Linus Torvalds is embracing AI tools for Linux kernel development, drawing criticism from some corners of the open-source community while arguing that the technology is a net positive when used by competent programmers.
Roy Schestowitz, editor of Tux Machines, described Torvalds' stance as 'Linus Torvalds Turns to Slop.' In the podcast Pivot to AI, podcaster David Gerard said 'Linux kernel goes full AI slop code.' Commenters on Tom's Hardware also voiced skepticism, with one user writing that the move was 'a publicity stunt from his AI overlords [The Linux Foundation].'
Despite the backlash, Torvalds told The Register in a note that the argument against using AI for commit messages is 'well-intentioned, but I think it's entirely misguided. I really think we want competent people to use AI as a tool.' He said the real problem is 'clueless people using AI' who generate drive-by patches without engaging with the kernel community. He said he wants to discourage that, not people who use AI and 'truly want to learn.'
At the Open Source Summit Europe, Torvalds described 'vibe-coding' as 'a wonderful way to find joy in programming' and 'a gateway drug' to programming.
Long-time kernel maintainer Greg Kroah-Hartman told The Register the quality of AI-generated reports has improved. He said months ago the project received what they called 'AI slop', obviously wrong or low quality security reports. Then, 'something happened a month ago, and the world switched. Now we have real reports.' He said all open source projects are now receiving good, real AI-generated reports.
The deprecation of the 'Assisted-by' tag was announced by developer Sasha Levin at the Linux Kernel Maintainers Summit. The change was described as the maintainers' decision.
Torvalds acknowledged two distinct 'AI problems' with the kernel: a large codebase where AI finds real bugs, adding to the workload of already-overworked maintainers; and people who use AI to generate patches without understanding the code. He argued these are not problems with AI itself but problems 'exposed by AI.' He said the underlying issue is how people misuse AI, not the technology itself.
Red Hat's director of software engineering, Christian F.K. Schaller, created an open-source project called hw-harness that uses AI to reverse-engineer Windows drivers for Linux more easily. The project automates slow manual methods of figuring out how hardware behaves, potentially easing the development of Linux hardware drivers.
