Against better judgement, I ran it on my repo. Poking at the "Ask anything about the report" bot, it penalizes me heavily for not having "positive tracked outcomes", as I don't report back to Claude what the outcome is. It wanted me to do this:
"At the end of research-heavy sessions, add a tiny closure note:
Decision:
Use CacheMountStore with registry/local/GHA backends.
Why:
GHA cannot expose the same content.Ingester path, local import has discovery issues, registry can resolve by tag.
Proof / current artifact:
See files X, Y, Z. Subagent found A, B, C.
Next action:
Implement interface in package N. Do not add new cache-mount flags yet. Use mode=cache-mount on existing cache flags."
There's no need for this, Claude is not my task tracker.
It also penalized me for:
- Using Claude to introspect a codebase as throwaway work and ` not close with a crisp “acceptable / risky / copy this / avoid this” decision this session.`. No action was needed, it wasn't an actionable session!
- Using Fable for code review. `Several BuildKit subagent sessions produced research reports, but the transcript does not show whether those findings became implementation choices.` Yes, Fable launches subagents to code review. Valid findings get turned into fixes or WONTFIX.
- `The Cloudflare CI upload failure got narrowed well, especially after you supplied the exact timestamp and challenged the /v2 routing recommendation, but the session ended without a chosen next diagnostic owner or action this session.` There's no need to pick a next diagnostic owner when I'm a solo founder with no employees.
I'm choosing not to attach a report to my application. As a side note, nobody needs to see "my greatest crashout", it's embarrassing enough for me to see it for myself, much less show others.
I had the exact same experience. Claude is not my validation engine, I do not tell it when it has completed. I verify that in the code/project and end the session.
I also got this,
> A useful next habit is to end each correction with a concrete acceptance test, owner artifact, or stop condition: “write it into PROGRESS.md,” “make nix run .#bench fail until this is real,” “rerun this exact command,” or “do not proceed until these two choices are explicit.”
> You already do this well in the biggest penance sessions. Apply it to the smaller ones too.
Which I have found to be counterproductive in my personal work. Current models can generally infer acceptance tests of this level of granularity (not true for larger project-level prompts, but those don't produce good enough code for me yet -- even with specific acceptance criteria).
I also got penalized for using claude in read-only mode for the same validation reason?
> For read-only work, end with one of:
“turn the top finding into a PR-sized plan”
“mark these as accepted/rejected/deferred”
“write a cleanup checklist”
“give me the exact command I should run safely”
“stop, no action recommended”
No thanks, I'm literally just exploring the codebase. I don't want any of these.
> Based on Paxel’s own site, 1.2 million+ coders have so far uploaded their reports to YC.
I checked the Paxel website and it says this:
> So far, 1,543,553 sessions have been uploaded and analyzed.
The count is for sessions, not coders. I assume the tool uploads a lot of sessions from each person who uses it.
That’s a large number, but it’s not a million different people. I am surprised that so many people think it’s a good idea to download a run a program which gathers up their coding sessions and submits information about them.
I could be mistaken, but isn't Paxel a tool that YC *itself* built to understand how founders/applicants apply AI? If so, this feels less scary than some of the comments (i.e. not 3rd party).
Also, not trying to take shots, but should the title be "I got into YC Startup School by hacking it" instead? Isn't that different than the main YC program?
All in all, you did them a solid by finding and responsibly disclosing. Nice job.
For the longest time, the YC application included the question "What was your greatest (non-computer) hack?". They have always liked people who think of ways to work around existing systems. So it's no surprise that their response was positive. Also, all the principals are ex-founders, mostly engineers, who totally understand hacking culture.
Sadly, it looks like they took that question off the application though.
So let me get this straight... technical founders are providing limitless access to their IP, and this is supposed to be a positive indicator that they would be a good founder? Thats wild
Am I misreading this, or is ycombinator running a tool that transfers IP from an applicant’s machine to OpenAI and then transfers the generated summaries into their own servers. And that tool has some “webapp security 101” level vulnerability in it.
