TL;DR
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
Linus Torvalds has said he uses AI coding tools and supports developers using them for Linux kernel work. His position is that the tool itself should not be banned or dictate acceptance: contributors are responsible for understanding their work, and patches must meet the kernel’s standards.
Linus Torvalds says developers can use AI coding tools when working on Linux kernel contributions, but the code they submit must still meet the project’s standards. An XDA Developers report published Oct. 26 describes Torvalds as an AI user himself, while making clear that his support for the tools is not an endorsement of unchecked, AI-generated patches.
Torvalds, the original creator of the Linux kernel, has said he used a large language model to help locate and fix a bug, and to write a commit message describing the change, according to XDA. The report also says he used AI to produce a kernel patch. Those examples show that his position is not simply theoretical: he has incorporated AI into his own coding work.
The distinction in Torvalds’s approach is between using a tool and accepting its output. He has objected to rules that require contributors to declare AI use, arguing that the kernel should not impose restrictions based only on which tools someone uses. At the same time, he expects contributors to understand what they are submitting and to do their best.
XDA reports that Torvalds called a proposal to set rules around AI use “utterly stupid.” In the same discussion, he compared objections to AI coding tools with refusing to use an assembler or a text editor. The comparison expresses his opposition to banning tools outright; it does not remove the responsibility to review and maintain submitted code.
Open source · AI · Kernel development
Linus Torvalds’ healthiest approach to AI
AI tools can help developers work on the Linux kernel. Torvalds’ reported standard is clear: contributors may choose their tools, but they must understand and check every patch they submit.
“AI is a wonderful tool if you use it correctly and treat it as a tool.”Linus Torvalds, as quoted by XDA Developers
01 / What he has said and done
Use the tool. Own the result.
Torvalds’ stance is practical: AI can be part of the developer’s toolkit, while responsibility for submitted code stays with its contributor.
Help with a bug
XDA reports that Torvalds used a large language model to help locate and fix a bug in his own coding work.
Describe the change
He also used AI to help write a commit message explaining the change, according to the report.
Produce a patch
The report says he used AI to produce a kernel patch. Its use alone does not make that patch acceptable.
02 / The balance
Open tools, accountable contributions
The reported middle course avoids both a tool-based ban and unchecked acceptance of generated output.
More work for maintainers
Unreviewed code or inaccurate AI-generated bug reports can add pressure to people who must determine what is useful, correct and ready for review.
Review the work itself
Contributors should understand and check what they submit. Maintainers judge patches under kernel expectations; AI assistance does not bypass review.
03 / Contribution flow
From tool choice to kernel review
A simple chain captures the responsibility that remains at each step.
AI can be part of a developer’s workflow.
The contributor remains responsible for the work.
Code enters the normal contribution process.
Acceptance depends on quality and review.
04 / Public comments
Support for experimentation, with limits
XDA describes Torvalds as positive about AI when used correctly, including its potential to help newcomers enjoy programming.
“utterly stupid”
On a proposal to set rules around AI use, as quoted by XDA“AI is a wonderful tool if you use it correctly and treat it as a tool.”
Linus Torvalds, as quoted by XDA Developers“wonderful way to find joy in programming”
On AI and newcomers, as quoted by XDAWhat remains unknown
Impact is not yet measured here
The report gives no count of AI-assisted patches, acceptance rates, review-time effects or the scale of AI-generated bug reports. It also does not establish a new kernel-wide policy or formal review procedure.
05 / Key questions
What this position means
Does Torvalds support AI use in kernel work?
Yes. XDA reports that he has used AI tools and supports contributors using them, with responsibility for their submissions.
Will AI-generated patches be accepted automatically?
No. Code still has to meet kernel quality expectations and pass review.
Has he used AI to produce kernel code?
According to XDA, he used an LLM to help with a bug, a commit message and a kernel patch.
Must contributors disclose AI use?
XDA says maintainers moved to remove an LLM disclosure requirement. The account does not confirm the full current documentation or any later changes.
