It's funny how much of life is just about putting in the hours. Maybe we'd like to think it's about talent or brainpower, but data wins pretty much every time.
I don't think that's quite true.
Often one finds that the top performers in a field have some practice regime that stuns even other professionals.
I don't know how to evaluate the question of whether Paul Graham is somehow naturally better at this than others. There's clearly something about him that's different, or he wouldn't be doing what he's doing. But to guess at what those qualities are -- that endeavour is so vulnerable to confirmation bias it's practically void of meaning.
Or you could just shrug and say that by now he's seen hundreds of proto-startups. For whatever reason.
I wish VCs would take note.