Linus Torvalds Strikes Again: AI Tools Disrupt Linux Kernel Development, Inflating 7.1-RC5 Beyond Normal Bounds
Linus Torvalds Strikes Again: AI Tools Disrupt Linux Kernel Development, Inflating 7.1-RC5 Beyond Normal Bounds
- 60% of MD5 Password Hashes Can Be Cracked in Under an Hour with a Single GPU
- Dirty Frag: Root Access on Every Major Linux Distribution — No Patch, No Warning
- Ubuntu 26.04 LTS (Resolute Raccoon): The Most Ambitious Ubuntu LTS in a Decade
- Proton Mail: Data Transferred to FBI Again!
- How Close Are Quantum Computers to Breaking RSA-2048?
- How to Prevent Ransomware Infection Risks?
- What is the best alternative to Microsoft Office?
Linus Torvalds Strikes Again: AI Tools Disrupt Linux Kernel Development, Inflating 7.1-RC5 Beyond Normal Bounds
For the second week running, Linux creator Linus Torvalds has issued sharp criticism of AI-driven contributions — this time blaming them for an abnormally bloated fifth release candidate that undermines the kernel’s stabilization rhythm.
Linus Torvalds released the fifth release candidate for Linux Kernel 7.1 on May 24, 2026 — and alongside the code came a frank and frustrated announcement email. For the second consecutive week, Torvalds found himself railing against the growing influence of AI tools on kernel development, warning that the late-cycle noise they generate is actively harming the project’s long-term stability.
To the surprise of absolutely nobody by now, rc5 is pretty big. Quite a bit bigger than rc5’s have traditionally been. I’m not entirely happy about it.
— Linus Torvalds, Linux 7.1-rc5 AnnouncementA Two-Week Pattern of Frustration
This is not the first time Torvalds has voiced concern about AI’s role in kernel contributions. The frustration has been building for at least two weeks, each episode revealing a slightly different facet of the same core problem.
Why RC5 Size Matters
In the standard Linux kernel development cycle, a new kernel passes through seven release candidates — and sometimes an eighth if delays occur — before the final stable release. By the time RC5 arrives, the expectation is that the pace of change should be slowing down significantly. The focus narrows to genuine regressions: bugs introduced by changes in the current cycle, not historical issues that have been quietly lurking for years.
Torvalds described most of the RC5 changes as “totally trivial stuff to random drivers.” While individually small and low-risk, their sheer volume creates what he called “pointless churn” — unnecessary noise that increases review and merge costs for maintainers, and introduces a collective, non-trivial probability of regression even if each patch is benign on its own. He noted that a patch which has even a small chance of causing a regression is not acceptable when the goal is to freeze and stabilize the codebase for final testing.
A typical Linux kernel release cycle consists of 7 release candidates (occasionally 8). By RC5, the community expects code churn to be declining sharply. Most changes at this stage should address regressions introduced in the current cycle — not non-critical fixes to long-standing issues that would be better deferred to the next merge window.
AI Agents in the Code — for Better and Worse
The Linux 7.1-rc5 release cycle has seen AI coding agents, including GitHub Copilot and Claude Code, actively contributing fixes across the kernel’s codebase — touching graphics drivers, security patches, and C code throughout. Torvalds acknowledged that AI tools can be genuinely useful, but stressed they are only beneficial when they don’t create “unnecessary pain” and “pointless make-believe work.”
The problem, he argued, is structural: many contributors are independently running the same automated tools over the same codebase, generating overlapping reports and fragmented patches. The result is a late-stage surge of low-value contributions that individually seem reasonable but collectively overwhelm maintainers and bloat the release candidate beyond its intended scope.
Torvalds Issues a Directive
To push the development pace back toward normalcy, Torvalds has required developers to more carefully evaluate their pull requests before submitting them during the RC phase. The core test: Is this fix addressing a regression introduced in the current cycle? If not, should it simply wait for the next development window?
He stated plainly that he will “start being a bit more hardnosed” about rejecting noncritical pull requests going forward. Developers can expect tighter scrutiny during the remaining release candidates as the kernel team works to bring Linux 7.1 in for a stable landing, currently projected for mid-June 2026.
What Actually Merged This Week
Despite his frustrations, the fixes merged into Linux 7.1-rc5 are wide-ranging and cover many real areas of the kernel:
The release also includes Intel P-State and AMD P-State driver fixes, additional HP and ASUS laptop support via x86 platform drivers, and a broad sweep of sound subsystem improvements. Torvalds acknowledged that none of this is inherently “scary” — but insisted the timing and volume are the problem, not the fixes themselves.
What’s Next
With Linux 7.1 stable expected around mid-June, the kernel community is entering its final weeks of the development cycle. If Torvalds’ pushback succeeds in reining in the late-stage churn, the next few release candidates should grow smaller and more focused. If not, an eighth release candidate — a signal of a troubled cycle — becomes more plausible.
For now, Torvalds’ message to AI-assisted developers is clear: use the tools, but know when to hold back. Finding a bug with an AI agent is only half the job. Deciding whether now is the right moment to fix it is the other half — and that judgment, he implies, is still very much a human responsibility.
