Even Linux is embracing AI now, and drowning in it at the same time

Linux runs most of the internet, nearly every Android phone, and a large share of the world's servers and supercomputers. So when the person who created it takes a position on artificial intelligence, it is not just one developer's opinion; it sets the tone for open-source software broadly. In mid-2026 Linus Torvalds took that position, bluntly, and it captures the double-edged reality of AI in software: genuinely useful, and genuinely overwhelming. Here is what he said, what is happening beneath the headline, and why it matters far beyond the kernel.

Torvalds plants a flag

The clarity came, as it often does with Torvalds, in a mailing-list reply. In a July 2026 discussion about connecting the kernel's patch-tracking system to an AI-assisted code-review tool, he pushed back on developers uneasy about AI's role, writing that AI is "just a tool ... and clearly a useful one," and that Linux "is not one of those anti-AI projects." To those who wanted the project to reject AI outright, his answer was characteristically blunt: "Fork it. Or just walk away."

By August, as The Register reported, he had gone further, describing large, AI-assisted kernel contributions as increasingly the norm rather than the exception. Coming from the person with the final say over the kernel, this is about as close to official policy as open source gets: AI is welcome in Linux, full stop. That matters, because plenty of maintainers across open source have wanted to draw a harder line, and the most influential figure in the field just declined to.

The flip side: the maintainers are drowning

Here is the part the triumphant headline skips. Welcoming AI has a cost, and it is landing on the people least able to absorb it. The kernel's development is famously sustained by a relatively small number of maintainers, some volunteers and many employed to do it, but finite either way, who review every proposed change. And in 2026 they are being flooded with AI-generated submissions: bug reports and patches produced by people pointing AI tools at the kernel and forwarding whatever comes out.

The problem is not that AI code is always bad; it is the volume and the signal-to-noise ratio. Maintainers describe security and bug channels becoming near-unmanageable as automated, often duplicated or low-quality reports pour in faster than humans can read them. Each one still has to be triaged by a person, because a real vulnerability buried in the noise cannot be ignored. AI made it trivial to generate plausible-looking bug reports far faster than review, or the nascent AI review tooling, can keep up. The bottleneck, human attention, got worse, not better.

Why this is the whole of open source in miniature

Zoom out and Linux is living through the tension every open-source project now faces. AI genuinely helps: it can draft patches, spot certain classes of bugs, and lower the barrier to contributing. But open source runs on a scarce resource, the time and judgement of maintainers, and AI floods exactly that chokepoint. It multiplies the easy part (producing code and reports) while leaving the hard part (reviewing, verifying, deciding) as slow and human as ever.

That is why Torvalds' "it's just a tool" is both correct and incomplete. A tool that lets one person generate a hundred patches is only a net good if someone can review a hundred patches, and in these projects, that someone is the same overworked maintainer as before. The projects that thrive in the AI era will be the ones that figure out how to let AI help reviewers, not just contributors, and how to filter the flood before it reaches a human. Linux, characteristically, is running straight at the problem rather than pretending it away.

Why it matters

Linux is the canary here. If the project with the deepest bench of maintainers and the clearest leadership is both embracing AI and straining under it, every smaller open-source project is facing the same squeeze with fewer people. The outcome will shape the software almost everything else is built on: whether AI makes open source faster and more accessible, or quietly burns out the maintainers who hold it together. Torvalds has answered the first question, AI stays. The harder question, how to keep the humans in the loop from drowning, is the one the whole ecosystem now has to solve. For more, see the Software section and our look at the software supply-chain security overhaul.

Linux and AI in 2026, at a glance

Torvalds' stance	Linux "is not one of those anti-AI projects"; AI is "just a tool"
To objectors	"Fork it. Or just walk away."
What's adopted	AI-assisted patch review; large AI-assisted contributions now common
The strain	Maintainers flooded with AI-generated (often duplicate/low-quality) bug reports
The core tension	AI multiplies code produced, but not the human review it still requires
Why it matters	Linux sets the tone for all of open source facing the same squeeze