It's probably true that the YC interview process is deeply flawed, but OTOH there is no known better way to select worthy investments yet. Good startup business ideas are not only rare but also impossible to identify a priori. The only way to tell whether a business idea is good or not is to actually try it out in the market and see what happens. (This doesn't mean that any crazy idea is good, but it does mean that some good ideas seem crazy before they are tried.)

If there were any methodology that could identify worthy investments with anything close to high accuracy, there would not be stock markets and there certainly wouldn't be a VC industry.

VC investment works differently from other forms of investment due to the outsize gains on the rare successes. This market requires many comparatively small bets with limited losses and a few low-probability but huge payoffs. Compare this to a Warren Buffet style investment methodology of a few large bets in well-established companies and industries with high probability of moderate payoff. Neither is "right" in an absolute sense; it depends on the investor's risk tolerance and the nature of the opportunities.

The VC business model is to apply some coarse-grained filters to screen out obvious scammers and gross incompetents, then give the remainder of the cohort some money, knowing in advance that 99% will fail completely. The remaining 1% will pay for the losses on the 99% many times over.

Moreover, in YC's case, the amount of money they invest is so small compared to the massive returns on their few successes that they can now afford to take many of these small risks on "two guys who seem smart." Most of them will fail, and they know that going in. It doesn't matter. They're betting that sooner or later, they'll wind up with another AirBnB or Reddit that will pay for many thousands of these small bets.