upbeat quartz
Hey all! Quick question for people who use AI coding agents (Cursor, Copilot, etc.) at work.
When an agent gets stuck, loops, or starts producing sketchy output, what do you actually do next?
Do you retry it, dig through logs, take over manually, ask someone on Slack, ping on-call, etc.?
Looking for real workflows you’ve used recently!
pale apex