> To the companies: You don’t need HackerOne anymore. The tokens to build your own in-house platform cost less than single year of HackerOne.
You know, the biggest thing that HackerOne delivers is a universal payments system that requires absolutely no efforts from companies. Have you tried to manually pay hackers from around the world? It is a laborious process involving trying to find what providers are compatible and what forms of money go where. It is extremely taxing to handle this. HackerOne provides real, tangible value in not making people think about how precisely to pay a hacker and in what currency. No amount of tokens solve the accounting problem, and it is foolish to imply otherwise.
This and the pre-triage are the only reasons we even use a bug bounty platform.
If paying out bounties was easy I would do it all via email; but as you said it’s almost impossible to do (unless you are maybe bigcorp and have a team just for that)
> universal payments system that requires absolutely no efforts from companies.
Indeed. I use a third party company (not HackerOne) to handle our bug bounty and the primary reason is so they handle all the payment hassles, I don't need to be involved. They also handle all the screening for false positives, which in the AI age are exploding. I also don't want to deal with that.
In general I lean towards building in-house, but this is one area I'm happy to oursource all the busywork.
I know Joel well and think a lot here is both accurate and well written. I led the Yahoo bug bounty program from 2023-2024 and was involved in it from about 2021. A major event that this glosses over is Covid which also happened right around this time as well. Covid killed travel (and budget) which in turn made it impossible to do the live events. A lot of companies ended up shifting to virtual live events which just never delivered on the same value, scale, or impact.
When COVID restrictions were lifted, travel and t&e budgets just never returned. Layoffs started happening and what were lavish, expensive events just couldn’t happen anymore. Hackerone charged for and likely made a lot of money on these events. I think a lot of what is talked about in the article is true but I think Covid is a big part of the why that led to it.
It is also worth remembering that the cost of travel itself, and the cost of venues itself has also increased substantially. As well as associated costs such as catering and insurance.
So in-person events have issues from both sides, those attending and those hosting.
You also do not mention corporate policies. Under pressure from investors, their employees and sometimes their home-countries, many corporates have also introduced environmental policies. So if you want the company to pay for your flight, you not only have to justify it financially, but you have to justify it environmentally too.
Sending the sales team on a paid vacation to a tropical paradise while the engineering product flounders is such a perfect representation of corporate rot it sounds like something out of a Mike Judge movie
Presidents club is a standard way to reward top performing sales reps across many industries. It doesn't indicate anything other than the company is trying to reward and retain their top sales reps. Engineers who find it distasteful should be happy to know engineers typically get way more equity than sales reps.
This one looks shocking on the outside indeed! It's however a sales HR practice: lower salary and commission, but use the alluring treat that top performers will have a special exclusive trip at the end of the year. Many are crazy for it.
It's more a reward for a competition-style work mindset.
I'm confused at the way they promoted it. Is there any way someone outside the company reading a Twitter post would consider this a positive thing for the product to be told what incentives the sales team get?
One of them, a remotely triggerable DoS vector got downgraded in severity. I got a token payment from the company, and 7 years later, it is still not marked as resolved.
Yeah I reported a j-frog vulnerability to Anthropic. It was downgraded to “informative” and they asked me to prove that I could exfiltrate data. I replied that exfiltrating data is against their program’s safe harbor policy and they just never responded. 2 months later the claude code source code leaked.
One of my only bug bounty payouts was a DoS against a site via their customer query engine. I was quite proud of it, and was relieved when they actually paid out a token amount.
It took down the entire application for all users and tenants, not just the tenant submitting the poisoned query.
I don't remember how much I was paid, a token amount for sure, but I was happy with any amount because it was a hobby and any payment was good for the CV.
Not only was there significant personal liability, but there had been multiple instances of hackers being criminally charged and sentenced to jail time for finding and reporting security vulnerabilities prior to this.
I don't think this is true, although it's a very commonly-held belief. Dan Goodin (I think?) wrote an article about this a long time ago, and was only able to come up with a few examples, and none of them fit this fact pattern.
What is true is that it is much less legally risky to test someone else's computer than it was 10-15 years ago. People forget that's what you're doing when you look for web vulns! The DOJ has had a norm over the past ~many years not to prosecute good-faith vulnerability research, even though strictly speaking it contravenes CFAA directly. But "risky on paper" is the most you could say about doing that kind of testing back in 2010.
I can think of 4-5 other situations from around that era (~2012) where people were at least charged and needed a lot of help to navigate the legal proceedings to avoid jail time.
In 2010 it was more than risky on paper.
