Math PhD career track

A problem with doing a PhD after working for a few years is that in pure math at least, people’s ability to make serious new contributions to the field seems to decline rapidly with age. Even 25 is getting on a little, and by 35 pure mathematicians are over the hill.