Linux drops AI rule Torvalds says banning it is stupid
The world of open source development just got a little less bureaucratic. The Linux kernel community has dropped a requirement, removing the need for contributors to explicitly declare when they used AI tools in creating kernel code.
This change, implemented earlier this year, signals a shift in philosophy regarding how modern tools integrate into the development process. Previously, the project had suggested that acknowledging the use of advanced coding tools—by adding an Assisted-by tag—was important for tracking the evolving role of AI in kernel development.
However, maintainers ultimately decided that this attribution was unnecessary. The change was prompted by concerns raised by some contributors regarding the administrative burden and the perceived lack of value in such a disclosure.
The reasoning behind the decision came from Linus Torvalds, the kernel’s creator and lead maintainer. In response to the call for the tag’s removal, Torvalds clarified that adding such a tag does not provide valuable information to the maintainers regarding the quality or contribution of the code itself.
Torvalds expressed a broader skepticism about imposing rules on developers regarding the use of tools. He questioned the notion that restricting the use of AI in coding is a logical path forward, suggesting that developers should be expected to understand what they are doing without explicit tool confessions.
“I don’t understand this whole idiotic concept of don’t use AI for X,” Torvalds stated. He argued that the focus should remain on the competence of the contributor rather than the specific tools they employ.
In Torvalds’ view, the goal of kernel maintenance is to distinguish between good code and less-than-good code. He posited that if a developer cannot demonstrate a deep understanding of the underlying machine, they should be limited in what they can submit. He compared the idea of restricting AI usage to hypothetical rules, suggesting that if a developer cannot write code in assembly, they should be prevented from submitting kernel work, emphasizing that true competence should be demonstrated through understanding, not tool restriction.
This move reflects a larger, ongoing conversation within the open-source community: how do we navigate the rapid integration of generative AI into highly technical fields? While the initial push for transparency was about tracking the evolution of the development process, the final decision prioritizes the core mission of the project—maintaining the integrity and quality of the code.
The change underscores a fundamental tension: balancing the desire for transparency with the practical reality of effective, high-quality software development in the age of artificial intelligence.
Buy on Amazon
Credit