2017-2018 is well into the established era and probably even the golden age of bug bounties when a lot of corporate and judicial thinking re: white hat cybersecurity had been shifted.
All good things don't last forever. A organization or company lasting forever with the same goal/mission while using the same methods is a statistical anomaly.
I am in this space. The reality is that the margins for a Bug Bounty Hunting platform are not good, triage is very expensive specially with all the AI slop that gets submitted now. You can hide it for a long time with VC money, but they need to diversify their product line to continue growing and compete against the AI pentest compagnies (which themselves will also diversify as AI pentest becomes a feature and not the whole product).
I agree. They have quality data to build an effective AI pentest product that is good enough, and they already have a good offering to bundle that into and satisfy enterprise demand.
Up and coming AI pentest companies need to have an exceptional product to get a chance to stand on their own and penetrate the enterprise market, otherwise their best scenario is an acquisition to get bundled into an established platform.
See also, the influx of spooks into various hackerspaces during the Snowden/Assange era. I truly believe there was an effort to subvert these communities, and thats what happened.
I once interviewed there, and it was the weirdest interviews of my life.
They literally asked me to prepare on the company mission and values.
The first round was about generic stuff where nothing much was asked. And ironically, despite transparency being their core mission, they didn't tell me I was rejected until I emailed them about a week later.
From the customer point of view: at a fortune500 I dealt a LOT with h1 (it was never H1) in the early days. Then we got a CISO who was mostly a showman. And at some point (which match the changes in leadership at h1 the article describes) the reports became all garbage and leadership (CISO and CTO) would talk about h1 hackathons with "top hackers flown from all over the world". Such a joke. The end result of those hackatons were 200 "internal host discovery" that were already reported internaly and teams always dismissed as "not worth fixing" and a single attack vector, usually from a brand new acquisition that was still going trhu onboarding. Pretty much never nothing relevant or actionable.
Last time I reported a DoS bug to HackerOne, the company behind the bounty tried incite me to commit a crime against them by DoS'ing their servers using the hack I had reported in detail!
I literally showed them their server taking over a minute to respond to my request. I even showed how the delay increased proportionally to the message size... Clearly doing more processing; classic DoS vulnerability... Doesn't leave much to the imagination! But they said they would not pay me anything unless I actually proved that it scaled and caused disruption of their service!
It seemed like they were baiting me into incriminating myself for a crime that they wanted me to commit against them. It's not even the first time that I've been baited by a software company into committing a crime against themselves. I never took the bait though.
To be fair “we will compensate you if you do X” sounds a lot like a contract so you’d probably be just fine in court. (Though likely wise to avoid the chance of a legal headache)
I don't understand the controversy at the heart of this post. H1 stated they don't use reports to train LLMs. Then they revealed they were using LLMs to triage reports based on previous reports. These two facts are not necessarily incompatible. It's entirely possible to use an LLM with a db tool installed to triage reports without using the body of the reports as training fodder. The article doesn't give any evidence that this was not the case. It sounds to me more like the OP already disliked H1 (for its sales practices and general enshittification) and the LLM issue was a convenient excuse to make a clean break.
> I don't understand the controversy at the heart of this post.
Did you miss this part from the article:
> They switched from talking about bug bounty programs, live hacking events, and how they could help you stay secure, to promoting their in-house AI security product and continuous security monitoring tool.
notably the in-house AI security product is trained on existing bug bounty reports.
> It sounds to me more like the OP already disliked H1 (for its sales practices and general enshittification) and the LLM issue was a convenient excuse to make a clean break.
that's pretty harsh to say when OP provided some very valid reasons, imho speaking as someone who's used HackerOne for over a decade.
And also the idea that H1 "training" models based on bug bounty reports is kind of a silly concern; frontier models have commoditized most of what was reported on H1, even at higher quality levels. H1 itself is a nonfactor.
> Co-founder Michiel Prins was allowed to leave the HackerOne dungeon to perform damage control with this absolute banger of an AI slop response: [...]
Wow, it's like he prompted for the most stereotypically AI response possible. There's a tired trope in every sentence going on for four whole paragraphs! I originally quoted it too but thought better and decided to snip it out because I'm pretty sure it would get my account flagged by HN's AI detection algorithm...
Imagine doing this article as a thorough writeup to provide feedback, rewriting this for like an hour before you post it.
And then you get an AI slop response like that in return where you can't even tell whether it was just a CEO not giving a damn...or a standard dumb chat bot with a stupid response.
