ProdigalFrog comments on Linux kernel goes full AI slop code, hides chatbot commits

[–] ProdigalFrog@slrpnk.net 1 point 9 hours ago* (last edited 2 hours ago) (3 children)

Your response to my comment mentioning that AI is contributioning to climate change was to suggest it was equivalent to HVAC workers being afraid of heatpumps, which is ridiculous.

If you think climate change is an impending global disaster, do you think building giant data centers that require extremely polluting on-site power generators is not a particularly bad idea, and should factor in heavily on how this technology is used or promoted?

It seems like most users and advocates of LLMs find the idea of limiting its use to just locally hosted models to avoid the majority of the negative externalities to be unthinkable.

permalink

fedilink

source

parent

hideshow 3 child comments

[–] flilare@piefed.zip 0 points 8 hours ago (2 children)

Climate change is an issue, but this hardcore stance with no nuance isn't helping. The extreme strawman slippery slopes don't help, either.

You've got a capitalism problem, not an AI one

Data centers don't require any of those things. It absolutely can and should be regulated (or even existing laws enforced.) Irresponsible implementation is an issue, but if the only people with any familiarity with the technology don't care to do any of that? Do you think there is a world where this genie goes back into a bottle?

AI is indisputably helpful in many situations. Most people who I know that use AI generally use local models and that can cover 95% of what people want or need to do with it. (You might have a bit of a blindspot there). But there is a place for the big models as well.

permalink

fedilink

source

parent

hideshow 2 child comments

[–] ProdigalFrog@slrpnk.net 1 point 8 hours ago (1 child)

The capitalism problem is also responsible for an inability to regulate AI, as the very people responsible for the regulations are lobbied to fuck by the AI corps to avoid that regulation.

That means the only power to fight back against them rests in the hands if individuals, who can refuse to use the corporate owned tools.

Most people I know are using corporate AI, and appeals to try local LLMs fall on deaf ears in my experience, with claims that they aren't powerful enough or that the hardware to run them is too cost prohibitive. If most of the people you know are running local models, I think that is an exception to the norm.

Would you support FLOSS projects officially restricting any AI contributions to those only made with locally hosted models? As I said, most times I've mentioned that idea in other threads, the pro-AI side was vehemently against it.

permalink

fedilink

source

parent

hideshow 1 child comment

[–] flilare@piefed.zip 0 points 7 hours ago

Absolutely not. Such a community resource is exactly where we need to be employing it. And for building the tools we can use to cut those corps out, that's not a reasonable solution at all.

Human code review, and not such a rush to throw out new "features" would alleviate those issues.

I don't have any data to support it, but I suspect a lot of that use relates to their work and not personal projects. Though I'll admit the Facebook pics and meme generation junk is almost certainly mostly done on corporate systems. But this could be done with reasonable systems as well.

That capitalism problem is so much bigger than AI, but its the problem, not AI.

We would not be anywhere at all as a species if we rejected any advantage, but now with civilization we can do it responsibly. Even if the early days are the Wild West. I can support you or anyone who wants to avoid it, but I can't support a total rejection.

permalink

fedilink

source

parent