MS in the past. Might be forward thinking to keep the relationship amicable going forward.
Giving the benefit of the doubt to Microsoft it seems like a pretty complete write up. If the changes weren’t material that might have been part of it too.
And he’s 16. Parents might have had a say.
Also he's 16. I'd definitely be less willing to push back (especially on something like this) at his age
At that age in the mid 2000s, I would be foaming at the mouth to tell a FANG company to eat dirt over something like a disclosure + unpaid bounty. But if he got the money and didn't negotiate... You live and you learn.
Of course not. You don’t push back on something like that. You hire a lawyer and let them do the pushing for you. If you contact the right NGO, they might even give you one for free.
There’s no way I’d sign any agreement with a company like Microsoft regarding an issue like that without a lawyer. I’d rather not disclose at all if those were the only options.
I don't know about you, but when I was 16, I never hired a lawyer, and none of the other 16 year olds I knew had either
I've seen hackers. Even if you do it right you're still going to get banned from the internet till you are 18 and dating Angelina Jolie is not the draw it was 20 years ago....
> dating Angelina Jolie is not the draw it was 20 years ago....
not not either. where do I sign up?
There are so many great one liners in that move. An underrated one is when Dade's mom opens the door, sees Acid Burn and says "now I see what all the fuss is about".
She’s glorious in the movie, though she does play the part of an extremely high maintenance woman uncannily well…
I mean when you're sixteen, you can sign a legally binding contract (might depend on the jurisdiction). Usually you have to be above the age of majority (18) or be emancipated before you can enter into a binding contract.
Not getting your life ruined by getting sued by a trillion dollar company sounds like a pretty good motivation
Speaking of being sue, I once found a hole in a healthcare system. I solicited legal advice and the response was overwhelmingly against reporting. The company didn’t not have a bounty program. So I shut up knowing my info is sitting there exposed. I wonder what’s the consensus on this? Do people normally report it or not?
On some forms I was advised to report hippa complaint. I can’t remember if I did. This was 10 or more years ago.
I think my take would be to report it anonymously (via support/marketing/etc.) and in the report strongly encourage them to create a path for security reports (either via a bounty or just a security email). When fixed or actually dismissed post publicly on anonymous channels, with a section saying how they could have created a channel for these reports.
>Do people normally report it or not?
if you are reporting it for moral/ethical reasons, but are worried that the company will have a shitty response, report it anonymously to the company and any relevant regulatory bodies.
but most companies, even without a bounty program, are happy to receive reports. a lack of a bounty program isn't really an indicator of how they will treat reporters, as most companies are ill-equipped to have a proper bounty program.
otherwise, and i know this is an unpopular opinion here, but i'm a big advocate of just posting the exploit publicly when a company makes it hard/scary to report it to them.
My issue was knowing that reporting would allow them to possibly discover my identity. Discovering the vul in itself may instead be considered unauthorized access.
don't worry you are popular with me
i think the answer is simple: they're a kid, and microsoft bullied them.
reminds me of a former job in which my goodbye letter was heavily editorialized...
Would love to hear more.
I will tell my story: I found a huge accounting error that allowed for several employees to embezzle funds. I disclosed it to the main stakeholders and some people went to jail. The accounting team still refused to acknowledge that the mistake was theirs, so I left on principle.
A few weeks later, I was talking to an old colleague and they revealed that the story that came from stakeholders was that the accounting team found the error, and I left in disgrace (despite that not being my department?).
So if you find an embarrassing mistake and you do not control the narrative, you have to proceed very carefully.
Had one like that. I quit and they blamed me for a project failure after the fact. That pissed off a couple of colleagues who thought it might happen to them. They also quit leaving them entirely without a software team. Set them back ten years because they didn’t make the market in time.
Write everything down and make sure everyone knows you’re writing it down. Saves a lot of hassle like that.
> the story that came from stakeholders was that the accounting team found the error, and I left in disgrace (despite that not being my department?).
That sounds like defamation to me.
_I am jack's complete lack of surprise_
This is a company that has engaged in heavy monopolistic behavior and violating US anti-trust laws with reckless abandon. They have even pioneered the infamous "embrace, extend, extinguish" strategy.
M$ will never change. The modus operandi is always the same, regardless of which ever lame MBA douchebag is running the shit show.
