Predicting a Lean community schism: human-readable proofs vs AI-generated slop

jacobaustin132 · x · 2026-09-07

In the wake of Claude's Fermat's Last Theorem formalization, @jacobaustin132 argues the most scalable path is a big graph of theorem statements anyone can burn Claude/Codex tokens on, since AI cares less about code quality. With no rush to merge such proofs into Mathlib, he predicts a coming schism between the "human-readable Lean" camp and the "let AI slop dominate as long as it proves everything" camp.

Original post →

More from AGI Musings

AGI Musings channel →