@mirabilos@toot.mirbsd.org yeah.. If you look at projects with permissive AI policies, you basically can't run anything built and maintained in the modern era anymore. The Linux kernel is open to it, too, as is systemd. Worse yet is that the folk who think the BSD options rely on stuff being ported over by the now tainted Kernel, and OpenBSD and FreeBSD both lack policies against AI contributions and* at least in the case of OpenBSD has already quietly let a bit of slop in (through tmux, if memory serves. Yet another tool tainted).

It's honestly very demoralizing to see all these projects open up to these problematic tools. But I still consider those projects that hard-pivot to serving the slop market to be much much worse.

Ultimately I know that LLM-based tools are here to stay, and all this legal ambiguity around output of it will not ever get solved. The words "too big to fail" come to mind here — or the old Mutually Assured Destruction idea, where no large organization is going to be willing to sue another for using slopware trained on their copyrighted works, as they themselves then open themselves up to the same kind of litigation.

It's just one more example of how the people that run these large organizations don't have to adhere to the same rules we mere peasants do. Par for the course.

And, well, I'd still prefer to stick with the less-slop-riddled options like Linux and all. It'll hopefully remain a viable option for myself and many others for decades to come. Hopefully nobody is going to look at the Kernel and go.. Hey let's slop port this to Rust in a week, and have that somehow end up being the thing all major distros switch to 😅

I sure wish Debian would clarify the project's stance on LLM usage, though. For my own peace of mind.