Y Combinator

Y Combinator's published Startup School material is unusual for being grounded in its own data rather than in received wisdom. Its talk on startup ideas draws on an analysis of the top 100 YC companies by valuation and how each of them actually arrived at its idea, plus the patterns visible across thousands of rejected applications.

That second source is the more valuable one and is rarely available anywhere: the specific mistakes that lead capable founders to bad ideas, described by the people who read the applications.

The framing is candid about its own limits — nobody, YC included, knows which ideas will succeed, and execution matters at least as much as the starting idea. The stated goal is only to stack the odds by starting from a more promising one, which is a more defensible claim than the certainty this genre usually offers.