Torvalds to anti-AI critics: Fork the Linux kernel or pipe down

Linus Torvalds has drawn a hard line in the silicon. The Linux creator told the kernel mailing list in no uncertain terms that the project is not an anti-AI safe space, and anyone who thinks otherwise should take their ball and go home — or fork the code. The blunt directive was a direct response to pushback against Sashiko, a Linux Foundation-backed AI tool designed to automate code reviews by pulling patches straight from mailing lists and Git repos and running them against various LLMs.

Torvalds didn’t mince words. “Linux is not one of those anti-AI projects,” he wrote, adding he would “absolutely put my foot down as the top-level maintainer” on the matter. The debate kicked off when developer Roman Gushchin flagged a pervasive anti-LLM attitude that was kneecapping Sashiko’s utility for overworked maintainers. Torvalds acknowledged that some folks in the community hold that position, but made clear it doesn’t represent the project’s stance. For him, this isn’t an ideological purity test — it’s about what works.

He conceded that AI can be “painful,” generating extra triage work or surfacing embarrassing bugs nobody wants to look at. But his counter to the skeptics is almost refreshingly pragmatic: pretending the tools don’t exist is a worse strategy. “Anybody who doubts that clearly hasn’t actually used it,” Torvalds said, noting that AI’s usefulness is much more obvious now than it was even a year ago. The kernel’s north star remains technical merit, he stressed, “not fear of new tools.”

What’s striking here is how Torvalds treats LLMs the way he’s always treated hardware architectures or version control systems — as tools to be evaluated, not movements to join or reject. He won’t force anyone to use AI, but he also won’t tolerate people trying to talk others out of it. The message to the community is classic Torvalds: if you don’t like how the project is run, open source gives you exactly one meaningful recourse.