Linus Torvalds Embraces AI-Powered Code Review, Rejects Anti-AI Stance in Linux Kernel
Key Takeaways
- ▸Linus Torvalds has explicitly committed Linux to using AI-powered code review tools, rejecting broader anti-AI sentiment in open-source communities
- ▸Sashiko, Google's AI code review tool, identifies 53.6% of bugs in kernel patches, finding errors human reviewers miss despite existing review processes
- ▸Torvalds views AI as a pragmatic tool like any other development utility, taking a hardline stance that those opposed can fork the project
Summary
Linus Torvalds, the Linux kernel's creator and maintainer, has publicly rejected anti-AI sentiment in the open-source community and declared that Linux will continue using AI-assisted tools like Sashiko, a machine learning-powered code review system created by Google engineer Roman Gushchin. In a forceful response on the Linux kernel mailing list, Torvalds stated that "Linux is not one of those anti-AI projects" and made clear that developers who object to AI tools can either fork the project or leave. This position contrasts sharply with other major open-source projects like Gentoo Linux, Curl, and Ghostty, which have implemented strict limitations or outright bans on LLM-generated code contributions.
Sashiko is an opt-in code review tool that analyzes kernel patches across multiple stages and claims to identify 53.6% of bugs in proposed patches—reportedly exceeding the effectiveness of human-only review of the same code. The tool operates in an advisory capacity, flagging issues through comments rather than taking autonomous action. When developer Laurent Pinchart suggested implementing additional triage of Sashiko's output based on Software Freedom Conservancy guidelines, Torvalds directly countered the proposal, arguing such restrictions would undermine the tool's effectiveness and represented an overly ideological stance.
Torvalds emphasized that AI should be viewed pragmatically as a development tool no different from other utilities, noting that its usefulness is now "no longer in question." He highlighted that Sashiko "keeps finding embarrassing bugs" that human reviewers had missed, and challenged resistance to AI adoption by observing that "it's not like natural intelligence is always all that great either." The statement represents an evolved perspective from Torvalds, who dismissed AI tools as overhyped as recently as 2024 but now champions their practical utility for improving Linux kernel quality.
- This marks a significant shift in Torvalds' position—he dismissed AI tools as overhyped in 2024 but now champions their demonstrated effectiveness
- Major open-source projects diverge sharply on AI: Linux embraces it while Gentoo, Curl, and Ghostty restrict or ban LLM-generated contributions
