Use Claude to draft a kernel fix and review it for races
Linus Torvalds had Claude produce a Linux memory-management fix, tests, docs, and review, while he handled benchmarking.
- What they did
- Linus Torvalds asked Claude to work through a Linux kernel memory-management issue end to end. Claude produced the code, unit tests, documentation, and a debug hook, and also did the first-pass review of its own change. Torvalds kept the benchmarking step for himself.
- How it went
- A race condition still made it into the final result. No benchmark figures or deeper technical details were given, and the writeup cuts off early.
- Tips
- Self-review is not enough for concurrency bugs; have a separate check for locking and ordering issues. Keep a human in the loop before anything is merged or shipped.
Try it yourself with Claude
I’m fixing [bug or feature] in [repo/path]. Draft the code change, tests, docs, and any useful debug hooks. Then review your own patch for race conditions, locking mistakes, and other concurrency hazards. After that, pause and show me the diff summary plus any risks or open questions before I approve anything that would be merged, submitted, or applied. I’ll run the benchmarks and final checks myself.
Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.
Get 3 agent uses like this every morning
With the day's top AI news. Free, and one click to unsubscribe.
Related agent uses
Use parallel agents to build a compiler
Coordinate many coding agents in containers and back them with tests and an oracle while they work on one compiler repo.
Build an anonymous board for AI hopes and worries
They used Claude Code to create a site where people can post anonymously, with one model sorting entries and another finding themes.
Have agents co-build a reference system with role splits
A solo founder divided an AI-assisted build into roles and shipped a 120-model mental-models framework with an MCP server.
Build new agent features inside a guarded Docker sandbox
A Claude-based agent updated its own code in Docker, with every edit checked, committed, or undone automatically.
Use Claude as an async coding partner
They paired with Claude and Jules while coding locally, then copied changes back and forth instead of sharing a live editor.
Use spare AI quota to draft reviewable PRs
A founder pointed an agent at goals and task lists so it could turn leftover Claude and Codex quota into pull requests for review.