The AI plan looked perfect — so it built a fish tank
jimmykoppel · x · 2026-11-02
Jimmy Koppel shares a funny agent failure: the AI's plan read 'a line of buttons at the top, and a bass control below', the user approved it — and it built a fish tank. A vivid illustration of how an apparently sound plan goes wrong through ambiguous interpretation, and why plan-approval steps in agents deserve scrutiny.
More from coding & agent
- One line in your system prompt is silently killing your prompt cache hits — Arindam_1729 · 2026-11-02
- mitsuhiko on Python JIT: agents could crack it, the blocker is grind not humans — mitsuhiko · 2026-11-02
- 13 Claude agents replaced a $300K marketing department, driving 52.4M views — aryanXmahajan · 2026-11-02
- Engineer ports entire Python codebase to Go with AI in just days — dosco · 2026-11-02
- Databricks engineer's outsider take on Anthropic's Scion: 7 people, 5 months, borrowed ideas — ricklamers · 2026-11-02
- Thorsten Ball: drop branches and PRs, push to main — the industry is catching up — RealGeneKim · 2026-11-02