Agents build Lean models of every CPython 3.15.2 function, validated on 50,000+ generated inputs
ChrSzegedy · x · 2026-11-10
A viral project describes a multi-agent formal verification pipeline for CPython: for each function in CPython 3.15.2, an agent writes a Lean model mirroring the C or Python source. The model is run against python3 on 50,000+ generated inputs until they agree, properties are then proved about it, and a second agent reviews every function — combining LLM agents, differential testing, and automated proving at interpreter scale.
More from coding & agent
- 10 open-source repos that review your code with AI, from Alibaba to Anthropic — victor_explore · 2026-11-10
- The dialectics of AI coding: LLMs can't write prod code unsupervised — neither can humans — ShenRaphael · 2026-11-10
- SteerCamp's multi-agent crew holds daily standups, reporting progress and blockers like employees — pramodk73 · 2026-11-10
- Running Codex for two days to reimplement an outdated Android book's projects and retake screenshots — burkov · 2026-11-10
- Unverified: Fable 5.2 tops FrontierCode 2.0 at 49.1%, half the cost of Opus — FinanceYF5 · 2026-11-10
- Mnemos memory system gives Claude cross-session mood continuity, dev claims — RileyRalmuto · 2026-11-10