Something that I've really not seen mentioned, is that they are also (and still!) ignoring the licensing that Microsoft has around VSCode.
VSCode is built on open source of course, but the the OSS version is "Code - OSS". When Microsoft builds it and releases it, it becomes VSCode with the Visual Studio trademark and what not.
PearAI's Code fork is using the real VSCode marketplace which has strict "This has to be used with Visual Studio products" (well, and some other MS / GitHub bits), so they can't use that. If you look at other VSCode-adjacent editors, they all use open-vsix.org instead.
They also use extensions that are licensed in the same way:
https://trypear.ai/blog/wsl-setup
They have instructions on setting up the WSL extension...which has a "This can only be used in Visual Studio Code" License too, so it can't be used in their PearAI fork.
You can see some examples of the terms of use / LICENSEs here:
https://cdn.vsassets.io/v/M190_20210811.1/_content/Microsoft...
https://marketplace.visualstudio.com/items/ms-vscode-remote....
It just shows a complete lack of regard for licensing...
The "dawg I chatgpt'd the license" tweet was absolutely egregious.
Honestly it also reflects equally badly on Coinbase if they were paying these guys $300k.
One positive outcome is that this might have cured a lot of people's imposter syndrome.
I imagine coinbase has a legal department and these guys were paid to code.
Using chatgpt to write a license is a ridiculously dumb thing to do since there are so many public templates to choose from, it sounds like they're just overly enthusiastic about AI.
I think there's a line where AI enthusiasm becomes disdain for human expertise. If you can't be bothered to learn about licenses, what else are you just going to assume chatgpt can figure out for you?
I've interviewed for Coinbase and it was a shit show. The recruiter initially said my interviews went well but then never got back to me, even after me reaching out multiple times. Plus there are lots of reports from people working there that things are held together by strings and glue, and the work life balance is terrible.
even working face to face with relatively incompetent people has never helped with my imposter syndrome.
If anything this is a stark reminder of the hiring bubble we were in the past few years..
The bad reflection goes both ways: Working for Coinbase already reflects badly on anyone no matter what they're paid.
Credit to Garry for unequivocally apologizing and shouting out https://amplified.dev . The whole PearAI situation was a bad look, and it made it seem like YC had not done their appropriate research before investing.
I doubt this really did permanent damage to their reputation (I don't know if anything can at this point), but this move was definitely necessary to placate anyone who was paying attention.
> it made it seem like YC had not done their appropriate research before investing
It still seems like that. Moreso, even.
Developer tools startups generally don’t make money. Seems like a pretty risky bet all round.
But this post doesn’t provide any details about Pear.AI outside of them scrapping their stolen code, which is all they had?
This says too little of substance. What errors were made in YCs assessment, which will be corrected? What is the path forward?
People have weird ideas about what YC is and does. People apply to YC: you can go look at the application. They get gajillions of them. They downselect applications to interviews. The interviews are brief. They downselect interviews to offers to join a cohort. Every company in the cohort gets a group partner that they do regular "office hours" meetups with. There are internal events and talks and stuff. There's a standard investment, and demo day. That's it.
YC doesn't operate YC companies. They're investors. They aren't your boss. They're not looking at your code. They're almost certainly not doing any meaningful technical due diligence. These are things that happen, to an extent, in priced rounds that happpen after your seed round (A-round technical due diligence is also a joke, but whatever, stipulate). The whole point of YC is that they're investing in big batches of startups, at a scale that is a big multiple of what an A-round VC does in a whole year.
Tan is apparently apologizing because he spoke up on behalf of a batch company, or repeated some misinformation TechCrunch published. But it is very weird to ask them how their selection process would change to account for this happening.
> YC doesn't operate YC companies. They're investors.
YC also offers mentoring and coaching services, which goes beyond what a traditional investment firm does. That's part of their appeal, anyway.
> They're not looking at your code. They're almost certainly not doing any meaningful technical due diligence.
Maybe they should?
If their job is to find talented, and hopefully also ethical and honest, people, then doing a basic check like ensuring that the project they're being pitched was actually authored by the candidates should be part of the vetting process.
If, OTOH, their job is to bet on anyone who's likely to give them the highest ROI, regardless of their personal values, then they don't need to do any of this, and cases like PearAI should be acceptable. But that's not what YC claims to be about.
> The whole point of YC is that they're investing in big batches of startups, at a scale that is a big multiple of what an A-round VC does in a whole year.
Maybe that's part of the problem then. When you're trying to fund and mentor hundreds of companies in a few months, it's impractical to think you'll be able to successfully screen all of them. But the startups must grow...
Having seen/participated in it firsthand several times, I don't believe priced, institutional A-round technical due diligence would have turned any of this up. It is not realistic to think YC would have.
