How about going SETI@home with LLM training? Everyone contributing their own computing power to train better, more transparent local models. Using properly licensed, publically available datasets. Not sure how viable this approach would be, but that would make forking of open-source models possible, maybe?

Olmo exposes all of its training code and training data, including post-training etc.

Olmo:
https://huggingface.co/allenai/Olmo-3-1125-32B

Also Apertus:
https://www.apertus-ai.org/

I've trained a tiny GPT model on my computer before. It's shitty, certainly, very shitty actually, but essentially anyone can train an LLM, if you got either enough time or calculation power.

These things are certainly made addictive by design, like TikTok, YouTube, and modern social media.

And even if the Internet itself is useful, these addictive-by-design stuff is so harmful that it is likely that the net effect is negative.

His line seems to be ‘don’t submit shit code.’

And he's been 100% consistent with that since the start.

I like the car analogy. I never had a car, and the last time I called to use one, I had had eye surgery and needed to get home.

There are many aspects of the analogy that match quite well:

• some types of motor vehicles like ambulances and trucks are clearly useful but most uses are not
• cars have a massive negative environmental impact (220 grams of CO2 per every kilometer, not to speak of toxic air pollution, photochemical smog, and particulate matter that appears to cause dementia)
• especially in cities, car feel much faster than they effectively are
• cars destroy cities and walkable communities, to such a degree that European cities like Paris come to new life after restricting cars
• cars are extremely convenient, such that they become an addiction quickly, and after a while you don't even believe you can't walk two miles on your feet. You even lose your sense of walking distances.
• cars massively increase the daily distance you travel, because of Marchetti's constant (humans have a time budget of one hour travel a day), which changes your whole daily routines, leads to more distant work places, urban sprawl in housing, and after a while you are really physically dependent on them
• using cars is bad for your own health because it leads to lack of exercise, which leads to multiple common and deadly diseases
• cars do actually cost a lot of money which is sucked up by a corrupt industry and even more corrupt fossil fuel producers

Accept something along the lines of a positive account given by an experienced software engineer - it’s useful, it presents risks but also it presents new ways to get certain work done well in much less time etc.

I agree to the utility of finding security bugs but the general utility of using LLMs is often overblown. There are studies that say that 95% of companies do not see any measurable positive impact.

The Linux kernel is likely a special case. But for other open source developer communities, it seems to have some negative impacts, too. And the full effects on maintainability of code will only be visible after some years.

It doesn't work reliably, though. It had a race condition he missed due to LLM-induced brainrot.

And you can also have at home your own LLM that you can supply your own power.

I feel like everyone here would like the same overall: not killing the planet, not build more data-centres in the middle of communities, and last but not least having a tool to assist and not to generate unreviewed slop.

We do need to develop now a way to protect the human curated content, that is the part that we want and care about.

Neocities is an interesting subject in this matter, most people there reject LLMs and have their own hand crafted page and content:

https://neocities.org/

Ah okay, thank you. I’m not educated enough on the internet Linuses, so it was a genuine question.

A sob story about the poor widdle smol bean petit bourgeois, making a proprietary closed source game to rent seek off of.

Why should I care about this class enemy?

You class traitors never care about actual working class people.

The comparison between AI and compilers is kind of stupid though. When you program in a given language, you're writing out the entire logic and behavior of your program. Not to mention that the compiler is a deterministic machine, it will always behave in a certain manner and produce predictable results. On the other hand, using AI, you just write a sentence. Then it "creates" a whole bunch of code that will not always be the same, even if you always use the same model and prompt. Not to mention all of the detrimental effects AI has, cognitively speaking, when you start to rely on it. On top of this, all AI models were trained on unfathomable amounts of stolen information.

bot

Scratch a copyright advocate, and a dehumanizer bleeds.

You could probably trace how it comes up with the probabilities of the next token. Though the meaning of all the weights are probably going to be lost on you. Once you've got your probabilities and you've limited to some top number, it's just a dice throw at that point.

