Shayan Rais
e9851f926b
Adopt YY-M-D naming scheme (year-month-date, month as numeric) for chronological sortability. Update image references in the code-review / test-time-compute post. Co-Authored-By: Claude <noreply@anthropic.com>
1.8 KiB
1.8 KiB
Code Review & Test Time Compute — Tips from Boris Cherny
A summary of insights shared by Boris Cherny (@bcherny), creator of Claude Code, on March 10, 2026.
| ← Back to Claude Code Best Practice |  |
1/ Introducing Code Review
New in Claude Code: Code Review. A team of agents runs a deep review on every PR.
- Built for Anthropic's own team first — code output per engineer is up 200% this year, and reviews were the bottleneck
- Boris has been using it for a few weeks and found it catches many real bugs he would not have noticed otherwise
- When a PR opens, Claude dispatches a team of agents to hunt for bugs
2/ Test Time Compute & Multiple Context Windows
Roughly, the more tokens you throw at a coding problem, the better the result. Boris calls this test time compute.
- Using separate context windows makes the result even better — this is what makes subagents work, and why one agent can cause bugs and another (using the same exact model) can find them
- Similar to engineering teams: if Boris causes a bug, his coworker reviewing the code might find it more reliably than he can
- In the limit, agents will probably write perfect bug-free code — until then, multiple uncorrelated context windows tends to be a good approach
