I like this interview because I was in a really good mood when Peter shot it. The startups had just finished the second (of three) demo day sessions, and they had all just nailed their presentations. I was so relieved. You can see from their body language in the background that they're all happy too.
Well, I've been doing it full-time for 5 years now. And at the scale we operate on, 5 years is a lot of data. 172 startups is several times the number a VC partner would deal with in his career.
That's interesting. I just finished a biography a friend of mine wrote about his life as an accidental cheese-seller, and he mentions that he became an expert in just a few years, simply because he tried everything every vendor had to offer. Pretty soon he was at least as expert as any vendor, and often more knowledgeable.
It's funny how much of life is just about putting in the hours. Maybe we'd like to think it's about talent or brainpower, but data wins pretty much every time.
My point is more about being good at something naturally. For example, I think Michael Arrington was always a natural to write Techcrunch, even if he never started it. Tiger Woods seems to be natural at golf (not that he didn't work hard, but many people work hard). I believe Paul Graham is a natural at Y Combinator in the same way.
Often one finds that the top performers in a field have some practice regime that stuns even other professionals.
I don't know how to evaluate the question of whether Paul Graham is somehow naturally better at this than others. There's clearly something about him that's different, or he wouldn't be doing what he's doing. But to guess at what those qualities are -- that endeavour is so vulnerable to confirmation bias it's practically void of meaning.
Or you could just shrug and say that by now he's seen hundreds of proto-startups. For whatever reason.
I mean VCs should invest in a higher number of startups, even if it's done by allowing less money per startup. At the very least they should have a simpler application process, with a clear decision time frame. These improvements would be hugely beneficial to startups, and would take just a bit of effort on the part of VCs/investors.
We have protections against arbitrary dilution, by e.g. issuing extra stock or doing a new round that values the company at $1. We've never needed them. What I meant by the "no soup" remark was that we don't care that much about legal protections; it is a more powerful deterrent that we would blacklist any VC who screwed us over-- or who screwed over any of the startups.
Creating this "repeat player" effect is awesome. Sometimes the best-drafted legal provisions can have loopholes. This kind of leverage makes everything much simpler.
Empirically, very much so. Stories of investors maltreating founders are not uncommon in the startup world generally, but they are extremely rare among the YC alumni.
That was nice. Although, the interviewer just talked too much. A good journalist knows how to get out of the way.
I have a question for pg, if he is reading this. You said that about 70% of the companies exit YC with funding or do not need further funding because they are already bringing in money. Do you usually return your investment at this point, or do you wait until a later VC round, or do you generally wait to the point where the founders cash out to return your investment?
Like most investors, we wouldn't normally get any return till what in the business is called a "liquidity event," meaning either an acquisition or an IPO.
Technology startups have never paid dividends. There are probably multiple reasons why not. There might be some way to tweak dividends to make them work for startups, but I know so little about that sort of thing that I wouldn't want to speculate.
On revision, thanks for deciphering my terse question.
As an aside, I did invest in one web startup which took the unusual step of returning the entire first round investment by way of dividends prior to an IPO. Very peculiar and not US based, so I figure your assertion holds.