It doesn’t upload code (it’s not meant to, anyway); it uploads coding agent transcripts. It’s meant to be a fun app like Spotify wrapped - that’s how it was described when presented internally.
The "automatically scan your entire system" is what's worrying. TFA:
Paxel has an incredible UX, everything it does is handled from a single one-liner curl -fsSL https://paxel.ycombinator.com/upload.sh | bash. It will automatically scan your entire system, prompt you for questions, and handle the upload — you don't have to worry about a thing after.
Though, as supply chain attacks have demonstrated, who knows what a simple tool like xz is even doing on my machine, at this point.
Sure - though that’s all generated by the coding agents. So it shouldn’t be uploading original hand-written code (unless the coding agent is running on a legacy codebase).
Obviously it’s not something everyone is going to be comfortable sharing, but YC is transparent about what is shared and how it’s used.
Edit:
I now see you have to explicitly run it in every directory you want shared - i.e., it doesn’t t scan your whole disk looking for stuff to upload. So there’s every opportunity to limit what it uploads to only what you want to share.
It's very common nowadays to claim that something doesn't have access to X when it does in fact have system level access to X, but the application code just doesn't use X (or a prompt is being asked not to use X).
Trustworthy people will ask you to give them the right amount of access, they won't ask you for infinite access and ask you to trust them that they won't use it. Would you trust a contractor that asks for root, and your master password when they could have requested a specific user or some specific files?
Sure - and people are able to audit what the software is doing and call it out if it’s operating in a way that’s unintended or inappropriate, just as has happened here.
Edit:
I’ll add this point as people always seem to want to raise the spectre of YC or the AI platforms stealing people’s code or ideas via these sorts of tools.
The value of any code or ideas is immaterial compared to the value of identifying talented people who can then be offered opportunities such as attending Startup School, being funded by YC, getting jobs at YC-funded companies or YC itself.
It’s one of pg’s oldest dictums, that ideas alone are worthless, and only working products and the people that build them are worth anything. Nobody at YC thinks any differently about that these days.
>>Sure - and people are able to audit what the software is doing and call it out if it’s operating in a way that’s unintended or inappropriate, just as has happened here
1. Yes you can audit client side and network interface, but not server side.
2. does 'curl $URL | sh' invite users to audit? Or to trust blindly? It's the equivalent of presenting a contract with 'sign here' stickers and pushing you to sign it quickly like it's nothing. Does not signal trustworthiness.
>I’ll add this point as people always seem to want to raise the spectre of YC or the AI platforms stealing people’s code or ideas via these sorts of tools.
>The value of any code or ideas is immaterial compared to the value of identifying talented people who can then be offered opportunities such as attending Startup School, being funded by YC, getting jobs at YC-funded companies or YC itself.
>It’s one of pg’s oldest dictums, that ideas alone are worthless, and only working products and the people that build them are worth anything. Nobody at YC thinks any differently about that these days.
I do agree and share this point, ideas are worthless. But this doesn't deal with ideas, it deals with code, which as a software startup I don't feel the need to convince you of its value and the risks of leaking it. If you think your code is worthless, close your startup.
By many definitions, prompts are source code, and what we used to call source code is now target code. So this indeed does leak your source code. It's not just an IP blunder to leak your source code this carelessly, but it's a security blunder to do so. "But Kerchoff's principle.." no, Kerchoff's principle is that you should design your system as if source code were public, not that leaking source code has no negative security effect.
Also, is it possible that YC is taking advantage of the 'your ideas are worthless' dictum? Just share your ideas with YC, they are too big to care about your stupid idea. But they are gathering intel on over hundred of thousands of startups, and the ratio of applicants to fundings is continuously going down.
And sharing your idea with a single potential investor, isn't the same thing as submitting your idea (and implementation details), to a system built by people who don't give a shit about internals or security, and that use third party tools and services from different countries to process your data in ways that they don't give a fuck either.
