It’s hard to predict what will happen to the Linux structure after Linux Thorvalds AI-comments: will we have a non-AI Linux fork, will people live with it or will people move to BSD?

mosgaard@uddannelse.social

kevin@dice.camp

@mosgaard I admit that I didn't expect this and thought the OS would be my safe haven from the AIpocalypse. Oh well.

kats@chaosfem.tw

@mosgaard I'll switch to BSD as soon as the DAW ecosystem is there.
I'm a SVR4 chick, and it's a short enough hop from there to BSD.

leeloo@c.im

@mosgaard
Peopld (>90%) will live with it,because momentum.

Unless Valve replaced Linux support with FreeBSD support.

carstenboll@mstdn.dk

@mosgaard I don't think the majority of people will care a lot as long as we're not seeing quality issues.

svabonius@mstdn.dk

@mosgaard do have a link? Haven't seen anything worrying lately.

dazo@infosec.exchange

@mosgaard

I'm not a huge fan of LLMs myself. But it's still early in its development phase.

But what I'm really not a huge fan of is this "anti-LLM hysteria". To me that hysteria is like calling for banning of cars because they pollute so much. While at the same time ignoring that the pollution by driving fully electric cars is close to none.

I do acknowledge that most LLMs today are resource hungry. And yes, that gives a good reason to be cautious. Plus, Deepseek is known to provide reasonable results without the same resource requirements. As LLMs are developing further, they will improve - also in the resource consumption. If everyone stops now, LLMs will not mature properly either.

And LLM output is just as good as the input it is given. LLMs are like paintbrushes in this regard. The outcome is highly relates to the person holding the brush. Inexperienced users will produce very bad output. Experienced users will end up with quite nice results.

That someone uses LLM to certain tasks doesn't mean the outcome is slop. It might be slop, if the prompt writer is sloppy. LLMs are tools. Nothing more, nothing less. Garbage in, garbage out. But also the inverse.

That's the pragmatic view I see Torvalds takes. Use LLMs if you know "how to hold it", but use it responsibly.

I see little wrong that. We all have left punch cards behind to do programming and testing. Nothing wrong to embrace new technology and tools when it seems generate good results when used correctly.

folfdk@helvede.net

( I'm not @mosgaard but)

@svabonius

Since @sjvn already commented on this thread I'll link to this:
https://www.zdnet.com/article/linus-torvalds-puts-his-foot-down-tells-anti-ai-programmers-to-fork-it/
which explains the controversy fine I think.