Terence Tao argues Lean-verified proofs humans can't understand shouldn't be published

arash_zaghi · x · 2026-09-02

Responding to a discussion on LLM-generated mathematical proofs, arashzaghi raises the philosophical question of whether human understanding matters or we're just enriching the knowledge graph. He cites Terence Tao's recent position that proofs verified by Lean but incomprehensible to humans should not be published — and admits he's torn.

Related event: Terence Tao Sparks Debate Over Unreadable Machine Proofs(3 posts)→

Original post →

More from AGI Musings

AGI Musings channel →