The problem is not with tech due diligence.
YC can never say enough that they invest in people, that only the people matter, etc. "Founders, Founders, Founders" is YC's answer to Balmer's "Developers, Developers, Developers".
That they decided to fund bros that simply copy-and-paste others' code (!) and are happy to declare publicly that they "chatgpt'd the licence" (!!) and "can't be bothered with legal" (!!!) is... alarming?
> I don't believe priced, institutional A-round technical due diligence would have turned any of this up
I've only done it once but doing a cursory check of the commits and who owns the copyrights etc is pretty fundamental during due diligence.
I'm skeptical. I don't think this "vetting" argument is a real thing. And, again: you're talking about a priced A round, where the legal costs alone are into the six figures. For YC? Come on.
You're making it sound like this is complicated.
It takes minutes to go through a public Git commit history.
It takes minutes to find this specific issue in hindsight. It is laughable to claim that you can do a meaningful audit of a codebase in minutes.
Good technical due diligence from reputed organization costs something like $50-100k. Good legal/accounting due diligence costs the same amount. Which is >20% of what YC invests. Even if 20% are committing clear fraud, it is worth it to pay them. Actual numbers are likely lower.
I can't help thinking that there has to be something inbetween doing nothing and going all the way.
Like sitting down with them for five minutes and talking, to see if they say shit like "dawg I chatgpt'd the license".
The "that's it" is the problem here. The lack of due diligence on YC's part caused a PR disaster and ideally that's something they would want to avoid.
YC's way of operating is globally optimal. The cost (in money and time) of adequately vetting every company they've ever invested in would dwarf the cost of this little brouhaha.
Even in later rounds due diligence skews non-technical. E.g. for Series A it mostly takes the form of "What licenses do your OSS dependencies use?", and (for B2B SaaS) "We want to talk to a half-dozen of your customers, preferably of different sizes and from different industries, and ideally one that's renewed".
PR disaster is a bit of an exaggeration, outside of a fraction of HN no one cares or has noticed.
True, dream world has been already forgotten.
“Don’t blame them, they just haphazardly throw money at projects without vetting them properly”
Well uuhhh, maybe they shouldn’t? Would a bare minimum of due diligence not be ok lol?
This is a good and thoughtful response. As much as we’d like to get mad at it, there’s not much more to do.
I had the same reaction. I'm usually quite cynical about these things and was ready to scoff at it, but I thought that was about the most you could really accomplish in an apology.
The key to me was that it wasn't focused on how sorry YC is, it focused on how great of a team Continue is. Far too many apology posts are just another way to get attention, so it was nice to see that attention directed at the party receiving the apology.
Ideally it would have a section on next steps and how YC will prevent such events in the future, but definitely preferable over a corporate nonapology.
We are absolutely taking steps internally to prevent this.
This is great. But as I said in another comment, you guys don’t owe us anything, we’re not YC shareholders.
But YC is so prevalent in the hacker community that is kinda associated with Open Source, is like we feel entitled to ask for explanations when in reality it’s just a private business. It’s an interesting phenomenon.
I think it’d be great if you could keep transparency in order to keep encouraging founders to apply to YC, but we have to learn that’s just an option.
But we’re not YC shareholders, they don’t owe us anything. I understand the point, it’d be nice to read how they might prevent it, but it’d be completely informative.
Of course they don’t owe us anything. But when you back something that ends up being a big problem people generally won’t trust your judgement until you explain the how and the why. You are perfectly at liberty to not do that, but then you get to deal with the consequences of that.
You do know that Garry banned a bunch of people who criticised him over this.
And that when told about PearAI violating the terms of the license said to the effect "who cares it's open source" as though none of this actually matters and it's just a bunch of nerds over-reacting.
It's arguably worse than what PearAI did in the first place.
Subtle touch here in linking repeatedly to Continue and not Pear. Really nice implicit message on what Garry's final view of the situation was after getting more information.
The article reads as diplomatically throwing PearAI under the bus, which may be a conflict of interest.
Whatever everyone thinks, YC is playing to win. And they make winners
>>if you don’t make mistakes it means you are not trying hard enough. << Elon Musk
Big picture, people who never heard of PearAI or continue dev just did. I did. I tried it. It’s as good as cursor
And guess what? whatever side you are on, YC and those startups won either way. More developers heard about each one of them and more tried it.
Counterpoint: PearAI was launched as a trivial copy of another product and framed by its creators as the greatest thing since sliced bread.
It's obvious that these people are less interested in making something useful than they are in making something that will generate enough hype to get another round of funding.
Well I guess only time will tell.