Yeah, but your compiler wasn't made with massive copyright infringement, stealing much needed water and rampant corruption, was it?

Compilers are not always deterministic, there is a lot of "smart" code in compilers to automatically optimize code and that can mean that whole parts or functions are replaced with what the compiler thinks is best.

I think there's another question you need to keep in mind. "Am I finding it harder to convince myself to do the work when it's important enough?"

Even with the small amount of LLM dabbling that I do on low importance tasks, the hump I have to get over to actually write code myself is getting higher. I'm considering a blanket exclusion of the tech because of it.

I'm done. This is a pointless exercise. You're refusing to engage on the core point and keep dancing around on semantics. None of this makes an LLM like a compiler.

I already own an EV. And there definitely are people who are very vocal about the issues with ICE engine cars. Personally, I'd rather we go back to smaller communities within biking distance, but that's just me.

Dude, this isn't a story about a closed source game dingus. Have you ever even tried to make a game? Most games can be decompiled or reverse engineered pretty easily. And as long as we live in a capitalist society, you have to play the game, even if you don't want too.

Dunno why you come in here with comments like that....oh, a .ml account that explains it. Can't even engage with people in good faith. Do you seriously go around calling people class traitors everywhere you go? Bcz that ain't going to make you any friends. I am about as far left as it gets, so I find it funny that you are calling me a class traitor.

I am basically an anarchist in my political views, or a socialist, so, yeah, super funny to me to be calling me a class traitor.

What are you even trying to say? I am not advocating for copyright, maybe ask questions first instead of calling people names. Man, I am so tired of this space always being filled with people shitting on eachother for not being the correct type of leftist, or not far enough left.

I am basically an anarchist in my political views, or a socialist

You're not being shit on for being the wrong kind of leftist, you're being shit on for thinking vaguely vibing with progressive politics makes you a leftist. You are a capitalist defending copyright from a capitalist perspective.

Yes its hard to work with unpredictable tool but not every AI is completely unpredictable, it can be only bad prompt or stupid model or something similar but that is problem for people who use AI (or use AI incorrectly). If you want use AI you must force AI to do what you want and this is part of doing things with AI. If you don't want this you shouldnt use AI.

I think its not "AI is stupid shit" or "AI is awesome, it can do things faster". Its something between. For some case is AI good like for brainstorming but for others not like psychology but this is very questionable...

Oh boy, you are so wrong, you don't even know it. Participating in a system that forces you to participate doesn't make me not a leftist. If you want to start a revolution to start that communist or socialist utopia we all dream of, be my guest. I'll be right there with you supporting you when it's all said and done. But right now, that's not in the cards for me. And I'm not about to explain that to someone over the internet.

Devil's advocating for the sense of completeness here rather than bickering, but wouldn't two of the same compiler on two of the same machine with the same code still create the same output? That is to say, it is deterministic rather than stochastic but it just takes unpredictable variables into account when compiling?

I think the difference would be that a compiler will always give you something understandable and explainable even if it's opaque, but an LLM is literally using RNG to generate its output, so could come out with something completely left field. If it switches to Chinese that's one thing because you can just ignore that, but equally a hallucinated bug that looks right but is actually very dangerous could slip in instead.

It's literally copying everyone's work

Since you're down voted without any explanations... No this is a different guy. This guy wrote linux

Nobody moved the goal posts, you simply picked a terrible example but you're too arrogant to admit it.

Someone doesn't choose where they're born, even if that land was stolen. It's simply not their fault. But they do choose whether to create or use an LLM based on theft. So it's not remotely the same.

We also didn't choose whether our ancestors were slave owners, whether they committed various atrocities, etc. Same with other things our country did in the past. We're not responsible for any of that. But we can hold some measure of responsibility for things that happened once we were alive if it's based on how we voted.

