Your AI Agent Said Done: Why {status:success} Proves Nothing About the Goal
GlenBradley · x · 2026-09-21
Glen Bradley walks through a real network ops failure: a corporate WLAN mapped to VLAN 70 instead of 40. An AI agent proposes the fix, the controller returns {"status":"success"}, and the ticket closes — yet users still can't connect.
The core point: command success is evidence about the command, not the goal. After the change, auth might still fail, DHCP/DNS might be broken, guest isolation might be violated, or VLAN 70 may have been wrong but not the outage's cause. These are distinct facts that today's agents routinely collapse into one word: Done.
He argues the right sequence is proposal → accepted → verified state, and has written a longer piece on designing a control plane that keeps command state and goal state separate.
Related event: Agent Says Done, But Is It Fixed? Splitting State From Outcome(6 posts)→
More from coding & agent
- Photoroom reclaims ~1,000 idle GPU hours a week with SkyPilot across Slurm and Kubernetes — skypilot_org · 2026-11-12
- Google launches Gemini Decisions API: scores 1,000 questions per pass, output tokens free — dean_rie · 2026-11-12
- Zed v1.28 adds background auto-fetch for current and default branch remotes — dean_rie · 2026-11-12
- Indie Britpop AI music video built with Nano Banana Pro, Kling, and free lip-sync via WAN2GP + LTX 2.5 — cocktailpeanut · 2026-11-12
- Cloudflare Talk: Swapping Some Agent LLM Calls for Fast Models Like Jev Gives Big Speedups — michellechen · 2026-11-12
- Engineer Curates 15 AI Agent Design Patterns Across Single-Agent, Multi-Agent and Feedback Tiers — MaryamMiradi · 2026-11-12