Linus Torvalds Defends AI in Linux

In July 2026, Linux founder Linus Torvalds issued a strong statement on the Linux kernel mailing list, demanding that developers opposed to AI-assisted code review either accept AI tools or fork the project. This stance contrasts sharply with his criticism of AI as 'hype' two years ago.

In his email, Linus emphasized that AI-assisted review can greatly improve kernel code review efficiency and reduce human review bottlenecks. He himself has started using AI tools for patch review and believes this is the inevitable direction of kernel development evolution.

This shift is not an isolated event. In recent years, the Linux kernel community has gradually introduced automated testing and static analysis tools, with AI seen as a continuation of this trend. Linus' tough stance indicates he no longer views AI as optional but as necessary for the community's survival and growth.

Linus' 'fork or leave' remarks directly target AI opponents within the community. He believes that opposition from a minority should not hinder the entire community's technological progress. In his email, he wrote: 'If you don't like AI, then fork your own kernel and stop getting in the way here.'

This logic is based on Linus' consistent philosophy of open-source community governance: technical decisions should be led by actual contributors and maintainers, not delayed by opponents through veto power. He cited historical cases in Linux kernel development where similar controversies (e.g., C vs. Rust) were resolved through practical verification.

However, the introduction of AI is more controversial than language choices. Opponents point out that AI review may introduce hard-to-trace errors, and model training data may contain biases or vulnerabilities. Linus acknowledges these risks but believes they can be controlled through continuous improvement and human review, while outright rejecting AI would only leave kernel development behind the times.

Linus' statement sparked heated discussion in the Linux kernel community. Some senior developers publicly support it, believing AI tools can indeed reduce review burdens; others worry that Linus' coercive attitude may damage the community's collaborative culture.

Notably, opponents do not entirely reject AI but question the maturity and transparency of current AI tools. Some developers proposed setting up an 'AI review sandbox' for voluntary testing rather than mandatory adoption. But Linus believes gradual adoption leads to fragmentation and prefers a full rollout at once.

Currently, no core maintainer has announced a fork, but multiple mailing list threads show some developers are evaluating whether to switch to other kernel branches (e.g., Zircon or FreeBSD). This event could become a turning point in open-source community governance: when a founder enforces a technical direction, how is community cohesion maintained?

Linus' stance shift is not an isolated case. From GitHub Copilot to Google's AI code review tools, AI is permeating all aspects of open-source development. As the world's largest open-source project, the Linux kernel's attitude serves as a bellwether.

However, AI introduction also brings new problems: how to avoid 'black-box' code review? Does model training data include GPL code? These questions have no clear answers yet. Linus' 'deploy first, improve later' strategy may accelerate technology adoption but could also plant compliance risks.

In the future, the Linux kernel community may need to establish norms for AI tool usage, including model transparency, review result traceability, and boundaries of human review. Linus' ultimatum is just the beginning of this debate.