Csaba Szepesvari on AI math proofs: Lean kernel and compiler bugs undermine trust
CsabaSzepesvari · x · 2026-09-07
RL researcher Csaba Szepesvári discussed the reliability of AI-formalized math proofs.
Key points:
- Formalized proofs rest on a chain of trust: a correct compiler, a bug-free Lean kernel, and trustworthy Mathlib plus its extensions — "relatively trusted" isn't bug-free.
- Formalization introduces many new definitions and proofs; one mistaken axiom means you've proven something else entirely.
- On reflection, he added that his worry may be overblown.
Related event: Claude's FLT Proof Ignites Math Community Scooping-Ethics Debate(26 posts)→
More from AGI Musings
- US senator slams AI industry for proposing 50-death threshold for 'catastrophic incident' — peterwildeford · 2026-10-27
- Replit CEO Says AI Makes All Software De Facto Open Source; Aakash Gupta Asks What Moats Remain — aakashgupta · 2026-10-27
- Understanding the AI consciousness debate: which theory of mind is behind it — msharmas · 2026-10-27
- Functionalism: if consciousness is just what the brain computes, silicon can run it too — msharmas · 2026-10-27
- iPhone case with 85,000 reviews in days: everything fake in the age of abundance — tekbog · 2026-10-27
- Whether AI can be conscious depends on your definition: five frameworks for thinking about it — msharmas · 2026-10-27