But all AI output is slop. It’s social spam. That’s what it’s for. It’s slop! Say “slop” every time
This writing is far, far too emotional to be considered anything but an opinion piece
But all AI output is slop. It’s social spam. That’s what it’s for. It’s slop! Say “slop” every time
This writing is far, far too emotional to be considered anything but an opinion piece
That is not what slop means. It's not ai slop if the developers understand the code and know exactly what they doing, audit everything.
Incorrect. "Slop" is anything generated by AI.
Please don't spread misinformation.
Incorrect. “Slop” is anything generated by AI.
Please don’t spread misinformation.
There is no definition that is correct for everyone. Therefore this is not misinformation, just because you have a different definition. Slop means if its in example vibe coded. But the Linux developers don't do slop, even when using Ai.
Please stop labeling a different opinion as misinformation.
Oh, this asshole again. They write like a twitter thread, and with about as much depth.
Linus is saying that using an LLM doesn't absolve you of taking responsibility for what you publish.
I am extremely anti-LLM for basically everything but even I can see this moral panic for what it is. Sadly not the first time David Gerard (the author of this blog) has fallen into it.
Ctrl-F "slop" - 32 matches
You know how we call AI output "slop" because it's low-effort and kinda generic? Yeah.
Actually, I just discovered, the person who wrote this bullshit article is the first to submit it in the crosspost.
And he's banning people from the community who ask questions about it.. Or point out it's bad info lol, so, it seems obvious its more about attention than proper journalism
Why are you banning people who dispute your claims. Everyone should be downvoting the cross posted article with a passion. He's running a dictatorship (basically seems like another "madthumbs")
For anyone who wants to see how comprehensive the oversight is on Kernel development: https://www.kernel.org/doc/html/v4.17/process/submitting-patches.html
But yeah, patches don't go straight to Linus or straight to master, and, it certainly doesn't get approved automatically by an LLM or anything like that. If you're lucky after the initial reviews, they almost certainly go to other testing trees first (and generally multiple). Then you also have staging branches, etc.
Oh it's one of the opinionated clueless douchebags from that instance, explains a lot.
For anyone curious how comprehensive the process is, the real info is in https://www.kernel.org/doc/html/v4.17/process/submitting-patches.html
Lmao at this completely misguided comment.
Torvalds’s first publicly released chatbot code was in January this year, when he vibe-coded a project in Python because he didn’t know enough Python. The programming language kids learn in high school. The guy who runs the kernel couldn’t get his head around Python.
Linus would run circles around whoever wrote this article. Ever hear of "domain knowledge"? It's perfectly reasonable to expect an old school engineer well-versed in low level code like C to see high level scripting languages like python as fairly alien.
… If you are a competent developer, he said, an LLM makes a great tool, but only after that effort has been put in to set it up. He resisted suggestions to post his LLM-training materials, saying that each developer needs to go through that process separately.
This is the vibe-code-two-step. "AI" is useful because it helps inexperienced users to code. But you also have to be an expert to use it properly.
AI is useful when you can put it to a task and validate the result.
For many tasks that’s visually seeing a button you can click.
For other tasks you need to be an expert to get value from it and really understand.
If you provide tools and context, you get a lot better results. For most of my core uses I’ve built decent tooling, but I wouldn’t want to share or be responsible for maintaining that for everyone else.
I genuinely don’t think it’s useful for learning unless used like stack overflow, but it’s not designed or trained to do that, it’s trained to take the drivers seat regardless of instructions.
You do realize that all code for the kernel is code reviewed and tested by so many people right? The patches are probably some of the most human reviewed code on the planet....
This feels like the time when every armchair expert became experts on immunology...
top 50 comments