A quote from Addy Osmani, Morgan Westlee Lunt, and Qing Wang

Why it matters

The Claude Code team has rethought code review by questioning whether a human actually reading code is a meaningful safeguard. In an article by Addy Osmani, Morgan Westlee Lunt, and Qing Wang, the team argues that asking whether a human reviewed a change is an imprecise proxy for the real question: what could go wrong with this change, and which check, automated or human, is best suited to catch it. Reframing the problem also shifted the timing of review. Previously, review functioned as a gate at the end of development, with a colleague examining a pull request after the code was written. Under the new approach, review runs continuously as work proceeds, compared by the authors to spellcheck. Most changes are reviewed and fixed by Claude at least once before a human looks at them, which catches routine mistakes and preserves human attention for higher-level questions and scrutiny. The piece matters because it documents a working team moving from human-gated review to automated continuous review, a model other engineering organizations may weigh as AI coding assistants become standard practice.