It's fine to work on any problem, so long as it generates interesting mathematics along the way - even if you don't solve it at the end of the day.