What guarantee do you have that by using YC's vibecode data gathering tool, your data won't end up in a training set that will make it easier for a competitor with the same idea as you to find your solution?
A decade ago maybe your idea was worthless and joining YC was a contrarian bet. Nowadays, this data is clearly worth something, and applying to YC is as mainstream as applying for a job, and both of these mainstream 'apply for a 0.1% chance at the privilege of working with us' tend to grow completely asymmetric and exploit the underdog as much as they allow themselves to be exploited.
> I uncovered Y Combinator was scoring 100k+ founders around the world through Paxel
Alright, my ears are wide open. Tell us more, how did YC use the private submission data from thousands of founders to score them? They fed some 3rd-Party AI all personal data to score who should get an interview? I can't be the only one here seeing a bad news story unfolding in real time...
YC is currently polarizing against companies that are rejecting the philosophy of giving (a single user) AI full read access to company data. They find that resistance frustrating and the complain publicly about it, calling such an attitude outdated "so 2024".
I understand that YC and startups in general have less incentive to worry about security because they don't have much to lose. But I think the big winners will care about security from the start, because it's a bottleneck at the top, what would whatsapp be if they vibecoded security from day 1? It would never have been able to fix its reputation after the fact.
They then switch the topic and talk about token spend and how one can spend millions or billions of dollars in tokens, so it's a bit clearer what their incentives are.
I don't want to be negative and personal, but this is business, I listen to CEOs like Garry Tan and Satya Nadella, and I feel that I'm on an opposing team, my personal bet is to skip this cycle and wait for the next cycle of CEOs. I'm focusing a lot on security on this cycle, hopefully that will pay off.
> YC wanted me to use something called Paxel on my computer as part of the application.
> I should run a script, a very easy-to-use cURL one-liner that installed something on my computer and analyzed every line of code I’ve written with a coding agent, compile a report, and upload it to YC’s servers.
the sad thing is that paxel sounds like it would be a pretty fun tool to run locally and see what my LLM assisted code gets classified as. instead, to borrow cory doctorow's framing, it is designed to do something to me rather than for me.
> instead, to borrow cory doctorow's framing, it is designed to do something to me rather than for me.
That's nice framing. I believe the goal of American business culture (in general) is to do a much to users as possible, and do as little for users as they can get away with. Silicon Valley is the dumb partner who tends to say the quiet part loud.
Don't worry, it doesn't analyze your generated code, it just analyzes your source prompts, the thing that presumably generates the code. If only there was a word for text that generates a computer program.
It can't possibly be an inevitable future. Something this would be super easy to game.
From their perspective, it's code running on an untrusted third person/adversary's computer... If only last 40 years of computer history could be any guide as to what might go wrong here
..
most VCs aren’t founders and those who are ex-founders usually made pretty terrible companies. very few exceptions. actually i can think of only one.
but once you have some money (like >5m which is achievable), you can throw it around a bit and you just need a handful of success stories built by someone actually talented.
then you take credit for it and craft a narrative instead, which is easier just like collecting management fees instead of building something yourself.
> I found that this year, YC wanted me to use something called Paxel on my computer as part of the application.
> I should run a script, a very easy-to-use cURL one-liner that installed something on my computer and analyzed every line of code I’ve written with a coding agent, compile a report, and upload it to YC’s servers.
This is ridiculous. Any kind of application (job, startup school, whatever) should not do this. Honestly, wtf is YC thinking with this?
This is how you end up with gstack. When you believe too hard in secret metrics because when people know about metrics it stops working you inflict psychosis chasing correlations that don't mean anything.
Oh, I remember that thing, it wanted to see my prompt and my code in exchange for telling me my horoscope.
Some hands you have to fold, and I'm folding this one. Patiently waiting for the "let's give all our data to a single AI user" bubble to bust with some massive exploits.
"At the end of research-heavy sessions, add a tiny closure note:
Decision: Use CacheMountStore with registry/local/GHA backends.