We also don't choose anything about our body at birth, attractions, genetics, etc. Yet people judge and condemn people for those as well. People who do judge others for those things are garbage, IMO. Yet that's the kind of gross argument you're basically making here.

Judge what someone chooses, for their actions, not for what they didn't have any say in. And, again, people choose whether or not to use an LLM.

Downloading this comment is not authorized.

Then about the “steal” argument:

You wouldn't download an argument, would you? Or would you? What, you're reading this in your browser without authorization from me?? FBI, open up!!!

With an undercurrent of being polite but firm toward mistakes from newbies, while occasionally telling oldheads 'You should fucking know better!'

I'm also not a fan of IDE that lock you in too much... but I get your point. It's been pointed out to me that I have a bit of an allergic reaction to corporate.. shenanigans.

RNG is pseudorandom. Every computer program is deterministic in some sense, even if the seed is obfuscated or dependent on ephermal environment variables. Being able to predict the output is very different from determinism.

are you trolling?

I checked your history and you seems to genuinely interact but you're sentence makes no real argument (more like a joke).

But maybe you haven't read the rest of the conversation so here some points:

• You can and are encourage to download and use free software. Same with comments.
• You are not allowed to monetized free software (hence the GPL v2 vs v3 debate). To be honest also with my comments.
• The term stealing could indeed be confused with the modern usage of fascists who want to fight against piracy but it was used for a good century by communist to talk about capitalists stealing surplus value made by workers.

Im not saying ai economy wont crash or anything like that, im just saying it's a new tool that most people will use incorrectly just like they use old tools incorrectly too. We can still recognize correct uses.

cognitively speaking

That's not what the damn study says.

People who wrote an essay with an LLM did not show signs of having having written it themselves, on account of, they hadn't. Just like I couldn't tell you the remainder of a long division I put in a calculator, because it went straight to a mantissa. I didn't lose the ability to to math. (At worst, I don't wanna, now that I have a calculator.)

The only long-term aspect of this study was a dozen people coming back months later and switching sides. So even pretending it's asking the right questions, the tabloid interpretation should expect to see the brain-only essayists fumble it and the LLM essayists sharper than ever. But that's plainly not what happened.

They didn't even measure essay quality. It was just about EEG activity, incidental recall, and self-reported degree of ownership. But that nerd shit doesn't get clicks, so every fucking headline is like Does chatbot is brain damage?! or Permanent cyber-psychosis ten minutes!! or If men learn this, it will implant forgetfulness in their souls. They will cease to exercise memory, calling things to remembrance no longer from within themselves, but by means of external marks. It is no true wisdom that you offer your disciples, but only the semblance of wisdom, for by telling them of many things without teaching them you will make them seem to know much while for the most part they know nothing. And as men filled not with wisdom but with the conceit of wisdom they will be a burden to their fellows.

Oh sorry, that last one was Plato arguing against the detrimental cognitive effect of books.

They are very different, actually. A compiler has a well-defined task: it should take any valid code written in a particular language and generate machine code or intermediate code that does exactly what was instructed in the input code, in accordance to the rules of the language. The exact sequence of opcodes it generates from a given input may not be completely predictable, because there are multiple ways of achieve the same result and different compilers will make different choices in different circumstances, but there should be a clear relationship between the behaviour of the output and the instructions in the input, and if there isn't, that's a bug, pure and simple. There's just not that much room for hijinks, which is why even a proprietary compiler like Visual C++ is more strict about following the standard these days; being the outlier risks irrelevance. But that's only possible because there's a clear, shared idea of what a compiler should do.

