How AI Tools Are Changing Kernel Development Practices
Linus Torvalds spent Sunday afternoon reviewing code for the upcoming Linux 7.3-rc2 release, a routine task he performs most weekends. The kernel maintainer shared a lighthearted comment during a recent interview, suggesting artificial intelligence might have contributed to the release's size without proper human oversight. His remarks came during a conversation at the Open Source Summit in November 2025, where he discussed development practices with fellow engineer Dirk Hohndel. The exchange highlighted ongoing tensions between automation and manual review in open source workflows.
Breaking news
Town AI Secures $108 Million Series B, Valued at $1 Billion After $55 M Pre‑Series
Apple Prepares for Major Product Launch Event on October 28
Apple Releases Release Candidates for watchOS, tvOS and visionOS 27.2
Apple Limits iPhone Duo and iPhone 18 Pro Max to Reservation-Only Sales in Japan and South KoreaTorvalds explained that while automated tools assist in generating code, the responsibility for verifying changes ultimately rests with human maintainers. He noted that the scale of modern contributions can make thorough inspection challenging, especially when relying on AI-assisted submissions. His joke underscored a genuine concern about maintaining quality control as development processes evolve. The Linux kernel project continues to balance innovation with stability, relying on a global network of volunteers to vet each patch before integration.
What Safeguards Exist Against Overlooked Changes
The maintainer pointed out that AI-generated code snippets are becoming more common in developer submissions, though they often require significant refinement. Torvalds emphasized that the project does not prohibit such tools but insists on rigorous testing and human validation. He recalled instances where AI-produced patches contained subtle logical errors that only emerged under specific workloads. This reinforces the project's longstanding principle that all code, regardless of origin, must meet the same standards before inclusion.
When asked about preventing oversized or inadequately reviewed releases, Torvalds cited the multi-stage review system as the primary defense. Changes pass through subsystem maintainers before reaching his final approval queue, creating multiple checkpoints. He also mentioned that release candidates like 7.3-rc2 are specifically designed for broad testing across diverse hardware configurations. The process allows developers worldwide to identify issues that might not appear in limited internal testing, serving as a crucial quality gate before stable release.
Did Torvalds claim AI wrote Linux 7.3-rc2? No, he made a humorous remark suggesting AI might have contributed to the release's size without proper checking, but clarified that all code undergoes human review.
Frequently Asked Questions
How does the Linux project handle AI-generated contributions? The project allows AI-assisted code but requires the same testing and validation standards as manually written submissions, with maintainers responsible for final approval.
What is the purpose of release candidates like 7.3-rc2? Release candidates enable widespread real-world testing by the community to catch bugs and performance issues before the final stable version is published.
