Slashdot has reported:

Linux kernel contributors no longer need to disclose when they use large language models to produce code, following a decision at the Maintainers Summit on October 8, 2026, in Prague, Czechia. A message to the Linux mailing list announced the change: submissions no longer require the “Assisted-by” tag, so developers can submit AI-generated code without identifying it as such.

The change prompted correspondence about removing the disclosure requirement. Linux creator Linus Torvalds argued against banning AI tools, comparing such a ban to getting rid of text editors on the grounds that they might help someone write a commit message without mistakes. “Do you not see how UTTERLY STUPID those kinds of rules would be?” he asked, adding that he expects people working on the kernel to “understand what they are doing.”

“Quite honestly, in my experience, the people I respect and call intelligent are not the ones that shy away from using new and interesting tools,” Torvalds wrote. He has also said he uses AI to help understand patches that are difficult to follow.

This week, Torvalds posted a patch developed with help from an AI assistant. The development branch aims to address a longstanding problem involving heavy cache contention when many CPUs read the same file, a situation that can arise with shared libraries, databases and web servers. Several developers had tried to solve the issue over the years, with varied results, but none had produced what was described as a suitable all-encompassing fix.

Torvalds said his proposed solution produced a 23-fold improvement in a dedicated synthetic benchmark. He cautioned that the real-world effect would be much smaller, saying that “no real load” would see those figures. On Tuesday, he submitted a series of 10 patches.

Torvalds used Anthropic’s Claude Code for the work. The New Stack reported that the assistant wrote the implementation, tests and initial documentation, and reviewed the code under Torvalds’s direction. Torvalds ran the benchmarks himself, despite having previously criticized developers who claimed AI wrote 99 per cent of their code.

Reviewers have since found a race condition and raised concerns about memory reclamation. Torvalds handed further development to Kiryl Shutsemau, a Linux kernel developer who had previously worked on the same bottleneck. The kernel submissions policy change and the patch work both reflect Torvalds’s stated expectation that developers understand the code they submit, regardless of the tools used to produce it.