Fixes for bugs found by artificial intelligence tools helped make the next Linux kernel unusually large, and Linus Torvalds said Sunday that the way those fixes now arrive is "a much healthier model," The Register reported. For companies that run Linux on servers, cloud systems and appliances, the shift suggests more AI-found fixes will reach them through scheduled releases rather than a stream of late patches.
What Happened
In his announcement of Linux 7.4-rc1 on the kernel mailing list, Torvalds closed the two-week merge window, the period when new code is accepted before weeks of testing on release candidates. He called it big: almost 20,000 commits, or individual code changes, against what he described as a more common figure of about 15,000.
Torvalds said he was much happier to see AI-triggered work land during the merge window than to have it show up mainly as extra fixes during the release-candidate phase. He wrote that he hopes this means AI-found fixes are becoming part of normal development, and noted that many patches are now suggested by AI as well, "although they often get rewritten." He was cautious about the outcome: "But we'll see how the -rc's look."
He offered other reasons for the size. About a third of the patch is AMD graphics-chip register definitions, which he told readers to ignore, and roughly half of the rest is drivers. He also noted that 7.4 is expected to be the next long-term support (LTS) release, a version maintained for years, which can draw extra work. The Register, citing Phoronix, said the release also adds security improvements for virtual machines and support for new processors from AMD, Apple and Google.
Why It Matters
AI tools are finding kernel bugs faster than before, and every fix has to move downstream to Linux distributions, cloud providers and the makers of devices built on Linux. The Primary Wire reported in October, citing The Register, that one Debian kernel update listed 1,313 CVE identifiers, partly because of AI-assisted bug hunting.
The same kind of tools can help attackers. Last week Anthropic said one of its unreleased models wrote a working exploit for a critical Linux kernel flaw in under an hour, BleepingComputer reported. If fixes are folded into regular development, security teams can plan around LTS kernels and vendor updates instead of chasing urgent patches.
What to Watch
The weekly release candidates will show whether the testing phase stays calm, the measure Torvalds himself set. Also watch whether major distributions and appliance makers adopt 7.4 as their next long-term base, which would carry its batch of AI-found fixes into enterprise systems for years.
