Learning that my guess as to what might work or not is just as accurate as anybody else's, that is, not very accurate.

I used to think that somehow kicking around ideas on places like HN made the ideas better, when in fact all they did was create ideas that hackers thought were cool. Different thing entirely.

You only know an idea is good or bad once you start executing. The market is always different from what you imagine (or want it to be)