Making AI-generated code safer to ship. Consulting: advising, building, coaching on AI
I build Mindrealm: it catches the slop your AI agent ships confidently, then makes it fix it. 75+ deterministic checks across Go, Rust, TypeScript, and Python.
Markdown supported · ⌘↵ to post
Mateo Ruiz
One distinction I think more teams should make is between "tests passed" and "requirements verified." AI agents are getting very good at satisfying existing tests, but that doesn't always mean they've solved the original problem or handled the edge cases. Treating generated → reviewed → accepted as separate stages creates much healthier engineering discipline than assuming a green CI pipeline is the finish line. The acceptance ledger idea is also interesting because it captures why a change was accepted, not just that it was.