Claude Fable produced a counterexample to the Jacobian Conjecture
TL;DR Highlight
A tweet claims Claude Fable found a counterexample to the Jacobian Conjecture
Who Should Read
Developers curious whether LLMs produce genuinely useful results in math and search problems
Core Mechanics
- The article body is missing
- Only the title and comments were provided.
- The title says Claude Fable produced a counterexample to the Jacobian Conjecture.
- Only the title is verifiable
- How the counterexample was found is not in the source.
- No prompt or reasoning trace is in the source.
Community Reactions
- Surprise at how small the counterexample is
- One commenter recalled people brute-forcing polynomials with about 16 variables and 80-700 terms, guessing a degree lower bound near 200.
- They found a degree-7 result astonishing by comparison.
- Opinion that it could have been found in 1997
- A comment said a grad student could have found it with about a 3-day computer search.
- Another view was that few people seriously searched because intelligence did not scale.
- Verification and process sharing are the bottleneck
- One commenter doing math vibe coding for months said verification takes longer than making discoveries.
- Several comments wanted the full conversation or reasoning trace.
- View that this is advanced brute force
- Counterexample searches commonly use brute force.
- The commenter said using LLMs as better search seems right.
How to Apply
- Re-verify LLM math results yourself
- When an LLM gives a counterexample or proof, check it with code you run.
- Cross-check large results using multiple methods.
Terminology
Related Papers
The Judge Can Debias Itself: Awareness and Self-Debiasing for Reliable LLM-as-a-Judge
Warning an LLM judge about bias or letting it self-critique improved 20 of 21 model–bias combinations.
90% of US working hours are spent on tasks not covered well by AI benchmarks
Mapping 29 benchmarks to job tasks shows 90% of US working hours lack good coverage
Postgres 19 delayed after AI reviewers find more bugs than humans can fix
LLM code review surfaced so many bugs that Postgres 19 slipped to November and 8-10 features were reverted
Show HN: Gemini 4 Argon plays StarCraft
LLMs wrote StarCraft bots in one hour; GPT-6 Astra scored highest.
A Plausible Explanation Is Enough to Mislead Multimodal Judges
Adding plausible explanations to fixed answers cut MLLM judges' wrong-answer rejection by 12.7–13.8 points.
Neural networks are clean
Looking at neurons directly shows most have one main job, challenging superposition.