But, he said, it took six months to set up the tool to not produce bad responses. 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.

I mean just because I don’t copy other people’s neovim or emacs dotfiles 1:1 doesn’t mean they’re not very helpful when putting together my own config. After all one of the great things about open source is that when you share some config or code others can benefit from it without having to go through the same struggle.

Torvalds suggested adding an AGENTS.md file that makes generating kernel code expensive. He ends every session with a lengthy cleanup pass that takes a couple of hours to run, executing tasks like ensuring that the code matches the comments. That amount of work cannot be done with a free LLM account. I asked whether that risked cutting off bright new developers that we would rather hang onto; Torvalds answered that the kernel project only wants serious developers. That includes having access to enough resources to function in this world.

I thought the point of said agents file was to increase efficiency, not deliberately make things expensive and punish people who don’t get infinite free tokens.

Linux is of course free to run dev however they want and will survive no matter what given who backs them, but it is very depressing to see someone who I’d looked up to as I got into tech take such an attitude. Making things harder for new contributors while worrying about succession also seems ridiculous.

score 1 · canonical

I find it funny[1] that the Rust for Linux people had to fight tooth and nail for years to get Rust approved for a small portion of the kernel yet LLMs got their filthy tentacles everywhere this quickly.

Not that I wasn’t expecting this outcome, when slopshops like Microsoft and Nvidia are throwing copious amounts of money and “assisted-by” patches at you there’s not much else you can do (as it happened to Rust itself, funnily[1] enough).

[1] funny in a sad way.

score 1 · canonical

Sounds like a big middle finger to all new contributors, be their contributions be made with AI or not. Unfortunate.

score 1 · canonical

The requirement to add Assisted-by tags to patches created with LLMs will be removed, but maintainers still want to know about how a patch was created.

This just seems like a bad idea? At least before there was a requirement to be pointed at, now it’s ??vague asks??

score 3 · canonical