Linus Torvalds has used AI to author code for the Linux kernel, marking the first time the world's most prominent open-source developer has employed artificial intelligence to generate production-grade kernel patches. The announcement, reported by ZDNET on October 14, 2026, details how Torvalds worked with Anthropic's Claude Code to tackle a longstanding memory cache bottleneck that slows multi-core processors. The patches represent a shift from skepticism to practical adoption for a developer who once viewed AI-assisted coding as experimental at best.

The performance improvements from the AI-generated patches proved substantial in testing, though limited to a specific workload. When eight threads executed 16-byte pread() operations against a single page on Torvalds' four-core, eight-thread Intel Core i7-8565U laptop, throughput jumped from roughly 9.9–10.4 million reads per second to 30–33 million reads per second. Kernel cycles consumed per read dropped from approximately 2,950 to 725 in the same multi-threaded scenario. Single-thread reads showed smaller gains, declining from about 452–454 cycles to 408–409 cycles per read. The problem being addressed—contention in Linux's page cache when multiple threads access the same cached file—had produced a shared cache line that accounted for 99.92% of measured cross-core load according to Hit in Modified cache line metrics.

In his Git submission, Torvalds wrote that "The code, the test programs, the review behind the list above and the first draft of this text were done with an LLM coding assistant (Claude Code) under my direction." The Linux creator's October 9 remarks following a conversation at Open Source Summit Europe in Prague were notably enthusiastic: "I used to think it was just not very good—but now I'm at the point where I actually really like using AI," adding that "I think AI is a wonderful tool if you use it correctly and treat it as a tool, and I find it makes programming much more enjoyable." However, he cautioned that "you need to be very careful using AI when you're doing something real and important." The report notes Torvalds' stance evolved from December 2023, when he saw AI mainly as a bug-finding aid, through late 2025 when he became interested in maintenance and patch review, to January 2026 when he experimented with AI coding on a personal Python project.

The technical approach trades immediate memory return for reduced contention on shared reference counters. Rather than returning memory removed from the page cache to the allocator right away, the patches delay reuse until a read-copy-update grace period passes. This allows a reader protected by an RCU read-side section to copy data straight into a user buffer without grabbing a folio reference, avoiding repeated updates to a shared reference counter that forces processor cores to compete. The solution addresses an issue known since 2024 that previous fix attempts haven't resolved. Yet significant engineering challenges remain unresolved—Torvalds described the memory reclamation issue as "the part of this series that is least solved" in his submission. The new code path also introduces a regression when a read targets a user page not yet mapped, incurring an extra failed exception before reverting to the existing implementation.

As of October 13, the branch remained outside the mainline Linux 7.3-rc6 release, and whether it will be merged—and on what timeline—remains uncertain. The report emphasizes this doesn't make Linux as a whole run three times faster; the results describe experimental outcomes from a particular small-read workload on Torvalds' personal laptop during Linux Plumbers in Prague. Machines with more CPUs or multiple sockets should see greater gains, though nothing larger has been tested. The experiment puts both sides of Torvalds' AI position on display: Claude Code helped produce kernel code, tests, and review materials with promising benchmark results, but also unresolved memory-reclamation questions and performance regressions that must be addressed before the experiment becomes production code. The turning point is clear—the world's foremost kernel maintainer is now generating new Linux kernel code with AI assistance, something that would have seemed impossible even a year earlier. For enterprises evaluating whether AI coding tools belong in critical infrastructure workflows, watching how the kernel community handles this merge request may prove more instructive than any vendor benchmark.