I don't think anyone can realistically expect to find a google-scale diamond in the rough at every single YC batch. "Frighteningly Ambitious Startup Ideas" I suppose is about a process of collective unconscious iteration that would yield such a diamond every 5 or 10 years.