Linus Torvalds to 'Start Being More Hardnosed' About 'Pointless Pull Requests' – Some of Which Come from AIs

Linus Torvalds, the creator and principal maintainer of the Linux kernel, has signaled a significant shift in his approach to managing pull requests during the kernel development cycle. In his weekly state of the kernel update announcing the fifth release candidate (rc5) for version 7.1, Torvalds expressed frustration with what he described as "pointless pull requests" – many of which are being generated or influenced by AI code review tools.

The Linux kernel development process follows a structured cycle where Torvalds opens a two-week "merge window" for new features, followed by several release candidates (typically rc1 through rc7) that should progressively move toward stability. During this later phase, the focus is traditionally on identifying and fixing regressions – issues where new code breaks existing functionality.

However, in his recent update, Torvalds noted that rc5 for Linux 7.1 is "pretty big. Quite a bit bigger than rc5's have traditionally been," and expressed dissatisfaction with "totally trivial stuff to random drivers" being submitted at this late stage.

"These things are 'fixes', sure, but at the same time a lot of them are simply so irrelevant that I think they'd be better off in a linux-next tree and get merged during the merge window," Torvalds wrote. "And yes, several of these series were triggered by AI code review."

The explicit mention of AI's role in generating these pull requests represents a notable development in the ongoing conversation about artificial intelligence's impact on software development. Torvalds has previously expressed concerns about AI-generated bug reports, noting last week that "the continued flood of AI reports has basically made the security list almost entirely unmanageable, with enormous duplication due to different people finding the same things with the same tools."

The implications of Torvalds' new stance extend beyond mere personal preference. By declaring rc5 "too big" and stating that large release candidates "are not conducive to long-term stability," he is highlighting a fundamental principle of software quality assurance: late-stage changes, regardless of their apparent triviality, introduce risk.

"Trivial fixes may be trivial, and have a pretty low chance of causing problems, but 'low chance' is still not 'zero chance,'" Torvalds explained. This caution reflects the reality that even minor code changes can have unforeseen consequences, particularly in a complex codebase like the Linux kernel with millions of lines of code and countless interdependencies.

For developers contributing to the Linux kernel, Torvalds' message is clear: exercise greater discernment when submitting pull requests during the release candidate phase. He specifically asked contributors to "start looking closer at your pull requests, and ask yourself: 'Is this really a regression or serious enough that it shouldn't just go into the development pile?'"

This advice applies particularly to those using AI tools for code review. While AI can certainly help identify potential issues, Torvalds' comments suggest that uncritical acceptance of AI-generated suggestions – especially for non-critical fixes during late development cycles – is counterproductive.

The Linux kernel's development process is widely regarded as one of the most successful examples of open-source collaboration, with thousands of developers contributing code from around the world. Torvalds' role as the final arbiter of what code makes it into the mainline kernel has always been crucial to maintaining quality, but his recent statements indicate that the scale and nature of contributions – particularly those influenced by AI – are changing the dynamics of the process.

For organizations and developers contributing to the Linux kernel, this development underscores the importance of understanding the development cycle and respecting the established processes. Submitting non-essential changes during release candidate phases, regardless of whether they were flagged by AI or human review, adds unnecessary complexity and potential risk to the release process.

As the Linux kernel continues to evolve and AI tools become more prevalent in software development, Torvalds' approach may represent an important case study in how to maintain quality control in an increasingly automated development landscape. His willingness to push back against what he perceives as problematic contributions, even when they come well-intentioned developers using modern tools, demonstrates the ongoing importance of human judgment in software development.

For those interested in contributing to the Linux kernel, the official documentation on kernel development practices provides detailed guidance on how to effectively participate in the process while respecting the development cycle.

The Linux kernel's development model, including its approach to handling contributions and maintaining stability, continues to serve as an important reference for large-scale collaborative software projects worldwide. Torvalds' recent statements highlight the challenges and considerations that arise as new technologies like AI intersect with established development practices.