Linus Torvalds says he is now using AI tools to help develop and review Linux kernel code, presenting the approach as a way to make work more enjoyable and certain outputs more readable. ZDNet reports that Torvalds is using AI to develop kernel code, while Phoronix describes him continuing that practice alongside public commentary about how AI fits into his workflow.

Across outlets, Torvalds frames AI as a tool rather than an authority, saying he has moved from earlier skepticism to actively liking AI for programming support. At Open Source Summit Europe, Slashdot relays his view that developers should be careful when using AI for “real and important” tasks, particularly in a high-stakes environment like kernel maintenance. He also acknowledges that AI can help identify vulnerabilities and improve the kernel, but that it can overwhelm maintainers with poor-quality or superficial bug reports.

Phoronix adds that Torvalds used Claude Code to draft code and create supporting tests for a kernel I/O-related improvement, and he also used AI/LLMs to make patches more decipherable. He says this process helped uncover more long-standing bugs, while emphasizing Linux progress through incremental, steady changes rather than sweeping revisions.