Dev Uses AI-Assisted TLA+ Formal Verification, Finds Hole in His Own Nethack Fix
davidbau · x · 2026-09-28
Developer davidbau shares how, with AI assistance, he's learning formal methods: after an interactive wire protocol kept misbehaving during server hot-updates, Claude suggested using TLA+ to formally verify the protocol design. He had already found a Nethack bug and — via formal verification — discovered a hole in his first fix, illustrating how easily issues slip through manual assumption-checking.
Related event: MIT Researcher Uses AI and Formal Verification to Uncover NetHack Bugs(3 posts)→
More from coding & agent
- Agents flip tooling choices: SCAD and ThreeJS beat GUI CAD and Unreal on speed — _Stocko_ · 2026-09-28
- Hono creator Yusuke joins VoidZero, stays Cloudflare developer advocate — cnakazawa · 2026-09-28
- Microsoft Research runs 1K+ coding agents with no central orchestrator, lifting pass rate to 55% — omarsar0 · 2026-09-28
- MongoDB launches Agent Skills to guide AI coding agents on schema and indexing — TheTuringPost · 2026-09-28
- Dev builds flight-search MCP that queries 200 sites so Claude can book for you — Calm_Cartographer324 · 2026-09-28
- User finds Opus 5.5 usage on Claude a far better deal than Astra on Codex — RexDouglass · 2026-09-28