AI Agent Specula Finds 249 Bugs, but Expert Questions Its Approach

tianyin_xu · x · 2026-08-14

Specula, an agentic system that automatically generates TLA+ specs and model-checks code, found 249 bugs (207 new) across 48 open-source systems like MongoDB and Etcd. However, distributed systems expert Murat Demirbas criticizes its approach, saying letting agents drive key decisions is a fundamental flaw and it lacks composition. Author Tianyin Xu acknowledges the critique, calling Specula a useful tool but not the ultimate answer, and suggests putting humans back in the loop.

Original post →

More from coding & agent

coding & agent channel →