Formalizing CFSG at 3x RH scale: Lean is now the bottleneck for AI math

Vjeux · x · 2026-11-23

Vasily Ilin (Axiom Math) responds to Antoine Chambert-Loir's messy formalization effort, arguing the cleanup itself is valuable — everyone involved now has firsthand experience improving Lean code at scale.

Key points:

The thread is a rare ground-level report on where formal proof engineering hits its limits when applied to mathematics at maximum scale.

Original post →

More from coding & agent

coding & agent channel →