If we want to actually care , rather than spew off platitudes, break up big tech.
> awarded $5000
It's a little perplexing. Of course it's always a controversial topic since it's difficult to value an exploit, but whenever we read about these online, which probably goes through some survivorship bias, they seem pretty low.
On https://www.microsoft.com/en-us/msrc/bounty it seems the top is $100,000 or $250,000 depending which program this counts under.
What does HN think? Why would it be only $5000?
Every HN post regarding security exploits inevitably results in some comment saying the bounty is too low. I find it helpful to read previous comments by tptacek regarding bug bounties and market values:
https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...
This one probably has the best summary:
https://news.ycombinator.com/item?id=43025038
This is correct. I've made well over six figures over the last couple of years through bug bounty programs. I wouldn't spend months finding one bug. It's usually days or a week or two max.
$5,000 is actually pretty normal for a critical. I think the most I received for one bug was around $10,000, but this is the exception.
AI has also ruined the market. I'm a security consultant (where I make most of my money) and stopped bug bounty once AI slop reports made it impossible to actually get anything triaged in a timely manner.
Something else many don't know is that with all of the major platforms, customers can see the bugs before they are even triaged by the platform team. Many companies are now taking advantage of the long triage times and fixing the bugs before the researcher can get paid. It's then marked as a duplicate and the researcher gets nothing.
> Many companies are now taking advantage of the long triage times and fixing the bugs before the researcher can get paid. It's then marked as a duplicate and the researcher gets nothing.
Isn’t that just a speedrun-encouragement for selling the exploit not to the one offering the product, but an attacker offering more (in this case, >$0 is not difficult to exceed) instead?
If money is your motivation, you probably wouldn't consider disclosing the bugs in the first place.
As I understand it, bug bounty awards are a rough proxy for "would nation-state actors be able to exploit this for operational purposes without getting caught".
Zero-click iPhone exploits that affect the current OS and also previous ones are worth hundreds of thousands.
thats the true market value of bug bounty awards
the unilaterally set awards by the affected corporation are far lower and based on the price of the researcher’s liability
Try 10-20 million USD for a zero click iPhone exploit.
I personally think it sets a bad precedent: you want to broadcast that valuable info will be treated as such. A hacker in the future will see this, compare the "precedent" price to the black market, and not even ask Microsoft the next time.
I think HN, like a lot of people in this industry, have strongly skewed perception of the actual importance of these bugs.
You could view the bounty prices as market evidence that most of this is rightfully treated as nothingburgers. I.e. the alternative to paying $5000 to some random person for this class of vulnerability research is not risking a trillion dollar hack the next day - it's just risking shmaybe some kerfuffle down the line, followed by fixing it through normal triage process. The bounty program is as much marketing as security, and $5000 is probably about the right price for marginal effort into sustaining the "we are treating security seriously" message.
In a way, the very existence of those bug bounty programs in large companies is evidence they don't see a reason to treat vulnerabilities seriously enough to proactively find and fix them in-house.
If security vulnerabilities would be anywhere serious as most commenters on-line seem to think, companies would pay hundreds of thousands for serious vulnerabilities, just to save a day before they get hit by them - on top of spending millions in-house to try and stay ahead of the attackers.
But they don't. Because most exploits are inconsequential and/or aren't being exploited much.
$5k is a literal penny for Microsoft. Give the kid $100k.
Per their market cap...
3.85 billion is actually closer to a "penny" for Microsoft.
Operating income for Microsoft was $128.5 billion in 2025.
https://www.microsoft.com/investor/reports/ar25/index.html
What's worse is the alternative is winding up like Aaron Swartz... (all he did was scrape PDFs for mostly public funded / tax funded papers) Which is even worse, I think I'd be glad to receive $20 for Starbucks instead of being legally chased for showing them they messed up.
There should really be laws for protecting security researchers who produce 0 harm and divulge / share a vulnerability with a service provider. I'd rather the floor be getting no money AND not going to jail or being sued.
"all he did was scrape PDFs for mostly public funded / tax funded papers"
He wasn't a security researcher. He broke into a room and used equipment to steal information. It wasn't just 'tax funded papers'. Companies invested millions of dollars into some of this research.
