Linus Torvalds Uses Claude Code for Linux Kernel

Linux founder Linus Torvalds has started integrating generative artificial intelligence directly into his kernel development workflow. Working on a recent optimization effort, Torvalds utilized Anthropic's Claude Code tool to construct eight experimental code changes intended to troubleshoot and resolve a performance bottleneck within the Linux operating system kernel.

According to project documentation, the Claude Code assistant operated under Torvalds' explicit oversight to craft the patch set. Torvalds confirmed in the documentation that the AI assistant generated the raw code, the accompanying test suites, the review materials, and the initial draft of the documentation text. Although these eight modifications are currently categorized as experimental and have not been officially merged into the upstream Linux kernel, the trial highlights a major shift in how core system infrastructure can be maintained.

This experiment represents a clear change of heart for Torvalds, who had previously expressed skepticism regarding the reliability of automated coding tools. Addressing attendees at the Open Source Summit Europe conference in Prague, Torvalds explained his shifting viewpoint on generative models. He admitted that he used to think the technology "just wasn't very good," but acknowledged that he has now reached a point where he can "really enjoy using AI" to assist his programming tasks.

For software engineers and open-source maintainers, Torvalds' practical application of Claude Code underscores the growing maturity of AI agents in low-level systems programming. While human supervision and rigorous testing procedures remain essential to ensure stability before merging code into main repositories, the adoption of LLMs by leading kernel developers indicates that AI assistants are shifting from novel conveniences to genuine productivity tools in complex software environments.