because those problems involve messy parts of the human experience. math/physics, basic science can be pretty divorced from human sociology/psychology/etc.

you don't need to think of the 2nd/3rd/4th order effects on society if you prove some theorem in number theory.