Code & dev Story
Linux 7.4-rc2: Linus Torvalds will stop pointing out that AI drives many kernel fixes
Linus Torvalds said on 8 November 2026, in his Linux 7.4-rc2 mailing-list message, that he will stop repeating that AI is behind a large share of kernel fixes, treating it as a given.
This story was written with the help of artificial intelligence, checked automatically, then reviewed by an editor before publication, under our AI policy.
What Torvalds said
Linus Torvalds made the remark in his mailing-list message announcing Linux 7.4-rc2, the second release candidate of the next kernel (Numerama). The message is dated 8 November 2026. He now treats it as settled that AI is behind a large share of the fixes going into the kernel.
In his words: "I think I'll stop repeating that, and we can just take it for granted." (LKML). In practical terms, this is a change in what he says in his release messages: he will no longer point out the role of AI each time. The available material does not mention any formal change of rules, and it does not quantify the share of fixes that come from AI.
The remark is therefore a statement of his own view, not a measurement. Readers should treat the phrase "a large share" as Torvalds's characterisation rather than a figure that has been published.
A big release candidate
The same message describes the release in blunt terms: "Ok, a big -rc2 to go with the largest ever -rc1." (LKML). Linux 7.4-rc2 thus follows the largest -rc1 ever published.
A large part of the diffstat, the summary of changes between the two release candidates, comes from automated Coccinelle updates. That detail matters for interpretation: the size of the -rc2 does not by itself indicate how many individual fixes were contributed, or by whom.
BPF: a contributor's claim
Alexei Starovoitov states that, for BPF, fixes are now arriving faster than new bug reports. Torvalds replies that this is great if true, but that most other areas of the kernel have not reached that point.
This is a contributor's assertion about one subsystem, and the available material gives no figures to back it. Torvalds's reply is conditional: he does not confirm the claim, and he explicitly limits it to BPF. The picture for the rest of the kernel, in his view, is different.
Maintainer workload and rules for AI agents
Numerama reports that kernel maintainers are finding it hard to keep up with a surge of AI-generated bug reports, many of them duplicates. The outlet adds that Torvalds had described dealing with them as needless agitation in May.
Separately, a documentation change in Linux 7.4-rc2 seeks to require AI agents to supply information when they submit a patch. The available material does not say which information would be required, or whether the change will survive into the final release.
Taken together, these points show two sides of the same trend as reported: AI is credited with a large share of fixes, while the volume of AI-generated reports weighs on maintainers. How the two balance out across the kernel is not established by the material available, and the BPF comment suggests the situation varies by area.
FAQ
What did Linus Torvalds say about AI and the Linux kernel?
He treats it as given that AI is behind a large share of kernel fixes and says he will stop repeating it.
When did he say it?
In his message accompanying Linux 7.4-rc2, dated 8 November 2026.
Will AI agents face new requirements?
A documentation change in Linux 7.4-rc2 seeks to require AI agents to provide information when submitting a patch.
Is the BPF claim confirmed for the whole kernel?
No. Alexei Starovoitov says fixes now arrive faster than new bug reports for BPF; Torvalds says most other areas are not there yet.
Sources
- https://lkml.iu.edu/2611.1/01530.html, lkml.iu.edu, (primary source)
- « On peut simplement le tenir pour acquis » : Linus Torvalds renonce à signaler les correctifs dus à l’IA dans Linux, Numerama,