We shouldn't support theft and he should have gotten some jail time/punishment for it.
"There should really be laws for protecting security researchers who produce 0 harm and divulge / share a vulnerability with a service provider. I'd rather the floor be getting no money AND not going to jail or being sued."
Too many 'security researchers' demand money or threaten to release the vulnerabilities.
I don't know anyone that got into trouble going through a legit bug bounty program.
I have not done a deep dive so I could be wrong, weren't these papers published (or soon to be)? Wasn't the one and only group with negative effects the Journals that could extract a fee? Especially with researchers usually happy to send a copy? This wasn't someone stealing trade secrets they weren't even secret.
Microsoft's bounty program and payouts are known to be pathetic, see that nightmare eclipse situation
The young researcher is a teenager. Getting your name out there, clout/influence/fame, is worth $1M to them.
It’s almost like most of you people have already forgotten your teenage years. And it shows. Doesn’t matter the decade or generation. A teenager , likely living with parents or grandparents, will always prioritize intangibles.
A mere 5 bands for full editorial control is quite literally peanuts for M$. It’s yet another case of abusing free labor, unfortunately.
You assume that whoever was responsible for the implementation of this feature cares in the slightest beyond “it seems to work”.
His comment didn’t assume that at all.
Physical security vendors usually stop exactly there. I can't count how many vendors responsible for badge systems shut off the Windows Firewall just because it’s easier than adding a rule.
Honestly that series of blog posts that came across here semi-recently about what an utter disaster the Azure team is (through little fault of their own, mind you), combined with a lot of things I've read about the development behind Halo Infinite, have convinced me Microsoft is to avoided whenever possible. It's not even that they're too big to care, though that is an issue. It's that organizationally they are SO reliant on short-term contractors and junior devs plugging away with copilot, along with an ever more exhausted group of seniors who actually work for them but are continually disempowered, that I think it's safe to say anything they manage to ship is a minor miracle if it works at all, and it's certainly not going to have any guarantee at all of solid software engineering fundamentals.
And to be clear, this is not an issue with them using contractors, overseas or otherwise, or with their senior dev staff, or even with AI really. It's an issue with them organizationally being so incredibly penny-pinching, and so dedicated to shipping new shit versus fixing anything long term, constantly chasing new revenue and letting their existing offerings rot.
If a Microsoft product is good nowadays, it is literally a miracle.
You assume there is any actual reason to care more than that in this case.
There is an internal library at Microsoft that reliably avoids all these JWT problems - Microsoft Identity Service Essentials (MISE). Adopting MISE and upgrading to the latest versions of it have been part of the Secure Future Initiative (SFI) that can be read about in the news of previous years. Unfortunately it sounds like the service team intentionally deferred the compliance alerts they will have received.
There is so much work to do for SFI that is still being ignored. The only way that some of these services will update is by being the target of the red team, security researcher, or threat actor.
And all of that is definitely happening.
That being said, just last night I observed that the identity team is now opening up agent-assisted PRs against individual service team repos to force MISE adoption and upgrade to the latest version and best practices. I think that's a great thing because many individual service teams simply lack the bandwidth or knowledge. Prior to GenAI availability I wasted many cycles on this kind of work. While GenAI made it easier - internal source documentation still does not unambiguously address every use case. So having the identity team drive this now with the help of agent sessions initiated by them is great.
> {"alg":"none","typ":"JWT"}
I don't know how this ever became a thing that was allowed into the spec and then picked from the spec and implemented in various implementations.
Yea, and it was a big issue years ago. I can't believe there is any modern implementation that has it.... or another way to put this is, how old is this damned implementation they are using?
It should be retroactively removed the spec, and all implementations should remove the "none" algorithm. It is a huge security hole.
wow I am so surprised to hear once again how JWTs are terrible
https://www.howmanydayssinceajwtalgnonevuln.com/
Someone not verifying the signature at all is not a mistake where you can blame the JWT spec itself.
They did verify the signature, and it was correct according to the "none" algorithm.
Argh, I missed that it actually uses the "none" algorithm. Yeah, the existence of that option is extremely dumb and it shouldn't be possible to use that. I misread the post and thought it was a regular JWT, but they simply didn't validate it.
“Works as designed.”
JWT is complicated.