I'm not sure if founders are aware that these are tipping points in customer care where the people that care about your product and ecosystem will leave your company for good, and you're irreparably damaging your own reputation.
If I were OP I'd never ever touch anything with a 10ft pole that the founders will build in their lifetime, and I'd warn everyone I know in the community about it.
That's the damage they're doing with these AI optimizations to themselves.
There's a reason why everyone starts to hate your company right after your stupid chatbot was introduced.
tl;dr which begins 3,000 words in: employees were noticed to be leaving and it’s because a “fine tuning from user submissions” ai psychosis of yesteryear, except it’s amusingly happening in 2026 still. Investigation into the veracity of the claims.
From what I've seen in the bounty-related subreddits, AI is flooding bug bounty inboxes with low-value or meaningless reports, or straight-up hallucinations when people use smaller models (to turn a profit, you make lots of low-value bug reports and see who pays out).
This has a negative effect on humans doing their work with or without LLMs: curl shut down their bounty program, and GitHub just announced they're "restructuring" theirs.
The author of this post also makes a case that HackerOne hasn't been honest about LLM training and use, either to hackers or to their own staff.
You’re surprised that workers don’t like their work being used to remove the need to pay them for it in the future? Your idea of time saved for the worker is for them to lose their livelihood without compensation
I can understand coming down on either side of the question of whether these AI reports save or waste time. I cannot understand being surprised or ignorant about the existence or high level beliefs of either side.
> To the companies: You don’t need HackerOne anymore. The tokens to build your own in-house platform cost less than single year of HackerOne.
You know, the biggest thing that HackerOne delivers is a universal payments system that requires absolutely no efforts from companies. Have you tried to manually pay hackers from around the world? It is a laborious process involving trying to find what providers are compatible and what forms of money go where. It is extremely taxing to handle this. HackerOne provides real, tangible value in not making people think about how precisely to pay a hacker and in what currency. No amount of tokens solve the accounting problem, and it is foolish to imply otherwise.
This and the pre-triage are the only reasons we even use a bug bounty platform.
If paying out bounties was easy I would do it all via email; but as you said it’s almost impossible to do (unless you are maybe bigcorp and have a team just for that)
> universal payments system that requires absolutely no efforts from companies.
Indeed. I use a third party company (not HackerOne) to handle our bug bounty and the primary reason is so they handle all the payment hassles, I don't need to be involved. They also handle all the screening for false positives, which in the AI age are exploding. I also don't want to deal with that.
In general I lean towards building in-house, but this is one area I'm happy to oursource all the busywork.
Many solutions nowadays.
https://www.payoneer.com/ is one of them. Of course this one is a bit racist depending on which contry you were born in.
Just pay them in stable coins. That's a solved problem.
I know Joel well and think a lot here is both accurate and well written. I led the Yahoo bug bounty program from 2023-2024 and was involved in it from about 2021. A major event that this glosses over is Covid which also happened right around this time as well. Covid killed travel (and budget) which in turn made it impossible to do the live events. A lot of companies ended up shifting to virtual live events which just never delivered on the same value, scale, or impact.
When COVID restrictions were lifted, travel and t&e budgets just never returned. Layoffs started happening and what were lavish, expensive events just couldn’t happen anymore. Hackerone charged for and likely made a lot of money on these events. I think a lot of what is talked about in the article is true but I think Covid is a big part of the why that led to it.
> travel and t&e budgets just never returned.
It is also worth remembering that the cost of travel itself, and the cost of venues itself has also increased substantially. As well as associated costs such as catering and insurance.
So in-person events have issues from both sides, those attending and those hosting.
You also do not mention corporate policies. Under pressure from investors, their employees and sometimes their home-countries, many corporates have also introduced environmental policies. So if you want the company to pay for your flight, you not only have to justify it financially, but you have to justify it environmentally too.
Sending the sales team on a paid vacation to a tropical paradise while the engineering product flounders is such a perfect representation of corporate rot it sounds like something out of a Mike Judge movie
Presidents club is a standard way to reward top performing sales reps across many industries. It doesn't indicate anything other than the company is trying to reward and retain their top sales reps. Engineers who find it distasteful should be happy to know engineers typically get way more equity than sales reps.
This one looks shocking on the outside indeed! It's however a sales HR practice: lower salary and commission, but use the alluring treat that top performers will have a special exclusive trip at the end of the year. Many are crazy for it.
It's more a reward for a competition-style work mindset.
I'm confused at the way they promoted it. Is there any way someone outside the company reading a Twitter post would consider this a positive thing for the product to be told what incentives the sales team get?
if i'm working for a company i sure hope they're sending their sales reps to a tropical paradise, and if they don't, i hope i don't own any equity.
