Critic: nobody can verify LLM-generated Lean formalizations, so Slop Lean isn't a solution
_onionesque · x · 2026-10-11
In a debate about LLM-generated Lean proofs, the author argues "Slop Lean" isn't a solution: much of the required mathematics isn't formalized yet, and the teams involved have neither the intent nor the expertise to check whether LLM-generated formalizations are actually correct.
The author mocks the idea of asking people to clean up 200-page LLM-generated dumps, quipping that the "TCS geniuses" should present the work properly themselves. Core point: formal verification is only valuable if correct, and unverified LLM formalizations may be wasted effort.
Related event: Critics warn against LLM-generated Lean proofs no one can verify(3 posts)→
More from Models
- Musk shows Grok researching and ordering Lego Star Wars kits in one prompt — elonmusk · 2026-10-11
- Google ships EmbeddingGemma 2: 740M multimodal embeddings that run on phones — dl_weekly · 2026-10-11
- Rumor: Grok 4.8 with 2.5T parameters (up 67% from Grok 4.6) may launch this week — mark_k · 2026-10-11
- TensorFold 1.0.7 writes each learned fact into ~10 new neurons, 4x faster with 3D view — HankYeomans · 2026-10-11
- Arena scores look close: OpenAI 88 vs Claude 83 means double the error rate — i_dg23 · 2026-10-11
- Rumor: Anthropic's next-gen Fable 5.5 can one-shot SVG generation — koltregaskes · 2026-10-11