Complexity is a spec failure in security issues.
It's that simple.
Does this extend to OIDC? I’m not knowledgeable on the topic but it uses JWT right? Is it also prone to poor implementation? If you just error on alg=none does that solve it?
Idk if that's not too much of an oversimplification, maybe more like JWTs are an indicator/enabler of architecture level bugs?
Other commenters are suggesting you can’t blame the spec for end implementation mistakes, except that’s one of the many issues - JWT being so error-prone is a problem.
I use JWT just for handling of tokens, because it’s so well supported, but I won’t use it for anything more than token storage _because_ it is so vulnerable to mistakes.
The fact that mistakes are so easy to make is indicative of poor design in the spec itself.
What is Antares? Can't find anything related to it except for the 1B model, which does not seem to be capable of autoresearch.
UPD: It's his personal bot.
> I also started building AI into how I hunt, which led me to develop Antares, my personal AI hackbot.
Pretty confusing, there is a Cisco Antares small language model... https://blogs.cisco.com/ai/introducing-antares-the-most-effi...
Guess everything is just going to be named after stars for awhile.
You’re going places kid :) keep up the good work
Possibly jail if the wrong people get upset.
Well college isn’t worth it according to the Thiel crowd so maybe this will be better.
Great way to use up that 6-10 years of vacation and sabbatical time.
This kid is 16?
Can’t wait to see what he’s up to in 10 years.
Seriously, who is responsible for "none" JWT tokens. It has caused so many critical security bugs over the years.
This is a pretty typical example of the security posture of microsoft, and yet people continually buy their arguments that open-source and therefore auditable alternatives are inherently less secure than their "trust me bro"
Geohot vibes
Wait until someone does that to your favourite cloud provider's customer data.
Entra has already had vulnerabilities where a user in one tenant could access global admin of every other tenant. Somehow everyone didn't immediately move off Microsoft's cloud platforms.
Next time you find a bug like that, offer it to the black market, you could make millions instead of measly salty peanuts
I bet someone already did that and didn't disclose it.
Now some blackhat somewhere can't afford their monthly Lamborghini payment.
If a 15yo can find it
Only $5K when you saved them millions. Pretty cheap!
You should have.
> Two quick notes first. The impact I describe is hypothetical. It’s what an attacker could have done with this access, but luckily I found the bug instead, reported it, and never touched any customer data or PII.
I am the last person to judge someone for using AI to help them write a blog post, but what I wonder is: Do people not read what the AI produces before putting their name on it, or is the AI writing style not obvious to some people, or do they just not care that it's obviously AI and bad style?
That.. looks like a normal sentence to me, and very probably one of the ones Microsoft required they add.
Did you give the article a once-over beyond that? It’s one of the decidedly not-AI lines.
It's not a "normal" sentence and is quite clearly produced by an LLM, it's a typical Claudeism so probably that. The entire post is also flagged by Pangram, so OP is correct.
Prefixing saying something with "Two quick notes first" is extremely common AI meta commentary. You may be right that it is something Microsoft insisted he put at the top, which would also explain the weird style.
From where do you think AI learned that?
Learning a phrase from humans doesn’t prevent a model from overusing it or making it a recognizable stylistic habit.
Weird user preferences during RLHF. Also how we got bulleted lists and emojis everywhere.
All AI semantic tendencies were “learned” from us. That doesn’t change anything.
All of the above, though mostly the last one.
Explain why easily 40% of HN posts are about AI, and when people actually _use_ AI for this task, they are vilified for (allegedly) using it.
He’s fine to use the AI for bug hunting but nobody wants to read AI’s bullshit slop and tone
I’m as tired of slopped up text as the next person but the example OP gave really isn’t that bad
I pasted the thread into ChatGPT and here is what it would like to tell you:
The comment explicitly says, “I am the last person to judge someone for using AI to help them write a blog post.” The criticism is about the quality of the published prose and whether anyone reviewed it before putting their name on it.
You can dispute whether that sentence is bad or whether it indicates AI authorship. But “why criticize unedited AI prose when HN talks about AI so much?” doesn’t identify a contradiction. Discussing a technology doesn’t imply endorsing every use of it, and criticizing its output isn’t the same as vilifying someone for using it.
— GPT-6.1 Sol