Forgot your password?
Close
typodupeerror
Linux

AI Makes Linux 7.4 Merge Window the Largest in Kernel History (phoronix.com) 27

Posted by EditorDavid from the large-Linux dept.
"On a Git commit basis, the Linux 7.4 merge window has been the busiest on record," writes Phoronix," with 19,671 commits since Linux 7.3: As of this morning in Linux Git ahead of v7.4-rc1, the kernel source tree is up to 41,605,120 lines. That's from 5,193,578 blank lines, 4,993,604 lines of comments, and 31,417,938 lines of detected code. Or nearly half a million lines of code added during the Linux 7.4 merge window, ignoring comments and blank lines [as detected by cloc ].
A historical graph of commits shows a clear spike for Linux 7.4, which Phoronix argues shows the definite recent uptick due to AI/LLM activity." With Linux kernel maintainers being "completely overwhelmed" due to both AI generated bug reports and patches/fixes, the Git commit activity has been surging as a result... The upward trend has continued building the past several kernel releases. Typically the merge windows saw 13~14k commits or less and now we are pushing 20k....

Linux 6.14: 10,034
Linux 6.15: 13,568
Linux 6.16: 13,805
Linux 6.17: 12,289
Linux 6.18: 12,910
Linux 6.19: 13,229
Linux 7.0: 12,468
Linux 7.1: 13,963
Linux 7.2: 14,395
Linux 7.3: 16,267
Linux 7.4: 19,671
And "we could still see a few more trickle in today for Linux 7.4 ahead of the expected Linux 7.4-rc1 kernel release later in the day."

UPDATE: "People are expecting 7.4 to be the next long-term support release," Linus Torvalds wrote today in his weekly state of the kernel update, saying that's also "another independent reason why some releases tend larger than others. So it may not *just* be the flood of AI patches."

The Register reports: Torvalds also seems not to be completely convinced that all AI submissions are entirely helpful, by describing this release as "Large, and full of fixes for issues found by AI (and more and more the patches have been suggested by AI too, although they often get rewritten)."

AI Makes Linux 7.4 Merge Window the Largest in Kernel History More | Reply

AI Makes Linux 7.4 Merge Window the Largest in Kernel History

Comments Filter:

  • OpenBSD (Score:1)

    by Anonymous Coward
    Here I come...

  • This is affecting everyone (Score:5, Interesting)

    by SoftwareArtist ( 1472499 ) on Sunday November 01, 2026 @03:44PM (#66404895)

    As an open source maintainer, I've seen the same thing. There's been a huge surge in activity recently, driven by a wave of AI generated and AI assisted bug reports and feature implementations. Some of them are useful. A lot of them are slop. The part about overwhelming maintainers is completely true. We're all suffering under the deluge.

    If you're thinking of running an LLM on someone's code base and sending them "fixes" for all the "bugs" it finds, please don't. You really aren't helping them.

    • If you're thinking of running an LLM on someone's code base and sending them "fixes" for all the "bugs" it finds, please don't. You really aren't helping them.

      Or, at the very least, do some work yourself to see what is valid and worthwhile, don't just blindly push things off onto someone else.

    • I'm fine with people sending AI-assisted PRs (especially if they are transparent), but what to do when AIs directly send emails through https://www.agentmail.to/ [agentmail.to] without any accountable human involved?

    • Re:This is affecting everyone (Score:5, Interesting)

      by MIPSPro ( 10156657 ) on Sunday November 01, 2026 @04:53PM (#66404983)
      I've been a C programmer for 30+ years. I've written code on over 30 platforms and 20+ compilers. I don't need their LLM-created patches because I'm already using better LLMs than they are and 50 miles ahead of them because I've been there since day one, just waiting for the day the LLM could code better than I could (happened in early 2026).

      You cannot help me with your LLM code because I've already helped myself. It's like some guy wondering in off the street to give the mechanic his advice on an engine rebuild. Just fuck off. We don't need you. Think of it like this. If europeans showed up with metal weapons and were going to give a few to the savages, who would be the most likely to make good use of those weapons first? Would it be the old women or little kids or would it be the tribes warriors, already use to fighting with weapons and now just given better gear? My response to these kids is "Bruh, fuck your submission." If you have ideas or usability fixes, great, let's hear them. However, you fucking piss-ants couldn't code before LLMs, you can absolutely shut the fuck up and stuff it, now.

      • Slopflooding is sabotage under guise of assistance.

      • I don't need their LLM-created patches because I'm already using better LLMs

        Have you gone back and applied it to all your old code bases? I mean I'm sure you as a coder probably are way ahead in the LLM game compared to script kiddies, but are you using it everywhere? Do you even have the time to do so? I suspect you probably have better things to do than to convert your life to an endless bugfixer.

        The reality is it doesn't matter how good our tools are if we don't apply them.

        • Of course I have. It's been absolutely trivial to do so. That happened even before they could out-code me because it was easy money. On to some more shit.

    • Re: (Score:2)

      by AmiMoJo ( 196126 )

      A lot of these patches are for security flaws that LLMs found, so can't be simply ignored because of the source. If one LLM found it, another can.

      It's just a difficult period we have to get through. Then it will be down to checking new code for flaws with AI, because it get deployed.

  • Welcome to the new world, I hope you enjoy your stay.

  • You can't seriously check this amount of code (Score:4, Insightful)

    by BitterEpic ( 10503015 ) on Sunday November 01, 2026 @04:40PM (#66404967) Homepage

    For a core component of many, many computers, I do not think Linux kernel development needs arbitrary speed gains in the amount of code contributed. The kernel should be uninteresting and realistically this amount of code is less likely to be closely inspected than previous code.

    Hell is built with good intentions. The answer to everything becomes, "more AI! Let's use AI to validate the AI!" Eventually we will find not only biases, but AI that is designed to inject vulnerabilities only under a specific set of circumstances. Maybe when the contributor is named Linus and he has the power to override everyone else..

    Eventually, everything is wrapped up in a nice alert that everyone ignores until the nuclear power plan melts down.

    • I suspect most of the updated "kernel" code is for device drivers.

    • The summary is pretty bad. Here is a link to a good article from 9to5linux [9to5linux.com] where they actually have Linus explaining what is going on. I'm guessing you don't understand how Linux development works, but nothing about the process is arbitrary. If it made it into the kernel there is a reason and it was looked at by human eyeballs. Nobody, and I mean nobody is saying "more AI! Let's use AI to validate the AI!". You are basically just being a bad conspiracy theorist.
  • You either think you are smarter than the Linux kernel developers or you don't. Feel free to comment about how Linux developers are clearly clueless idiots and this is a sign that they have no idea what they are doing, but just know you sound stupid when you do it.

Slashdot Top Deals

Another megabytes the dust.

Close