Recent and related:
The AI startup drama that's damaging Y Combinator's reputation - https://news.ycombinator.com/item?id=41722507 - Oct 2024 (48 comments)
"Our view is that PearAI should start over from scratch" - https://news.ycombinator.com/item?id=41712589 - Oct 2024 (4 comments)
YC criticized for backing AI startup that simply cloned another AI startup - https://news.ycombinator.com/item?id=41707495 - Oct 2024 (257 comments)
Pear AI founder: We made two big mistakes - https://news.ycombinator.com/item?id=41701265 - Sept 2024 (228 comments)
Y Combinator Traded Prestige for Growth - https://news.ycombinator.com/item?id=41697032 - Sept 2024 (244 comments)
Did I miss one?
Did not get much discussion on HN, but was substantial on X:
Our view is that PearAI should start over from scratch - https://news.ycombinator.com/item?id=41712589 - Oct 2024 (3 comments)
Ok, I've inserted it above. Thanks!
The thing about doing something like what PearAI did is that worst case, they get flamed, they apologize and make changes, but overall don't lose much. They still have the YC label, the money, and free publicity. In a few months nobody will care, so overall their decisions were a net benefit.
Has any startup founder ever really suffered due to bad publicity? Short of extreme cases like Theranos, worst case scenario is they get a golden parachute, enjoy a few months/years "soul searching" then emerge with another startup. However most of the time nothing really happens.
People only came after PearAI because their founders' arrogance was particularly egregious. A bit of preemptive covering their tracks and nobody would have done anything.
I agree with you, but what do you think should be done instead? What they did was misleading and dishonest, but it wasn't illegal. Should they be expelled from YC because of it? Maybe, maybe not.
I feel like endless persecution for "bad" behavior, part of the current cancel culture we're living in, is not healthy for society. People can make mistakes, and as long as they acknowledge and make an effort to correct them, they should be allowed to move on with their lives.
Well that's the thing, they only owned up to their mistakes because they were criticized. However what they did to any objective observer was a net benefit for them in the end, so it overall incentivized misleading and dishonest behavior as long as you can get away with it.
The only things standing between where we are and a low-trust kleptocracy are upholding standards of behavior, either through encouragement of good behavior or punishment of bad behavior.
Well, they could be kicked out of the program for one.
> What they did was misleading and dishonest, but it wasn't illegal.
Misrepresentation is illegal [0].
[0] https://www.investopedia.com/terms/m/misrepresentation.asp
People who say shit like "dawg I chatgpt'd the license" aren't healthy for society.
> PearAI has apologized, removed the offending repositories, and is working to correct all their mistakes.
If the issue at heart is a forked repo...then what's left?
Nothing. It sounds like all the problems have been fixed.
And now all we are left with is two random guys who have built nothing, have no novel idea and have given the middle finger to the open source community.
all their mistakes... that they are working to correct.
Why shouldn't YCombinator just drop PearAI from their program? I mean they literally stole another YC startup's work wholesale and effectively defrauded YC investors. I don't know much about startup land but it seems there's an absurd amount of forgiveness given to a company that is clearly fraudulent.
Who cares whether or not they received the standard amount of due diligence/vetting? The only reason a system like that can work is if participants act in good faith, and this company obviously didn't.
edit: Like I said, I'm not exactly knowledgeable about the world of tech startups. If i've learned anything though, it's that the money being given to obvious scam artists being excused as "vetting batch investments doesn't scale" has made me even more depressed and cynical about tech startups than I previously thought was possible.
> I mean they literally stole another YC startup's work wholesale and effectively defrauded YC investors.
Didn't YC know it was a fork of one of their existing investments? If they knew then it is bad, if they didnt know then it is bad. If they didnt know and continue to fund this startup then they are stupid.
> The only reason a system like that can work is if participants act in good faith
If you are a professional investor you need a better bullshit detector. If you fail any time someone acts in bad faith, it is not the game for you.
The investment world is full of functional sociopaths. You get used to it I guess. YC's bullshit detector would be blaring nonstop if that was enough to tank a deal.
YC funds startups based on their teams and ideas, and in some cases, even without the latter. Building a code editor is not a requirement to get accepted.
It's very possible that Pear AI even mentioned that they're a fork in their application, just like they mentioned it on their description on GitHub; you could say that they were being dumb by generating a license with ChatGPT, but at worst that's a stupid mistake, not fraud.
Does it really matter what it is? A team that is going to use ChatGPT to generate their license is likely to do a bunch of other ill-advised stuff.
I can promise you that after all of this outrage, they're going to be much more careful about everything they do.
To respond to your question directly though, yes it does matter; any accelerator that accepts hundreds of companies every year based on a 10min interview is gonna get some bad ones, but rescinding the offer to them is just really bad form if they didn't commit actual fraud.
