Overheard at YC's 2012 Demo Day: "I really liked 2008. It was a great year. People weren't trying to make money. They were trying to do something." PG is right that @ycombinator has a surprising legacy of critics, going back as far as 2007. The best batch was always the earliest batch. There's a simple reason for this: observability. The latest batch is inherently a list of nobodies, pursuing odd looking ideas. First, we get to know the loudest and most controversial companies. Only much later are companies recognised for actual progress and meaningful successes. So, the lazy read is that batch quality is constantly falling and it's only ever the older batches which were actually any good. In defiance of this, the program has posted consistent, impressive results. And it has done so by supporting companies across the spectrum, like @boomsupersonic, @oklo, @AstroForge, @opentrons, @initofertility and @noxmetals. Most importantly, YC produces a credential which is more earned than awarded. The YC stamp allows other investors to herd behind promising companies, rather than chasing Stanford grads, Meta alums, or whatever other "legible" signals the market dictates. It's easy to imagine that without YC, venture capital would be even more crushingly consensus-driven.