The eight filters that decide which problems mathematicians actually study

burny_tech · x · 2026-09-27

A widely shared observation lists the traits of problems mathematicians disproportionately study: compressible enough to describe, understandable within human working memory, expressible symbolically, connected to existing math, aesthetically attractive, publishable, teachable, and provable in reasonable-sized arguments.

The implication: mathematical research is shaped by human cognition and academic incentives rather than pure objectivity — often cited in debates about which problems are best suited for AI.

Original post →

More from AGI Musings

AGI Musings channel →