Debate: do AI-proved theorems count if humans can't understand them?

Stanford professor Anshul Kundaje questioned the value of AI-proved theorems no human can understand, but sytelus countered that mathematics exists independent of human comprehension and proofs compel acceptance—arguing human understanding has hard limits, as with FLT's years-long proof.

2026-09-13 ~ 2026-09-13 · 3 related posts

• sytelus: math isn't 'by humans, for humans' — proofs compel truth regardless of understanding — sytelus · 2026-09-13
• Debate: does it matter if AI proves millions of theorems no human understands? — anshulkundaje · 2026-09-13
• Do humans need to understand AI-generated math proofs? A debate on cognition limits — sytelus · 2026-09-13