Csaba Szepesvari questions whether AI formalized proofs are faithful to what mathematicians mean
CsabaSzepesvari · x · 2026-09-07
- Renowned RL researcher Csaba Szepesvari raises concerns about Anthropic's formalized math proofs: the math community's real worry is whether the formalization faithfully captures what mathematicians intend, not just kernel-checked correctness.
- He argues the formalizations themselves are the real value, potentially unlocking other proofs, but verifying faithfulness is again a specification problem with unclear effort so far. A reply adds risks of proofs bypassing the kernel or exploiting bugs, suggesting comparators could help.
Related event: Claude's FLT Proof Ignites Math Community Scooping-Ethics Debate(26 posts)→
More from Research
- DONIC hits 86.79% on bimanual dexterous tracking, crushing prior SOTA's 9.43% — liuziwei7 · 2026-10-27
- PSNR below 15 dB in papers likely means the metric was misused, researcher warns — kwangmoo_yi · 2026-10-27
- Ortet's first health eval: 34 open-weight models, 50 tasks, generic rankings fall short — thegautamkamath · 2026-10-27
- Scaling-fit discrepancy: output-head FLOPs are ~65% of compute at the smallest rung, ~3% at 535B — dlwh · 2026-10-27
- Marin's dlwh Explains the Scaling Law Fit: Estimating the Loss Floor Early — dlwh · 2026-10-27
- Marin's Open Scaling Law Extrapolates 300× and Predicts Loss Within 0.2% — dlwh · 2026-10-27