Financially, as if you were an investor. They're the people whose job it is to evaluate startups' prospects, and they care above all about two things: the founders and the market. The founders should be relentlessly effective, and the market should ideally be of a size that can only be obtained by riding on trends beyond the startup's control (but visible to few besides the founders, or the market would already be full). Joining the young Microsoft, for example was a bet on Bill Gates and microcomputers, both of which turned out to be very good bets.
As a hacker you may be able to judge market bets as well as or better than many investors. E.g. I think HN readers knew Dropbox was onto something before most investors did. So if you go wrong it will be in judging founders. For many hackers, especially the unwordly sort, it's hard to distinguish true Bill Gateses from mere good talkers.
I wish I could offer some advice about distinguishing, but that would take a whole essay. The best simple hack I can think of is completely self-serving, but I'll offer it anyway: piggyback on our filter. YC specializes in distinguishing between genuine Gateses and good talkers. We're occasionally fooled, but far less often than a typical hacker looking for a job would be.