Comment by vessenes

The blog post is dumb, in the extreme. Sven doesn’t know the PearAI founders, the key metric YC says they assess when they make a funding decision. Where they forked a codebase from is a deminimus consideration — as Gary says, if it’s in the license, it’s in the license. Maybe Sven thinks that’s unethical, although in this case it’s funny to complain on behalf of YC, “Hey, YC, your money went to an open source codebase, and then some more of your money got to use it! You guys suck!”

Comments here miss a lot as well, (although I agree that YC’s prestige days are over) — PG’s plan was ALWAYS to be able to do more. He lays out the reasoning in an essay maybe eight years ago — if your portfolio looks like 1-2 companies (now 3) out of 500 made 80% of the returns, what should you do?

There are basically three answers to this: Rejoice, Write 497 less checks next time, or write 1,000 more checks in the hope of getting a fourth.

PG’s head was at: write 1,000 more checks. I like this attitude a lot, and believe there are good social, macroeconomic and financial reasons to act this way. Of course you will put up with more failures in that mode — you already went and got the “good” ones — you are now picking ones you didn’t love that much in the sure knowledge that you’ll happily be wrong about one (or maybe even two) of them.

Especially when you’re considering what to do mid-ZIRP, this is I think the only rational strategy if you want to make more money and help the world see more cool things.

That said, one thing YC companies benefitted from immensely in the height of the prestige era was just that, prestige; it was a virtuous cycle in that getting in to YC guaranteed a quality seed and probably A round just on the name. In that way, it was like getting in to Stanford or MIT. Today, along with mentoring, one of the main benefits is the larger network internally, and this is a different thing. Possibly better, possibly not. I do think that if the prestige is needed for their model, they’re probably over scaled for today’s markets and venture money.

Of bigger concern, I would say should be the question: “what does pre-seed/seed even look like in five years?” — Much of what seed capital is for can be done with quality LLMs today — and I expect that trend to continue. We saw the rise of pre-seed firms in exactly this economic environment — through the middle of the dotcom one boom, a tech startup was EXPENSIVE — ten engineers, $30k sparcstations, long data center contracts — the cash was needed. YC started at an inflection point when open source tooling and availability was driving that cost down.

The next YC / early stage fund is going to look very, very different than the last one. And that’s okay! It will be fun to see what’s next.