I think what stops them is motivation and a realization that the information can be really valuable when reasoning about software.

Certainly I am making up for a deficiency of math during my early years now. I always dismissed all category theory as useless but increasingly I realize how important statistics, category theory, and a solid understanding of how to analyze algorithms is. Even if you never prove the time complexity of an algorithm, being able to approach new literature and come out with new insights for your engineering efforts is invaluable.

Which is what any civil or mechanical engineer could have told you about their career, I guess.