One very important thing that even the smartest people seem to have to learn from experience is taste in problems. There are juicy problems that tend to lead to interesting discoveries, and nasty problems that tend not to.