https://petbrains.dev/blog Build notes and essays from the Petbrains workshop — shipping real products with AI agents, documented as we go. New posts land here first. en-us https://petbrains.dev/blog/plausible-is-not-correct https://petbrains.dev/blog/plausible-is-not-correct LLMs generate plausible code, not correct code. The techniques that make AI coding agents work all reduce to one move: constrain and verify. Wed, 10 Jun 2026 00:00:00 GMT