I reported some exploits on hackerone.
Most got dismissed.
One of them, a remotely triggerable DoS vector got downgraded in severity. I got a token payment from the company, and 7 years later, it is still not marked as resolved.
I doubt my situation is unique.
Yeah I reported a j-frog vulnerability to Anthropic. It was downgraded to “informative” and they asked me to prove that I could exfiltrate data. I replied that exfiltrating data is against their program’s safe harbor policy and they just never responded. 2 months later the claude code source code leaked.
One of my only bug bounty payouts was a DoS against a site via their customer query engine. I was quite proud of it, and was relieved when they actually paid out a token amount.
It took down the entire application for all users and tenants, not just the tenant submitting the poisoned query.
I don't remember how much I was paid, a token amount for sure, but I was happy with any amount because it was a hobby and any payment was good for the CV.
I reported a security bug, it was all processed very quickly and I got paid. I doubt my situation is unique.
I think it would be the individual companies slowing things down, not the platform.
Most bounty programs won't pay for DoS at all.
Not only was there significant personal liability, but there had been multiple instances of hackers being criminally charged and sentenced to jail time for finding and reporting security vulnerabilities prior to this.
I don't think this is true, although it's a very commonly-held belief. Dan Goodin (I think?) wrote an article about this a long time ago, and was only able to come up with a few examples, and none of them fit this fact pattern.
https://news.ycombinator.com/item?id=16642155
What is true is that it is much less legally risky to test someone else's computer than it was 10-15 years ago. People forget that's what you're doing when you look for web vulns! The DOJ has had a norm over the past ~many years not to prosecute good-faith vulnerability research, even though strictly speaking it contravenes CFAA directly. But "risky on paper" is the most you could say about doing that kind of testing back in 2010.
https://m.slashdot.org/story/159162-- example circa 2011
I can think of 4-5 other situations from around that era (~2012) where people were at least charged and needed a lot of help to navigate the legal proceedings to avoid jail time.
In 2010 it was more than risky on paper.
2017-2018 is well into the established era and probably even the golden age of bug bounties when a lot of corporate and judicial thinking re: white hat cybersecurity had been shifted.
Maybe in the US but see eg germany: https://binsec.wiki/en/pentesting-guidelines/pt-legal-framew...
Doubt, I’d argue it’s the opposite given the term “vulnerability research” is being overloaded to include things such as F12 on a school website.
It got the executives it paid for
All good things don't last forever. A organization or company lasting forever with the same goal/mission while using the same methods is a statistical anomaly.
What is the corrupting force?
I am in this space. The reality is that the margins for a Bug Bounty Hunting platform are not good, triage is very expensive specially with all the AI slop that gets submitted now. You can hide it for a long time with VC money, but they need to diversify their product line to continue growing and compete against the AI pentest compagnies (which themselves will also diversify as AI pentest becomes a feature and not the whole product).
I agree. They have quality data to build an effective AI pentest product that is good enough, and they already have a good offering to bundle that into and satisfy enterprise demand.
Up and coming AI pentest companies need to have an exceptional product to get a chance to stand on their own and penetrate the enterprise market, otherwise their best scenario is an acquisition to get bundled into an established platform.
money happened. it corrupts all. once there is enough of it going around people lose all senses and just want more.
See also, the influx of spooks into various hackerspaces during the Snowden/Assange era. I truly believe there was an effort to subvert these communities, and thats what happened.
Bug bounty programs were overrun with low-effort slop nearly a decade before LLMs were introduced; I can't imagine what they're like now...
You don't want to. It's exactly as bad as you think. I would say ~90% of reports are false now as opposed to ~40% before LLMs.
I once interviewed there, and it was the weirdest interviews of my life.
They literally asked me to prepare on the company mission and values.
The first round was about generic stuff where nothing much was asked. And ironically, despite transparency being their core mission, they didn't tell me I was rejected until I emailed them about a week later.
Honestly, you could sub the other big Bug Bounty platform for H1 in this post and you’d be still extremely accurate.
From the customer point of view: at a fortune500 I dealt a LOT with h1 (it was never H1) in the early days. Then we got a CISO who was mostly a showman. And at some point (which match the changes in leadership at h1 the article describes) the reports became all garbage and leadership (CISO and CTO) would talk about h1 hackathons with "top hackers flown from all over the world". Such a joke. The end result of those hackatons were 200 "internal host discovery" that were already reported internaly and teams always dismissed as "not worth fixing" and a single attack vector, usually from a brand new acquisition that was still going trhu onboarding. Pretty much never nothing relevant or actionable.
