Urgent.News

What's breaking now, across thousands of outlets.

AI

Linus Torvalds posts his first AI-assisted patch to the Linux kernel, synthetic test clocks 23x performance leap

Linus Torvalds posts his first LLM-assisted patch to the Linux kernel — after issuing clear pro-AI directives, the bossman makes use of AI tooling himself

Linus Torvalds posts his first AI-assisted patch to the Linux kernel, synthetic test clocks 23x performance leap

Linus Torvalds, the creator of the Linux kernel, has posted his first patch assisted by an artificial intelligence system. The development branch targets a longstanding problem with excessive cache contention when numerous CPUs access the same file — a situation that may arise with shared libraries, databases, and web servers. Various attempts to resolve the issue were made in the past, yet none provided a comprehensive solution.

In early 2024, Torvalds previously proposed a fix for the problem and implemented it on his personal laptop. Recently, he ventured to use an AI-assisted approach to address the issue. Torvalds has now shared the solution within a development branch, asserting a remarkable 23x performance enhancement in a specialized synthetic benchmark. Nevertheless, he clarifies that the real-world implications will likely be considerably less pronounced, as no practical workload would experience such gains.

Given Torvalds' long-standing favorable attitude toward AI tools, particularly within the regulated and thorough environment of the Linux kernel, this development is significant. In July, he expressed his willingness to firmly oppose AI detractors, highlighting that an automated code review tool utilized by the kernel team surpasses human capabilities and has identified fundamental programming errors. The team has also chosen to cease requiring patch authors to disclose AI involvement when generating code.

Torvalds has consistently maintained that users should perceive AI as merely another instrument, akin to compilers or editors. He argues that people ought to strive to utilize these tools effectively and comprehend their proper application. While some may argue that only a few individuals in the world are capable of proficiently utilizing AI, Torvalds and kernel maintainers possess a unique advantage.

However, the emergence of highly proficient AI agents capable of identifying prevalent security flaw patterns has resulted in an overwhelming number of reports concerning Linux security, leading to the necessity of eliminating older hardware drivers due to their marginal installation base and frequent vulnerability exposure.

Written by urgent.news from Tom's Hardware's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at tomshardware.com →

More in AI

The link is the diagram: why I built FlowDiagram, a free diagram tool your AI can draw in

By Tuna, the maker of FlowDiagram Every engineering team I have worked with has the same graveyard: a Confluence page with an architecture diagram that was last true two re-orgs ago.

  • Tuna created FlowDiagram to address outdated architecture diagrams on Confluence pages.
  • FlowDiagram offers a unique share link feature, generating a URL with the entire diagram.
  • The tool supports AI-generated diagrams without requiring an API key or account.

More from Friday 16 October →