If AI can produce correct proofs cheaply, what still matters? A mathematician's take

RexDouglass · x · 2026-09-12

Pengchuan Zhang, a mathematician working in AI, untangles what remains valuable when AI can cheaply produce correct proofs. He distinguishes three layers: a proof can solve a problem, learning to prove can educate us, and inventing a proof is creative work. Math's value takes scientific, practical, educational, and creative forms — Fields medalists emphasized the educational and creative values tied to their own work. Every mathematician, he argues, should know which value they prize and what their math can deliver.

Original post →

More from AGI Musings

AGI Musings channel →