Linus Torvalds says that AI fixes drove most of the progress in latest Linux kernel

Linus Torvalds says that AI fixes drove most of the progress in latest Linux kernel — BPF saw more AI-written fixes than bug reports, possibly indicating a tidal shift for open-source maintainers tired of flood of junk reports

It would certainly be a welcome development for open-source projects everywhere.

Most open-source projects have meager resources, and the inrush of AI-provided bug reports and "fixes" has been a bane on their respective maintainers. However, the Linux kernel is well-manned, and code for it generally has a higher barrier to entry than in other places. Linus Torvalds thinks that the flood of AI bug reports may be settling, as the kernel's Berkeley Packet Filter (BPF) portion saw more AI fixes than reports.

Torvalds splits off the changes among three areas: graphics, networking, and sound drivers; networking and the BPF; and miscellaneous changes. He proceeds to state that "much of this was driven by various AI fixes," an apparent underscoring of his stance that AI-written contributions should be embraced rather than shunned — so long as they're subject to the same quality control.

The Linux maestro also relayed a note from Alexei Starovoitov, the BPF maintainer, who claims that this subsection actually saw a higher rate of AI-provided fixes versus reports. This might signal a tidal shift, at least where the Linux kernel is concerned, and would be exceedingly welcome news for other open-source maintainers everywhere tired of dealing with junk reports at the same time they have to handle a higher influx of legitimate bugs, thanks to AI-driven bug discovery.

In Torvalds' own words, "I think most other areas haven't reached that point yet, but it's at least a sign that we may be starting to see the tide starting to turn."

Anyone can make a pointless bug report, but it's hard to deny that the existence of powerful AI bots has prompted a renewed focus on cybersecurity most everywhere. Firefox fixed 423 security bugs in an April release, Chrome found and fixed 1,072 bugs in June — making for more fixes in 2026 than the project's previous eight years — and my personal "favorite," NodeJS, sadly fixed 50 vulnerabilities instead of announcing its closure.

Many open-source projects have complained about high-volume, low-quality AI-driven bug reports, with some even shutting down their bug bounty programs altogether, or placing additional verifications for issue reports. We'd venture that the Linux kernel has a much larger bug-triaging team than your average project, though, meaning it has more resources to sift through the junk.

This tentative conclusion popped up in a message sent yesterday to the kernel mailing list, wherein Torvalds briefly describes the overall shape of the kernel's 7.4-rc2 update. He says that despite the update looking large from a pure file-count perspective, it's actually mostly composed of one- and two-line fixes and is "very straightforward" to implement.

Get Tom's Hardware's best news and in-depth reviews, straight to your inbox.

Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.

Bruno Ferreira is a contributing writer for Tom's Hardware. He has decades of experience with PC hardware and assorted sundries, alongside a career as a developer. He's obsessed with detail and has a tendency to ramble on the topics he loves. When not doing that, he's usually playing games, or at live music shows and festivals.

- -Fran- As a Dev Lead myself and working with agents in several areas of the business and SDLC now, it's really amazing how well it works when you know what you're doing. It's a shame the CEOs and such have given the technology such a bad reputation, but it definitely works.

From what Linus needs to deal with and all the Kernel maintainers, the agents (or just multiple AI agents analysis) can help so much to diagnose and troubleshoot.

Be it local models or paid cloud ones, you can choose how to go about it, but in 2026, we need to get comfortable as software workers to understand and embrace how to use AI to our benefit. It's here to stay and saying otherwise won't change reality.

Not a paid advertisement either, LOL. I just find that, in late 2026, the assistants are working very close to "as originally intended" (or envisioned), which is great.

Regards. Reply