AI Use Meets Kernel Review Standards
The Linux kernel is maintained through contributions that have to be reviewed for quality and compatibility. Torvalds’s position suggests that the project’s response to AI-generated work should focus on the patch and its effects, rather than trying to police whether a contributor used an AI tool along the way. That approach leaves room for developers to adopt new tools without treating generated output as automatically acceptable.
The question matters beyond Torvalds’s own coding habits because his views carry weight in the Linux community. A blanket prohibition could be difficult to enforce and could generate disputes over how code was produced. Unrestricted acceptance, by contrast, could put more review pressure on maintainers if contributors submit work they have not checked. XDA’s account presents Torvalds’s position as a middle course: tool choice is open, but quality remains the test.
The report also notes concerns about a high volume of AI-generated bug reports burdening maintainers. That is a different issue from whether AI-assisted code can be submitted, and the supplied account does not provide figures or describe a specific policy response. Still, it points to a practical challenge: AI may help people produce and report more, while kernel maintainers must decide what is accurate, useful and ready for review.
Products Worth Considering
Logitech MX Mechanical Wireless Illuminated Keyboard, Tactile Quiet, Graphite
Logitech MX Keys S Wireless Keyboard, Graphite
Jackery Explorer 2000 v2 Portable Power Station with Two 200W Solar Panels
As an affiliate, we earn on qualifying purchases.
Torvalds’s Public AI Comments
Open-source projects have taken different approaches to the growing use of large language models in programming. As XDA describes it, some projects prohibit AI-generated code or commit messages, while others permit AI assistance under quality-control rules. The report does not name those projects or give a survey of their policies.
Torvalds has made several comments about AI use. XDA says he has spoken positively about using AI for projects and defended “vibe coding” as a way for newcomers to enjoy programming. The report also quotes him describing AI as a useful tool when used correctly. His support for experimentation does not mean he regards casual, unreviewed coding as suitable for kernel submissions.
A separate point in the report is that Linux maintainers moved to remove a requirement to disclose LLM use in kernel submissions. Torvalds received concerns about what that change would communicate to contributors and replied against setting a tool-specific rule. The supplied material does not give the date of that policy change or the full wording of the relevant kernel documentation.
“utterly stupid”
— Linus Torvalds, as quoted by XDA Developers
How Kernel Reviews Will Adapt
The report does not specify how many AI-assisted patches have been submitted, how maintainers assess their quality, or whether acceptance rates differ from other contributions. It also gives no details on the scale of AI-generated bug reports or their effect on review times. Those limits mean the practical impact on kernel development cannot be measured from the available information.
It is also unclear whether the kernel’s documentation or review practices will change further. Torvalds’s comments describe a preference against rules based on tool use, but they do not by themselves establish that every maintainer or contributor shares that view. No new kernel-wide policy, timetable or formal review procedure is detailed in the report.
Patch Quality Remains the Test
For contributors, the standard described by Torvalds is straightforward: they may choose whether to use AI, but they remain responsible for understanding and checking the work they submit. Kernel maintainers will continue to review contributions under the project’s existing expectations; the report does not announce a new AI-specific review process.
The next development to watch is whether Linux maintainers clarify documentation on AI disclosure or change how they handle submissions and bug reports. No such change is confirmed in the account. For now, Torvalds’s position is a statement about tool use and responsibility, not a guarantee that AI-generated code will be accepted.
Key Questions
Does Linus Torvalds support AI use in Linux kernel work?
Yes. XDA reports that Torvalds has used AI coding tools and supports contributors using them, while expecting them to understand and check their submissions.
Has Torvalds used AI to write kernel code?
According to XDA, he used an LLM to help find and fix a bug, write a commit message, and produce a kernel patch.
Does this mean AI-generated Linux patches will be accepted automatically?
No. Torvalds’s reported position is that tool choice should not decide whether work is allowed. The submitted code still has to meet kernel quality standards and pass review.
Is AI use required to be disclosed in kernel submissions?
XDA says maintainers moved to remove a requirement to declare LLM use. The supplied report does not provide the current full documentation or confirm whether any later policy changes have occurred.
Source: XDA Developers
