I do that, I try to find loopholes and edge cases to better define my mental model and then test them against the knowledge of people who know better than me.

I figure GGP is trying to do the same, and not sending in a paper to Nature right now.