Last time I reported a DoS bug to HackerOne, the company behind the bounty tried incite me to commit a crime against them by DoS'ing their servers using the hack I had reported in detail!
I literally showed them their server taking over a minute to respond to my request. I even showed how the delay increased proportionally to the message size... Clearly doing more processing; classic DoS vulnerability... Doesn't leave much to the imagination! But they said they would not pay me anything unless I actually proved that it scaled and caused disruption of their service!
It seemed like they were baiting me into incriminating myself for a crime that they wanted me to commit against them. It's not even the first time that I've been baited by a software company into committing a crime against themselves. I never took the bait though.
To be fair “we will compensate you if you do X” sounds a lot like a contract so you’d probably be just fine in court. (Though likely wise to avoid the chance of a legal headache)
I don't understand the controversy at the heart of this post. H1 stated they don't use reports to train LLMs. Then they revealed they were using LLMs to triage reports based on previous reports. These two facts are not necessarily incompatible. It's entirely possible to use an LLM with a db tool installed to triage reports without using the body of the reports as training fodder. The article doesn't give any evidence that this was not the case. It sounds to me more like the OP already disliked H1 (for its sales practices and general enshittification) and the LLM issue was a convenient excuse to make a clean break.
> I don't understand the controversy at the heart of this post.
Did you miss this part from the article:
> They switched from talking about bug bounty programs, live hacking events, and how they could help you stay secure, to promoting their in-house AI security product and continuous security monitoring tool.
notably the in-house AI security product is trained on existing bug bounty reports.
> It sounds to me more like the OP already disliked H1 (for its sales practices and general enshittification) and the LLM issue was a convenient excuse to make a clean break.
that's pretty harsh to say when OP provided some very valid reasons, imho speaking as someone who's used HackerOne for over a decade.
link to H1's "continuous monitoring tool" for the curious: https://www.hackerone.com/product/h1-continuous-testing
And also the idea that H1 "training" models based on bug bounty reports is kind of a silly concern; frontier models have commoditized most of what was reported on H1, even at higher quality levels. H1 itself is a nonfactor.
> Co-founder Michiel Prins was allowed to leave the HackerOne dungeon to perform damage control with this absolute banger of an AI slop response: [...]
Wow, it's like he prompted for the most stereotypically AI response possible. There's a tired trope in every sentence going on for four whole paragraphs! I originally quoted it too but thought better and decided to snip it out because I'm pretty sure it would get my account flagged by HN's AI detection algorithm...
I wonder if that's the golden handcuffed founder equivalent of blinking out SOS in morse code?
Imagine doing this article as a thorough writeup to provide feedback, rewriting this for like an hour before you post it.
And then you get an AI slop response like that in return where you can't even tell whether it was just a CEO not giving a damn...or a standard dumb chat bot with a stupid response.
I'm not sure if founders are aware that these are tipping points in customer care where the people that care about your product and ecosystem will leave your company for good, and you're irreparably damaging your own reputation.
If I were OP I'd never ever touch anything with a 10ft pole that the founders will build in their lifetime, and I'd warn everyone I know in the community about it.
That's the damage they're doing with these AI optimizations to themselves.
There's a reason why everyone starts to hate your company right after your stupid chatbot was introduced.
human slop
tl;dr which begins 3,000 words in: employees were noticed to be leaving and it’s because a “fine tuning from user submissions” ai psychosis of yesteryear, except it’s amusingly happening in 2026 still. Investigation into the veracity of the claims.
I'm surprised someone could get upset at AI triaging of bugs which would save everyone time.
From what I've seen in the bounty-related subreddits, AI is flooding bug bounty inboxes with low-value or meaningless reports, or straight-up hallucinations when people use smaller models (to turn a profit, you make lots of low-value bug reports and see who pays out).
This has a negative effect on humans doing their work with or without LLMs: curl shut down their bounty program, and GitHub just announced they're "restructuring" theirs. The author of this post also makes a case that HackerOne hasn't been honest about LLM training and use, either to hackers or to their own staff.
You’re surprised that workers don’t like their work being used to remove the need to pay them for it in the future? Your idea of time saved for the worker is for them to lose their livelihood without compensation
I can understand coming down on either side of the question of whether these AI reports save or waste time. I cannot understand being surprised or ignorant about the existence or high level beliefs of either side.
I'm sorry you don't know the difference between training, fine tuning, and context. But definitions matter especially in legalese.