I don't think it's obvious that they will be careful. IMO it's rather another reinforcement for "ask for forgiveness, not for permission".
Not to mention keeping them saves a lot of face.
Isn't this a thing that blew up like a week ago? After they were accepted into the program? If that's the case, what does this have to do with "vetting"?
I'm not sure I understand your confusion. Yes, any vetting that occurred would have occurred in the past. Clearly there wasn't as much vetting as some other comments i've seen might have hoped. Nevertheless, a system which requires minimal vetting (as suggested by others) only works if participants aren't literally fraudulent.
I'm just repeating myself though. That's all in my original comment.
Why should they drop them, though? The damage is done, the money is spent, and they can just as well wait it out now, to see if anything comes from it anyway. In the worst case, it is a failed investment, and they will have plenty of those, and the reason doesn't actually matter, except in learning from it for the future.
There definitely are things to get depressed about, I just don't think this is one of them.
LOL the desire to cancel is so strong. Dude, it’s YC’s money and they’re betting where they’re betting. They’re the ones being “defrauded” if any.
> They’re the ones being “defrauded” if any.
Yeah that's what I said.
Was doing a bit of searching to get context on this story and found an unrelated but interesting tangent... there's another unrelated startup named pear.ai that's been around for years.
It's in a different domain, but the name is basically identical. Are people just... not doing any due diligence around company names any more? At all?
> Are people just... not doing any due diligence around company names any more? At all?
Well, there was a famous tweet/post/("X"?), answering to the claims of illegal relicensing, reading, "dawg i chatgpt'd the license (…) we busy building rn can't be bothered with legal" – I guess, the answer may be indeed "no"…
X (formerly known as Twitter) Post.
i’ve been using “Xeet”
A "message at no-not-the-famously-complex-RPG-with-the-aliens"is always worth considering and probably on topic. ;-)
If no two companies could share a name, we'd have run out of names by now. (Case in point: I'm building a company called Stack.)
This is only true if you are very uncreative about names.
Companies aren’t even doing due diligence on their own internal naming schemes. Hello Google Chat number twelve.
It’s a trend. There was a wave of companies with unique and easily pronounceable names, then it was popular to drop a letter and add a “ly” suffix, later a phase of my/get/lets prefix. Etc.
It’s shaped by how we build and use digital products. E.g. domain names were once super relevant, today less so.
Perhaps in this case they don’t care; they’ll either survive and become the dominant pear, or they’ll get eaten.
They are not. It's currently hip to use a single word from the dictionary as your company name, no matter how many others are using it.
This drama kinda happened on a platform called "X" so...
I see similar name overlap criticism by people in every other HN thread. Are you of the opinion that once somebody names their company something, that name is forever unusable for a similar or different company ever again? I’ve seen both kinds of examples being criticized, even for acronyms! Mind boggling.
Pear is an anagram for Rape!
I didn't even have to use the Internet Anagram Server to perform that minimal amount of due diligence.
https://new.wordsmith.org/anagram/anagram.cgi?anagram=pear&t...
So PearAI is actually forking VSCode? I wonder if they'd have been better off building on the latest iteration of Eclipse Theia.
PearAI is forking Continue's VSCode extension, from what I can tell.
> Afraid of switching editors? No need, PearAI is a fork of VSCode and Continue so you'll feel right at home
Literally from the smack center of their homepage.
Apparently YC has over a dozen active partners. They don't coordinate tightly. They compete against each other for investments and often overlap verticals.
That's why you see 2, 4, or more redundant investments in the same space from YC. It's a mess but not uncommon.
Did you go through their commits?
Hint: https://x.com/minitru/status/1840549001797701984/photo/1
Here’s the commit that tweet dunks on:
https://github.com/microsoft/vscode/commit/78c1777cd12559745...
It came from the upstream VSCode repo, and merely removed the words “all rights reserved” which do not appear in the “official” MIT license.
This is a pretty poor apology.
Open source has been infinitely more valuable and important to humanity than the latest LLM or AI trend will ever be. Without it Linux or Apple would simply not exist. And every startup would be solely dependent on Microsoft for their future. It would've set back software decades and seriously harmed the take up of the internet.
And all of it hinges on everyone agreeing to play by the same rules which are codified in the licenses. And acting like it's no big deal that can be safely ignored is to frankly spit in the face of all of the selfless developers who contributed their code for no financial benefit over the many decades.
> Out of instinct to defend YC companies
Credit to Garrry for realizing he is very biased and offering a tepid "apology."
He's an idiot
ICYMI
> In 2024, Tan attracted criticism for a profane tweet wishing death on seven of the 11 members of the San Francisco Board of Supervisors. In a later apology, he said he was referencing a rap track.
https://www.sfchronicle.com/sf/article/garry-tan-tech-exec-a...