On November 2, 2026, Linux creator Linus Torvalds said the project had moved beyond its AI stabilization phase. The goal isn’t to reduce AI-generated code; it’s to bring those patches into the main development cycle before they bottleneck release candidates.

The problem: AI fixes arriving too late
The issue isn’t simply that AI contributes code to Linux. It’s when that work arrives in the development cycle. Torvalds said he’d rather see AI-triggered work enter the merge window than show up mainly as extra fixes during the release candidate period. That matters because stabilization should settle a release, not absorb a late wave of changes.
Fixes arriving then add review pressure just as maintainers prepare a candidate release, and Torvalds’s comments frame timing—not volume alone—as the more important measure of progress. Linux 7.4’s current merge window was reportedly unusually large, with almost 20,000 commits compared with a usual total of about 15,000. In coverage highlighted on November 2, Theregister reported that Torvalds saw the increase as a positive sign if AI-related work is becoming part of normal development.
Why Linus Torvalds prefers earlier AI integration
In Torvalds’s view, the root cause is a mismatch between when AI-assisted patches are produced and when they’re integrated. A late burst of fixes can make stabilization feel like a second development window. Bringing those changes into the regular merge process lets maintainers assess them alongside the broader flow of kernel work.
He called the current approach a “much healthier model.” That describes a workflow, not a claim that AI-generated code can be merged without scrutiny. Reporting also notes that machine-made patches are often rewritten, so human review and editing remain part of the process. The distinction sets a practical standard: the project can benefit from assistance without treating generated patches as finished work. His comments don’t establish a new formal rule for contributors or say how many commits in the window involved AI.
Three paths for AI contributions
There are three broad ways to handle AI-assisted changes. The first is to bring them in during the ordinary merge window, the approach Torvalds welcomed.
The drawback is that a larger window still means more commits to review, as the Linux 7.4 figures show. The second path is to defer AI-related fixes until the release candidate period. That may keep the initial merge window smaller, but it risks concentrating review pressure late in the cycle.
| Approach | Benefit | Trade-off |
|---|---|---|
| Merge-window integration | Earlier assessment by maintainers | Larger window means more commits to review |
FAQs
What did Linus Torvalds say about the Linux kernel’s stabilization phase?
Linus Torvalds announced that the Linux kernel has moved out of the stabilization phase, indicating a shift towards a more robust and dynamic development approach.
How does Linus Torvalds describe the current model for AI contributions to Linux?
Linus Torvalds described the current model for AI contributions as a “much healthier model,” suggesting improved collaboration and integration within the Linux community.
Read Next
- OpenAI Pledges to Buy Google AI Chips in First Move Away From Google, Nvidia, and Reliance
- Apple iPhone Duo vs iPhone Air: Which Should You Choose?
- ONDC DigiDukaan: How India’s 1.4 Crore Kiranas Are Going Digital, One Restock at a Time
Was this article helpful?
Your feedback directly improves future articles on this site.





