Skip to content- Sources: primary
- Summary: In a post quoted by Simon Willison, Addy Osmani, Morgan Westlee Lunt, and Qing Wang write that review now runs continuously while code is written, not as a gate at the PR. They call "did a human read the code?" an imprecise proxy for choosing the right automated or human check for each risk.
- Why it matters: This is a first-party account of a team moving routine review to an agent and keeping human attention for higher-level scrutiny, a model other teams can compare against their own review gates.