In the long rhythm of open-source development, even the most seasoned stewards encounter weeks that resist easy explanation. Linus Torvalds, the architect of the Linux kernel, found himself presiding over an unusually heavy second release candidate for Linux 7.3 — more commits, more fixes, more noise than the quiet that typically follows a merge window. No single cause emerged, and so, with characteristic wit, he offered the era's favorite scapegoat: artificial intelligence. The moment is small in itself, but it speaks to something enduring — the gap between the stories we tell about complexit
Torvalds jokingly blames AI for oversized Linux 7.3 release candidate
We'll obviously all blame it on AI, because it's an easy thing to blame
Mark
So Torvalds is saying the second release candidate is unusually large. What actually caused that?
Mimi
He couldn't find a single cause. Filesystems, graphics, networking, drivers—they all sent in fixes at the same time. It was just a busy period.
Luke
But he says it "didn't feel" particularly busy. How does he know it wasn't actually busier than normal, just in ways he didn't notice?
Mimi
Fair point. He's working from his own sense of the cycle, not from a systematic comparison of rc2 sizes across years.
Mark
And then he blames AI. Is that serious?
Mimi
No, it's a joke. He's poking fun at how the industry blames AI for everything now.
Luke
But the real story is that a week earlier, he'd made an actual mistake—the system upgrade during the merge window. That caused real problems.
Mimi
Right. That one he owned completely. Called himself a nincompoop.
Mark
So which is the actual problem—the system upgrade or the busy rc2?
Luke
They're separate issues. The upgrade caused rc1 to be complicated. The rc2 size seems to be just normal work piling up.
Mimi
And Torvalds is comfortable saying "I don't know why rc2 is big," which is honest. He could have invented a reason.
Mark
Does this affect when Linux 7.3 actually ships?
Luke
The source doesn't say. We only know rc2 is large and that there's more work to do before the final release.
Mimi
The size of a release candidate doesn't necessarily predict delays. It just means more fixes to review and test.
O Pulso
- Linux 7.3's second release candidate arrived bloated with commits from filesystem teams, graphics, networking, BPF, and driver trees — all at once, all without a clear common cause.
- Torvalds admitted the size caught him off guard, unsettling the expected calm that follows a merge window and raising quiet concern about what the full release cycle might hold.
- With no technical culprit to name, he reached for humor, nominating AI as the obvious scapegoat — a wink at an industry increasingly prone to blaming large language models for anything unexplained.
- Just a week prior, Torvalds had owned a different kind of disruption: a self-inflicted system upgrade mid-merge-window that he called the act of 'a grade A nincompoop,' creating debugging headaches he had to chase down alone.
- The two incidents together paint a candid portrait of kernel stewardship — one failure born of a human decision, another born of distributed complexity, both resolved, neither catastrophic.
In the long rhythm of open-source development, even the most seasoned stewards encounter weeks that resist easy explanation. Linus Torvalds, the architect of the Linux kernel, found himself presiding over an unusually heavy second release candidate for Linux 7.3 — more commits, more fixes, more noise than the quiet that typically follows a merge window. No single cause emerged, and so, with characteristic wit, he offered the era's favorite scapegoat: artificial intelligence. The moment is small in itself, but it speaks to something enduring — the gap between the stories we tell about complexity and the humbler truth that systems, like people, are sometimes simply busy.
Linus Torvalds announced this week that Linux 7.3's second release candidate came in heavier than expected — a "full fat" release, in his words, at a stage in the development cycle that usually runs quiet. The merge window had closed, and the typical lull had not arrived.
The commits piled in from multiple directions: filesystem teams, the graphics subsystem, networking and BPF contributors, driver trees. Torvalds looked for a single explanation and found none. A late EDAC pull had been overlooked during the merge window, but he dismissed it as minor. The kernel had simply needed attention in many places at once. "Nothing looks particularly odd," he concluded — and then, with nowhere else to point, he pointed at AI. "Whether that's really the cause or not, it's an easy thing to blame," he wrote, gently mocking the industry's reflex to invoke large language models whenever something resists explanation.
The week before had offered a sharper kind of accountability. When announcing the first release candidate, Torvalds had blamed himself directly for complications during the merge window. He had upgraded his system in the middle of the process — a decision he knew was unwise even as he made it. "I naively went 'nothing will go wrong,'" he wrote, before describing the debugging work that followed. He called himself a "grade A nincompoop" and moved on.
Together, the two announcements sketch an honest picture of how large software projects actually unfold: sometimes a specific human decision creates cascading problems, and sometimes complexity simply accumulates from the ordinary work of many teams. Torvalds' humor about AI is a small thing, but it points at a real tendency — the impulse to find a technological explanation when the simpler truth is that systems get busy, and people sometimes make choices they immediately regret.
Linus Torvalds, the Linux kernel's chief maintainer, announced this week that the second release candidate for Linux 7.3 arrived heavier than expected—packed with more commits and fixes than the typical quiet period that usually follows the merge window. In his regular Sunday update, he described it as a "full fat" release, the kind that doesn't normally show up at this stage of the development cycle.
The size caught him off guard. "This didn't feel like a particularly busy rc2, but it clearly was," he wrote. The second release candidate is traditionally when developers pause after the intense merge window and bugs take time to surface. Not this time. Multiple filesystem teams submitted fixes. The graphics subsystem sent in a substantial pull request with scattered corrections across the board. Networking and BPF contributors added their own patches. Driver trees arrived with updates too. No single culprit emerged from the pile.
Torvalds examined the contents for anything unusual and found nothing that stood out as the obvious cause. He noted a late EDAC pull that had been forgotten during the merge window, but dismissed it as minor. The real story was simply that many different parts of the kernel needed attention simultaneously. "Nothing looks particularly odd," he concluded, before pivoting to humor. Since no actual explanation presented itself, he suggested the obvious scapegoat: artificial intelligence. "It might be just random, but we'll obviously all blame it on AI, because whether that's really the cause or not, it's an easy thing to blame," he wrote, winking at the tech industry's current tendency to attribute nearly everything to LLMs.
This wasn't Torvalds' first moment of self-reflection about the 7.3 cycle. A week earlier, when announcing the first release candidate, he had taken direct responsibility for complications that arose during the merge window. He had performed a system upgrade in the middle of the process—a decision he immediately regretted. "I made it unnecessarily hard for myself by being a grade A nincompoop and doing a system upgrade in the middle, and then having to chase down an issue that caused," he wrote. The upgrade created problems he then had to debug, extending the difficulty of an already demanding period. He acknowledged the mistake plainly: "I knew I shouldn't have done it, but I just naively went 'nothing will go wrong.' And hey, it was all good in the end, I just made this merge window harder for myself than I should have."
The contrast between the two announcements reveals something about how large software projects actually work. The first release candidate's bloat came from a specific, identifiable human error—a decision made under pressure that created cascading complications. The second release candidate's size, by contrast, emerged from no single mistake but from the ordinary work of many teams fixing real problems across the kernel. Torvalds' joke about blaming AI for the second candidate's size reflects a broader pattern: when something goes wrong in tech, the impulse to find a technological explanation often overshadows the simpler truth that sometimes systems are just busy, and sometimes people make decisions they immediately regret.
Citações Notáveis
This didn't feel like a particularly busy rc2, but it clearly was.— Linus Torvalds, in his Sunday kernel update
I made it unnecessarily hard for myself by being a grade A nincompoop and doing a system upgrade in the middle.— Linus Torvalds, on his rc1 complications
Quer a matéria completa?
Leia o original em The Register ↗
Technology