AI-generated Lean proofs are sloppy and tangled — and that's a failure
bradneuberg · x · 2026-10-19
Engineer Brad Neuberg, quoting Ted Pavlic, argues that true intelligence means not just producing plausible solutions but disseminating them clearly — "if it's not disseminated, it's not science."
His core point: the 700 math proofs AI systems have produced in Lean are sloppy, complicated and tangled. We should hold AI agents to a higher standard — elegant, clean, explainable work, as we'd expect from a human genius. Incomprehensible code or proofs is a failure and a sign of sloppy thinking.
More from AGI Musings
- A few more OOMs of efficiency gains could unlock new financial modeling — xeophon · 2026-10-19
- Economist Daniel Rock: AI circles' intelligence worship is a stumbling block for scientific progress — danielrock · 2026-10-19
- Economist willcb calls Anthropic's econ team survey-derived modeling 'unserious' — willcb · 2026-10-19
- Diffusion paper author Jascha Sohl-Dickstein tells Stanford students the 'safe' academic path may be least safe — FinanceYF5 · 2026-10-19
- Stanford's Noah Goodman Shares First-Pass AI Safety Concept Map for Lecture — noahdgoodman · 2026-10-19
- AI safety researcher: safety concerns won't impede frontier AI revenue growth — menhguin · 2026-10-19