Also, startups are a big risk financially. If you try something that blows up and leaves you broke at 26, big deal; a lot of 26 year olds are broke. By 38 you can't take so many risks-- especially if you have kids."
We learn how to control our variance as we get older. (I'm 29, so I have no idea if I'm "young" or "old" by SV standards.) High-variance life isn't fun unless you hit an early home run. Otherwise, it's stressful and shitty. Experiences like getting fired because you were better than your CTO's-friend manager and made him insecure are pretty damaging, and they lead to self-filtering and variance-reduction as a survival strategy.
Give us an R&D environment that doesn't beat the shit out of high-variance people for 20 years, and you'll see a world where experience doesn't reduce variance so sharply, but delivers increases on the whole distribution.
So many current work environments have self defeating dynamics. You're not allowed to be better than your boss. When that gets thrown away (e.g. startups -- you have no boss) everybody can thrive. Well, they can thrive if they haven't been trained over a lifetime to be risk averse due to shitty management.
But given the "startup = growth" essay, where he says that only 1 company per cycle makes any difference to YC's returns, I can see the logic behind ageism.
Young people (I'm not one of them anymore) just have more unpredictable outcomes. They are the Black Swans.
The Black Swan theory is that the outliers shape our world. And that has been true in tech for sure -- Apple, Google, Microsoft, etc. are all anomalies. These companies grew enormously quickly and were all started by very young and inexperienced men. The two YC success stories that get pointed to -- AirBNB and DropBox -- were also started by the very young and inexperienced.
Older people are perhaps more likely to be successful. But there is less variance in their outcomes. There's the logic in investing -- once you're making money, you're less valuable. Because people know what you're worth. Before you make any money, people can ascribe crazy valuations to you. It will be wrong a lot of the time, but it doesn't matter because the one Black Swan event is what you're looking for. You only have to be right once.