Full Breakdown
Linus Torvalds Defends AI-Assisted Code Review in the Linux Kernel
7/18/2026, 12:19:16 PM
Core Decision and Rationale
Linux creator Linus Torvalds publicly reaffirmed that artificial-intelligence tools are permissible in kernel development. In a lengthy reply on the Linux Kernel Mailing List (LKML), he declared that the project is “not one of those anti-AI projects” and warned dissenters that they may “fork it” if they cannot accept AI-generated contributions. Torvalds framed AI as “just like other tools we use” and emphasized that its utility is now “no longer in question.”
Background & Context
Earlier in 2024, Torvalds dismissed AI-based tooling as overhyped, while several open-source projects—including Gentoo Linux, Curl, and Ghostty—imposed bans or strict limits on LLM-generated code. The debate resurfaced when developers began employing Sashiko, an opt-in, multi-stage code-review system that automatically comments on kernel patches. Sashiko’s creators, including Google engineer Roman Gushchin, argue that the tool can detect more than half of the bugs in submitted patches without taking autonomous action.
Key Figures & Groups
- Linus Torvalds – creator and top-level maintainer of the Linux kernel.
- Roman Gushchin – co-creator of Sashiko, representing Google’s involvement.
- Laurent Pinchart – developer who suggested triaging Sashiko output per the Software Freedom Conservancy (SFC) guidelines.
- Greg Kroah-Hartman – maintainer of the stable kernel series, noting the evolution from “AI slop” to reliable AI reports.
- Theodore “Ted” Ts’o – senior kernel maintainer raising nuanced concerns about mandatory AI support.
- James Bottomley – kernel developer emphasizing maintainer authority over tool choice.
- Software Freedom Conservancy (SFC) – organization that issued an AI-use policy recommending support for contributors who reject LLM-generated code.
Data & Statistics
- Sashiko reports a 53.6 % bug-detection rate on patches that have already undergone human review, a metric the project claims exceeds human performance.
- Its false-positive rate is estimated to lie “within 20 %,” though precise measurement remains difficult.
- Torvalds highlighted that the tool “keeps finding embarrassing bugs,” suggesting it uncovers issues missed in prior reviews.
Official Statements & Responses
Torvalds’ LKML comment underscored that AI tools are optional but welcomed, and that “the solution is to make sure those LLM tools help maintainers instead of just causing them pain.” He added that he will “very loudly ignore people who try to argue against other people from using it.”
Greg Kroah-Hartman corroborated the shift, noting that earlier “AI slop” has been replaced by “real reports” that are useful for security analysis.
James Bottomley clarified that contributors cannot veto the tools maintainers employ, stating that a patch will be rejected only if the tool’s use infringes on others’ rights.
Ted Ts’o raised questions about the practicality of obligating the kernel to accommodate every anti-AI stance, especially when automated backports or LLM-assisted bug analysis are involved.
Criticism & Opposition
Some developers remain skeptical of AI-generated code, citing concerns over quality, security, and the potential for “embarrassing bugs.” The SFC’s policy urges projects to “support, not just tolerate, those who outright reject LLM-gen-AI systems,” prompting debate over whether the kernel should actively accommodate such preferences. Ted Ts’o’s remarks reflect this tension, asking whether the community must “bend over backward” to satisfy all patch authors who object to AI tools.
Verbatim Quotes
- “I realize that some people really dislike AI, but this is an area where I'm willing to absolutely put my foot down [...] Linux is not one of those anti-AI projects, and if somebody has issues with that, they can do the open-source thing and fork it. Or just walk away.” — Linus Torvalds, Linux kernel maintainer
- “AI is a tool, just like other tools we use. And it's clearly a useful one. It may not have been that 'clearly' even just a year ago, but it's no longer in question today.” — Linus Torvalds
- “I will very loudly ignore people who try to argue against other people from using it.” — Linus Torvalds
- “Months ago, we were getting what we called 'AI slop,' AI-generated security reports that were obviously wrong or low quality.” — Greg Kroah-Hartman, Linux stable kernel maintainer
Why It Matters
Torvalds’ unequivocal endorsement signals that AI-assisted tooling is now an integral part of Linux kernel development. By positioning AI as a standard engineering aid rather than a controversial novelty, the Linux community may see accelerated patch review cycles and broader adoption of automated analysis, while still navigating legitimate concerns about code provenance and quality.