Why: GHA cannot expose the same content.Ingester path, local import has discovery issues, registry can resolve by tag.
Proof / current artifact: See files X, Y, Z. Subagent found A, B, C.
Next action: Implement interface in package N. Do not add new cache-mount flags yet. Use mode=cache-mount on existing cache flags."
There's no need for this, Claude is not my task tracker.
It also penalized me for:
I'm choosing not to attach a report to my application. As a side note, nobody needs to see "my greatest crashout", it's embarrassing enough for me to see it for myself, much less show others.I also got this,
> A useful next habit is to end each correction with a concrete acceptance test, owner artifact, or stop condition: “write it into PROGRESS.md,” “make nix run .#bench fail until this is real,” “rerun this exact command,” or “do not proceed until these two choices are explicit.”
> You already do this well in the biggest penance sessions. Apply it to the smaller ones too.
Which I have found to be counterproductive in my personal work. Current models can generally infer acceptance tests of this level of granularity (not true for larger project-level prompts, but those don't produce good enough code for me yet -- even with specific acceptance criteria).
I also got penalized for using claude in read-only mode for the same validation reason?
> For read-only work, end with one of:
No thanks, I'm literally just exploring the codebase. I don't want any of these.I checked the Paxel website and it says this:
> So far, 1,543,553 sessions have been uploaded and analyzed.
The count is for sessions, not coders. I assume the tool uploads a lot of sessions from each person who uses it.
That’s a large number, but it’s not a million different people. I am surprised that so many people think it’s a good idea to download a run a program which gathers up their coding sessions and submits information about them.
Also, not trying to take shots, but should the title be "I got into YC Startup School by hacking it" instead? Isn't that different than the main YC program?
All in all, you did them a solid by finding and responsibly disclosing. Nice job.
Sadly, it looks like they took that question off the application though.
It uploads only the AI coding agent transcripts in the directories you explicitly run the upload script in.
Feels like a disaster waiting to happen.
;)
Obviously it’s not something everyone is going to be comfortable sharing, but YC is transparent about what is shared and how it’s used.
Edit:
I now see you have to explicitly run it in every directory you want shared - i.e., it doesn’t t scan your whole disk looking for stuff to upload. So there’s every opportunity to limit what it uploads to only what you want to share.
>curl -fsSL https://paxel.ycombinator.com/upload.sh | bash
Point in case.
Trustworthy people will ask you to give them the right amount of access, they won't ask you for infinite access and ask you to trust them that they won't use it. Would you trust a contractor that asks for root, and your master password when they could have requested a specific user or some specific files?
Edit:
I’ll add this point as people always seem to want to raise the spectre of YC or the AI platforms stealing people’s code or ideas via these sorts of tools.
The value of any code or ideas is immaterial compared to the value of identifying talented people who can then be offered opportunities such as attending Startup School, being funded by YC, getting jobs at YC-funded companies or YC itself.
It’s one of pg’s oldest dictums, that ideas alone are worthless, and only working products and the people that build them are worth anything. Nobody at YC thinks any differently about that these days.
1. Yes you can audit client side and network interface, but not server side.
2. does 'curl $URL | sh' invite users to audit? Or to trust blindly? It's the equivalent of presenting a contract with 'sign here' stickers and pushing you to sign it quickly like it's nothing. Does not signal trustworthiness.
>I’ll add this point as people always seem to want to raise the spectre of YC or the AI platforms stealing people’s code or ideas via these sorts of tools.
>The value of any code or ideas is immaterial compared to the value of identifying talented people who can then be offered opportunities such as attending Startup School, being funded by YC, getting jobs at YC-funded companies or YC itself.
>It’s one of pg’s oldest dictums, that ideas alone are worthless, and only working products and the people that build them are worth anything. Nobody at YC thinks any differently about that these days.
