Linus Torvalds Endorses AI Coding Tools for Linux Kernel Development

“I realize that some people really dislike AI, but this is an area where I’m willing to absolutely put my foot down as the top-level maintainer,” Torvalds wrote. “Linux is not one of those anti-AI projects, and if somebody has issues with that, they can do the open-source thing and fork it. Or just walk away.”

The Debate Over AI-Powered Review Systems

The controversy was ignited by ongoing discussions regarding “Sashiko,” an agentic Linux kernel code review system. According to its creators, including Google’s Roman Gushchin, Sashiko is capable of independently identifying 53.6 percent of bugs that would otherwise be caught by human reviewers in later commits. However, the tool also generates “false positive” reports for non-existent bugs at a rate estimated to be within the 20 percent range.

Some developers, such as Laurent Pinchart, suggested that Sashiko’s output should be triaged before reaching patch authors, citing guidelines from the Software Freedom Conservancy (SFC). The SFC has recently stated that the open-source community “should support, not just tolerate, those who outright reject LLM-gen-AI systems.”

Torvalds, however, explicitly rejected the notion that the Linux community must accommodate those who demand a total ban on AI-generated contributions. He noted that while he will not force anyone to use AI, he intends to “very loudly ignore people who try to argue against other people from using it.”

Technical Merit Over Ideology

Torvalds emphasized that the Linux kernel project is focused on producing superior technology rather than adhering to social or “religious” agendas. “In the kernel community, we do open source because it results in better technology, not because of religious reasons,” Torvalds stated. “And so we make decisions primarily based on technical merit. Not fear of new tools.”

The maintainer acknowledged that AI tools are not perfect and can sometimes create additional work for maintainers by generating “embarrassing” bug reports or “painful” workloads. Despite these challenges, he argued that the appropriate response is to refine the tools to better serve maintainers rather than ignoring their existence. He further remarked that critics of AI should maintain self-awareness, noting that “it’s not like natural intelligence is always all that great either.”

Changing Perspectives in Open Source

The current stance marks a significant shift from Torvalds’ previous position. As recently as October 2024, he characterized AI as largely “marketing hype” and stated he would ignore it. However, he now views the utility of the technology as established, stating, “It may not have been that ‘clearly’ even just a year ago, but it’s no longer in question today.”

Other senior maintainers have corroborated this shift in utility. Greg Kroah-Hartman, a maintainer of the Linux stable kernel, noted that while the project initially struggled with low-quality “AI slop,” the quality of AI-generated reports has improved significantly as models have evolved. “Now we have real reports,” Kroah-Hartman said. “All open-source projects have real reports that are made with AI, but they’re good, and they’re real.”

Practical Implications for Contributors

The integration of these tools has raised questions about the boundaries of contributor autonomy.

While some projects, such as the Zig language project, have adopted strict policies against LLM-generated code, Torvalds remains committed to keeping the Linux kernel open to AI-assisted workflows. He concluded that the goal remains to ensure that these tools “help maintainers instead of just causing them pain.”

Find more reporting in our Tech section.