What a $900M Bank Error Teaches Us About AI Agents - John Feminella, Two Sigma

Shows how to grant agents real autonomy without blanket lockdown, using five concrete controls including purpose-scoped tool access and classifier-routed approvals to avoid approval fatigue. Useful for anyone designing agent governance.

• AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
• eval — A repeatable test for AI quality — a set of tasks plus scoring — used the way software teams use test suites, because model output is too variable to judge by eyeballing.

• videoHow to Test Your Agents on Real Tasks With Harbor - Xander Song, Arize AI
• videoAgents That Self Improve While You Sleep - Fuad Ali, Arize AI
• video300 Agents, 60 People: How We Secured Our Own AI Agents - Aiden Sochowski, Runlayer
• videoHow a Trading Firm Ships New AI Models the Same Day - Noyan Tokgozoglu, Optiver

• articleCrossing the Threshold: How OpenAI and Anthropic Approach Autonomy and Safety in Financial AgentsRichard MacManus
• videoReal AI Agent Attacks and What Would Have Stopped Them — Andrey Rybka, BloombergAI Engineer
• postWhy Bottom-Up AI Pilots Stall and Workflow Redesign Doesn'tAndrew Ng

Checking sign-in…

Loading comments…