I'm moonlighting, but I'm going after customers who move slowly. That is, if full-time founders decided to compete head-to-head, they might pump out features faster, but would still be stuck in the same long feedback cycle I'm in as far as knowing what features are and aren't worthwhile. For that kind of customer base you can afford to moonlight; in fact you should moonlight because you would burn through funding before getting traction.

If you're going after faster-moving customers, e.g. Techcrunch readers, you really need to go full-time. Otherwise once you get traction full-timers might imitate you, and then woo those fast customers with fast development.