Researcher: verifying 70–100-page proofs by hand takes months; AI could fix that
kfountou · x · 2026-09-13
A researcher describes the bottleneck in AI-assisted mathematics: three of his long-term projects sit unverified, each with 70–100-page proofs that would take months to check by hand — and even auto-formalizing them with Codex or Claude could still take months. Better AI verification, he argues, would dramatically speed up judging correctness.
Related event: RL Scholar Withholds Long Protests Over Poor AI Writing Quality(3 posts)→
More from Research
- AI formal verification: Aristotle proves memory safety of OpenTitan root-of-trust C code — satnam6502 · 2026-10-31
- MIT CSAIL's 'Machine Studying': agents that pre-study corpora instead of searching on-the-fly — eshorten300 · 2026-10-31
- ARC Prize 2026 closes in 3 days as Yi-Chia Chen tops ARC-AGI-3 at 75.56% — GregKamradt · 2026-10-31
- NVIDIA Nemotron 3.5 scales LLM self-verification, +9-10 points on agentic benchmarks — sarahcat21 · 2026-10-31
- Unofficial OpenTargets MCP Server Brings Gene-Drug-Disease Data to LLMs — modelcontextprotocol · 2026-10-31
- Researchers find 5 personality structures in LLM weight space, Big Five-like — luke_drago_ · 2026-10-31