Linus Judges AI Review by Technical Merit

Description

A black, terminal-like mailing-list screenshot shows the contact “Stephen Finucane <[email protected]>,” the subject “Re: Linking Patchwork with Sashiko?”, the date “Tue, 14 Jul 2026 20:06:14 -0700,” links labeled “[thread overview]” and “(raw),” a Gmail Message-ID, and an In-Reply-To address at `linux.dev`. Two repeated blue quoted blocks argue that a proposed restriction would make Sashiko’s goal of helping maintainers unachievable, say the general use of LLMs should be discussed instead of complicating each use case, and characterize reference “[1]” as broadly anti-LLM. Linus Torvalds opens with “Yes. And no,” rejects that stance as the Linux kernel’s position, says he will “put my foot down” as top-level maintainer, tells objectors they may fork or walk away, and calls AI a clearly useful tool whose usefulness is no longer seriously in question even though its eventual economics remain uncertain. He acknowledges maintainer workload and the pain of AI “finding embarrassing bugs,” rejects hiding from the technology, insists LLM tools must help rather than burden maintainers, says nobody is forced to use them but opponents may not prevent others, and contrasts imperfect AI with imperfect “natural intelligence.” He closes that the kernel is about technology; its open-source social dimension is a motivating side benefit rather than the project’s point, it is not a “social warrior” or religious project, and decisions must rest on technical merit rather than fear of new tools—a position prompted by integrating the agentic Sashiko patch-review system with Patchwork.