Robert Kingett

2 links below.

I know there was a big fight over the below, but I doubt just saying people are gonna be selfish and use it, and equating it to a thing like keyboards is the best way.

I do wish they would put up more of a fight but I guess even #Linux folk just give in to this AI nonsense. There are ways to ban it without tiring yourself out. https://github.com/torvalds/linux/blob/master/Documentation/process/coding-assistants.rst

And these folks maintain a list of projects that never use AI in their process, at all.

https://noai.starlightnet.work/

#AI #AIHype

linux/Documentation/process/coding-assistants.rst at master · torvalds/linux

Linux kernel source tree. Contribute to torvalds/linux development by creating an account on GitHub.

GitHub
11:39amWeb
Show threadD. Moonfire7h ago

@WeirdWriter I think the use of AI has already entrenched itself. And it is useful with small problems or hunting down obscure difficulties.

The consequences of using it and the frightening speed that folks accept those long term destructive aspects (water, power, the folks who live near AI data centers, the fact it all costs money that I don't want to spend) is just a mirror (at least for me) of Amazon and Walmart taking over small businesses. In my dad's words, "that's a you problem" which is horrible (I hated that answer of his).

As for banning it, I've seen how much work one of my developers goes to obscure his AI usage from me, even though the company encourages using AI, because he knows that I'm anti AI and I'm picky about code, and that makes it exhausting. Since Linux developers have a higher chance of having that hacker mentality, flat out banning it is just a challenge to hide it better.

Kind of like how there are clearly humans behind the AI only social network. They are only cosplaying as AI because it's a challenge.

I wish I had a better answer.

The fun ones, like checking in a skills.md file that insists that every comment have emoji are cute but someone is going to notice that no PR gets merged with emoji and tell their bot to remove them.

I think the header is trying to find a balance between accountability (much like the DCO) and not wasting more time and energy hiding it.

I also think Jellyfin's AI policy is trying to find that balance and is fairly well written in that regard.

Show threadKawazoe7h ago
@WeirdWriter Unfortunately, it's only possible to ban AI in coding contributions if you assume the submitter is well behaved, omniscient about their dev tools, and have good intentions. Otherwise, it's basically impossible to enforce, especially at this kind of scale.

While they'd have no way to enforce it fairly and the policy would risk being abused to silence valid contributions, I still think they should've blanket-ban it like forgeo did. At least, it would've gave them strong legal protections, compared to what amounts to a suggestion of respecting the GPL in their current policy. Now we risk loosing the whole kernel over this...
Show threadRobert Kingett6h ago
Yeah! I mean I know people will be shitty and submit LLM code anyway, so I guess they did the next best thing is put everything back onto the LLM user, but I just hate this timeline so much! Every mainstream app I use has drastically dipped in coding quality and my half baked theory is that their bosses, at these larger places, force them to use LLM to code and everything gets worse as a result. Like I'm 99% sure the YouTube app is vibe coded now. @kawazoe
Show threadAzuaron6h ago
@kawazoe @WeirdWriter Given that AI code is, essentially, public domain with no copyright protections, saying that "All code must be compatible with GPL-2.0-only" actually is a blanket ban against AI code. Unfortunately, it seems like they don't realize this.