Linus Torvalds says he uses AI to cover his weak spots

By Splitfeed EditorialLast updated Oct 12, 2026, 10:30 UTC

In October 2026, Linus Torvalds, the creator and principal maintainer of the Linux kernel, publicly stated that he uses AI tools to compensate for areas where he considers himself weak. Speaking across several outlets between October 9 and October 12, he described applying AI to make existing code more readable, and said the process surfaced additional bugs that had not been caught before. He framed the decision as consistent with advice he gives others, saying he "did what I preach," and stated plainly that he "really likes using AI."

The timing gives his comments more weight than they might otherwise carry. The Linux kernel community is currently working through an active debate about whether and how to regulate code contributions generated by large language models. Torvalds is not a neutral bystander in that conversation: his personal practices and public statements carry significant influence over the project's culture and eventual policy direction. His endorsement does not settle the policy question, but it does shift the terms of the debate inside one of the most consequential open-source projects in existence. The core dispute is whether LLM-generated code introduces reliability, licensing, or quality problems serious enough to warrant formal restrictions.

None of the three camps had published specific reactions to this story at the time of writing. The Pro-AI camp would typically treat a figure of Torvalds's stature endorsing AI tools as strong validation, particularly his claim that AI use led to concrete improvements like bug discovery. The Anti-AI camp would likely focus on the broader community debate rather than Torvalds's personal habits, arguing that his individual comfort with the tools does not resolve questions about provenance, copyright, or quality control for the contributions of thousands of other developers. The Middle Ground camp would probably echo the framing Torvalds himself used, pointing to targeted, task-specific use as the responsible model, distinct from wholesale reliance on generated code.

The argument will sharpen once the Linux kernel community produces formal guidelines on LLM-generated contributions. Any written policy, and the process by which it is adopted or rejected, will clarify how much Torvalds's personal endorsement actually shapes collective practice versus how much the community decides the two questions are separable.