A bunch of truly naive people trying "impossible" things sounds a bit like the Infinite Monkey Theorem.

However, what if they are grad students, or the best of the best product people/devs, which is what YC tries to harvest?

That improves the chances of the naive finding some actual disruption, does it not?