Well, what exactly did the AI do?

If Linus said, "fix the kernel" and the LLM generated 1,000 lines of code that he dropped in, saw it was working, and called it a day, then that would be a problem.

If Linux said, "Where is the problem?" and LLM told him and then he fixed it himself, that's not a problem.

I don´t have a problem with LLM even generating code, I have a problem with programmers not understanding their craft and publishing code they don't fully understand.

The problem is that practically you need to take a hard line because how are you supposed whether something is unexamined slop or if it's been fully understood by a human? The only way to be sure a human understood it is to be sure a human wrote it, so blanket generative AI bans make sense.

Headline is a neuron activation lightning rod. Torvalds used the model to debug, the code is a one word fix he probably wrote himself.

This is the actual commit:

https://github.com/torvalds/linux/commit/818bebeb63dd6bf5f4e07e145f6cdbace520a34c

It fixes a goof where Intel Xe drivers allocate a few kb of memory that is used by the system during boot that causes the video to stop working. The fix is round_down() the memory allocation instead of round_up() the memory allocation. That's pretty much the entire code change. round_up() -> round_down(). The commit message makes a lot of sense, and this bug is something you'd be spending a lot of time on manually debugging due to how insidiously subtle it is. Torvalds used the AI model to help debug the problem, but I'm pretty sure a one word change in code is not an AI generated code... and if changing one word is what the AI really did in the end, where do you draw the line at copyright? Challenging the copyright of individual words is impossible to enforce, because the word is used everywhere.

It's more like he used the LLM to do the grunt work of finding the issue, but he ultimately had to fix it because the thing is stupid and can't think. If anything it shows that these things are weak performing the skilled labor that real humans do on the daily. Still not ideal, mind you because it still is a problem that people might rely on these things to read and understand code. It's good that Linus had to fix the problem personally rather than have the unthinking LLM do it for him. Reporting on this correctly is important, as it lends credibility to the anti-LLM cause.

Jesus this seems like a psyop to fuck with us.

Reading these comments I have to contend that the halo effect is indeed powerful.

IMO it's immoral to use LLM's for any and every purpose and I think every slopper should be called out; no matter how accomplished they are.

Since Linus spends most of his time reviewing changes to the code in the kernel, he's probably one of the few people I'd trust to have properly vetted any AI code he used.

That being said, since you have no idea what code was scraped to generate the fix, it seems to me that it might not be able to be licensed under the GPL, though I'm no lawyer.

I guess Linux is turning into AI slop as well. God it feels like everything we hold dear is under attack these days. Is BSD any better or are they sycophantic AI bros too?

We already knew linus was pro-ai

Remember he said "trust me bro" /s /ns