I do agree and share this point, ideas are worthless. But this doesn't deal with ideas, it deals with code, which as a software startup I don't feel the need to convince you of its value and the risks of leaking it. If you think your code is worthless, close your startup.
By many definitions, prompts are source code, and what we used to call source code is now target code. So this indeed does leak your source code. It's not just an IP blunder to leak your source code this carelessly, but it's a security blunder to do so. "But Kerchoff's principle.." no, Kerchoff's principle is that you should design your system as if source code were public, not that leaking source code has no negative security effect.
Also, is it possible that YC is taking advantage of the 'your ideas are worthless' dictum? Just share your ideas with YC, they are too big to care about your stupid idea. But they are gathering intel on over hundred of thousands of startups, and the ratio of applicants to fundings is continuously going down.
And sharing your idea with a single potential investor, isn't the same thing as submitting your idea (and implementation details), to a system built by people who don't give a shit about internals or security, and that use third party tools and services from different countries to process your data in ways that they don't give a fuck either.
What guarantee do you have that by using YC's vibecode data gathering tool, your data won't end up in a training set that will make it easier for a competitor with the same idea as you to find your solution?
A decade ago maybe your idea was worthless and joining YC was a contrarian bet. Nowadays, this data is clearly worth something, and applying to YC is as mainstream as applying for a job, and both of these mainstream 'apply for a 0.1% chance at the privilege of working with us' tend to grow completely asymmetric and exploit the underdog as much as they allow themselves to be exploited.
Alright, my ears are wide open. Tell us more, how did YC use the private submission data from thousands of founders to score them? They fed some 3rd-Party AI all personal data to score who should get an interview? I can't be the only one here seeing a bad news story unfolding in real time...
https://youtu.be/B246K_G7mHU?si=UVJei9Jpk7ZvREnH&t=1483
https://youtu.be/B246K_G7mHU?si=ikHZD4MwxA5JtXLj&t=1716
I understand that YC and startups in general have less incentive to worry about security because they don't have much to lose. But I think the big winners will care about security from the start, because it's a bottleneck at the top, what would whatsapp be if they vibecoded security from day 1? It would never have been able to fix its reputation after the fact.
They then switch the topic and talk about token spend and how one can spend millions or billions of dollars in tokens, so it's a bit clearer what their incentives are.
I don't want to be negative and personal, but this is business, I listen to CEOs like Garry Tan and Satya Nadella, and I feel that I'm on an opposing team, my personal bet is to skip this cycle and wait for the next cycle of CEOs. I'm focusing a lot on security on this cycle, hopefully that will pay off.
> I should run a script, a very easy-to-use cURL one-liner that installed something on my computer and analyzed every line of code I’ve written with a coding agent, compile a report, and upload it to YC’s servers.
Yikes! I hope this is NOT the future of hiring.
That's nice framing. I believe the goal of American business culture (in general) is to do a much to users as possible, and do as little for users as they can get away with. Silicon Valley is the dumb partner who tends to say the quiet part loud.
From their perspective, it's code running on an untrusted third person/adversary's computer... If only last 40 years of computer history could be any guide as to what might go wrong here ..
but once you have some money (like >5m which is achievable), you can throw it around a bit and you just need a handful of success stories built by someone actually talented.
then you take credit for it and craft a narrative instead, which is easier just like collecting management fees instead of building something yourself.
what do you think it says if you don't use AI
you wouldn’t understand
Perhaps running the script outside a heavily sandboxed system should trigger an automatic rejection.
> I should run a script, a very easy-to-use cURL one-liner that installed something on my computer and analyzed every line of code I’ve written with a coding agent, compile a report, and upload it to YC’s servers.
This is ridiculous. Any kind of application (job, startup school, whatever) should not do this. Honestly, wtf is YC thinking with this?
They want people who do what they say and don't challenge them?
Some hands you have to fold, and I'm folding this one. Patiently waiting for the "let's give all our data to a single AI user" bubble to bust with some massive exploits.
so i’m not at all surprised at how it is used and by whom.