Critique of AI Lean proofs: formalization matters, labs leave verification to unpaid community
gerardsans · x · 2026-11-21
- gerardsans argues an AI-generated Lean proof is only meaningful if the original problem's formalization is correct; otherwise abundant proofs are symbol games disconnected from the real question.
- He contends the minimum bar is being able to understand and defend a proof — exactly what AI labs skip, leaving verification to an unpaid, unincentivized community.
- He mocks an Anthropic mathematician who admitted lacking the expertise to follow the math, calling it symptomatic of the vibe-coding era, and claims AI has no intelligence, only data coverage — hence needing a Lean runtime to verify every step it can't explain.
More from AGI Musings
- Counter to LeCun: LLMs May Skip Human-Level and Deliver Superhuman Intelligence — hbouammar · 2026-11-21
- Yoav Goldberg: many harness abilities shift from essential to marginally useful — yoavgo · 2026-11-21
- 'Frictional AI': Adding Resistance to AI to Prevent Human Skill Decay — JeremyNguyenPhD · 2026-11-21
- AI Is Testing What Economists Know About Bubbles — bloomberg · 2026-11-21
- Do executives owe any duty to future shareholders before AI's shoe drops? — AaronBergman18 · 2026-11-21
- Working with AI forces you to live both the maker and manager schedule at once — arjunrajlab · 2026-11-21