This was absolutely Paul's strategy -- read his essay reflecting on returns from their early portfolio (AirBnB and Dropbox). His take -- "we should have done another 1000 investments to get a third hit" (coinbase, it turns out. And he was correct).

The math for YC's seed stage returns is overall inexorably towards more checks being written - the checks are incredibly cheap. YC's IRR is currently roughly 176%, yielding an 81,000x return since 2005 -- another way to say it: a single $25k investment from 2005 could be deployed into 81,000 $25k investments in 2026. Or another way to say it, for every check they wrote, they're able to write another check off that investment every 8 months, compounding all the way back.

YC's main problem is that there aren't 81,000 even slightly viable startups in the world. It's true they've lost their "exclusive" vibe, but to be honest, when they started it was more "small and quirky" which briefly became "gold standard exclusive." It was never their original brand, and I don't think it was ever really part of the strategy.