GoKawiilAnthropic details how Claude Code team overhauled code review as AI writes 90% of code
Anthropic's Claude Code team published an account of how it adapted code review practices as AI-generated code volume surged. Merged pull requests to the Claude Code repository more than doubled between June and September, while the share of PRs later reverted fell by over 40%. The team says it shifted most initial code reading to automated Claude-based checks that run continuously, reserving human review for higher-level judgment calls.
GoKawiil's interpretation of the reporting above, not reported fact.
The shift suggests that as AI agents generate ever-larger shares of code, traditional line-by-line human review may no longer scale, pushing engineering teams toward automated, continuous checking models. Anthropic's internal data—daily changes in its monorepo rising from about 2,500 to over 30,000 between May and September—indicates the pressure on review processes is intensifying rapidly across the industry, not just at one company. How teams balance speed, error-catching, and shared team understanding could become a broader template for AI-heavy software development.
- Merged PRs to Claude Code's repo more than doubled from June to September while reverts dropped over 40%.
- Anthropic's monorepo saw daily code changes jump from ~2,500 in May to over 30,000 in September.
- The team now treats AI-driven continuous review as a first pass, saving human attention for higher-level scrutiny rather than routine checks.
Source: claude.dev — Addy Osmani, 2026-10-16
Published there as: “How the Claude Code team does code review”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.