When nobody says NO, who authorized YES? Rethinking AI agent governance
Portotify · reddit · 2026-10-12
A Reddit post challenges the assumption that monitoring answers the authorization question for AI agents. If an agent issues a refund, logs are complete and nobody intervenes — but nothing establishes what gave it permission for that specific action. Timeouts that auto-approve have the same flaw: designing a workflow to proceed isn't the same as authorizing every action it allows. The author isn't calling for human approval on everything — agents should act autonomously within their authority — but argues that treating the absence of objection as evidence of authority hollows out governance entirely.
More from coding & agent
- He built an AI production team with Codex: 15 clips went from half a day to 1 hour — hugobowne · 2026-10-12
- Harness Learning: RL-Trained Proposer Adapts Agent Harnesses at Test Time Without Weight Updates — pmddomingos · 2026-10-12
- Matthew Berman: I Don't Install Apps Anymore, I Just Tell My Agent to Do It — MatthewBerman · 2026-10-12
- Pedro Domingos: AI harnesses are the flavor of the month — pmddomingos · 2026-10-12
- How Do Junior Engineers Become Seniors Without Coding in the AI Era? — KlausCodes · 2026-10-12
- Garry Tan amplifies take: personal software stays ephemeral until agents maintain apps themselves — garrytan · 2026-10-12