But with an LLM, the input is any kind of text, and the output is something derived from a large, but finite, fuzzy possibility space, with no understandable relationship between the input and its output. This makes it difficult to reason about failure states. Let's say there's a bias in the model that makes it consistently favour using one library over another. Because most people will just take the path of least resistance, that will result in that library becoming more popular at the expense of the other, whose funding may dry up as a result, so this is an issue that affects more than just the users. But what is causing this bias? Is it there intentionally or just a result of the library already being more popular? Is the bias even there to begin with or am I just seeing it because of my own biases? It's all very hard to say. Just proving that the bias is there and what's causing it is going to be a lot of effort, as is removing it even with the cooperation of the maintainers. And if they're not cooperative and I'm on my own without the funding available to them, then forget about it.

Of course, biases don't just exist in AI models, like how QtCreator is clearly biased towards Qt. But it's right there on the tin, you knew what you were getting into from the start, and if you don't like it, there are any number of alternatives. But because of their high cost, there are only so many LLMs of a certain level of capability and they're all basically immune from accountability, because everything exists in this fuzzy space of reasonable doubt.

You constructed an entirely fictional argument about a study that you quoted. That's not what I was referring to. It's a pointless discussion though. None of these AI companies are financially sustainable (except for the shovel sellers, obviously).

But that's a static algorithm. It will always "optimize" the code the same way. Hence, it IS deterministic by definition. Maybe you should look up what the word means.

It's perfectly ok to use a new tool to do your work faster, more efficiently, and with greater precision if these improvements are justified by the cost of the tool. But if that tool is a coal-fired, water hoarding, ophan crushing, billionaire-owned kleptobot, your results need to be better than pure gold to be worth the true cost of using it.

You're right. Nothing has been stolen and then treated as public property in our society. I am fool, and you are a god.

Anyways

I don't care about Linus, Stallman's opinion is the interesting one.

Just don't ask him about age of consent...

there are abandoned tech certs for building datacenters that produce 110% of their own electricity.

Do you have any sources for this I could read?

Linus become like one of the boomers excited by some dumb ai slop on Facebook

But worse he doesn't care about how unethical is the present ai

I lost respect for this guy

Yes it was a joke, sorry. The propaganda video about "downloading a car" was an incredibly stupid but massive campaign and in hindsight incredibly successful in influenced attitudes.

I don't think any discussion is useful because I am fundamentally and radically opposed to turning intellectual ideas or works into "property" and it's abhorrent and obscene to me. There are other and better ways to fund or reward artist's work. Naturally I hate patents even more. They also delay market adoption for climate change solutions by 20 years. A crime against humanity! And no, that one is not a joke.

The worst thing that fuckai could lead to is help legitimize some kind of "IP tax" for use of AI models, which would as a side effect make it illegal piracy to use any open weight models. Then big tech gets to monopolize AI in the west. They can afford the tax. But we'll forever be limited and dependent in making use of LLM models.

No one driving a car today was alive when the gasoline powered car was invented, and none of the people who were were able to anticipate the terrible consequences of mass adoption of the technology.

The people driving cars today are doing so because they were born into a society that had already completely reorganized itself around cars, and that all but requires the ownership of a car to participate in that society. Meanwhile EVs have been prohibitively expensive to own for the majority of people (although that's finally changing) and so most people who wanted an alternative did not have one available.

Contrast that situation with AI where the terrible consequences are immediately obvious, cheap alternatives already exist and have for centuries, and the "benefits" of using AI are only marginal at best and only in some niche circumstances.

ethically trained models

Are there "ethical" models for people who believe the mechanization of creativity is an attack on human dignity?

Feel free to name whichever study you think says what y'all keep repeating, because I haven't seen that shit. Every time I go searching for some dire evidence of dain bramage from spicy autocomplete, it comes back to one of three poorly-designed studies that fail to distinguish this new tool from any other tool and categorically do not match their pearl-clutching news coverage.

It’s a pointless discussion though.

This is bad faith. Pivoting to unrelated criticism is pretense. You can't make some damning claim, glibly dismiss an effort to explain you're mistaken, and then gallop toward some other topic entirely. That's working backwards from a conclusion and failing to update your beliefs.

No shit these companies are boned. Local models will survive and proliferate on their demonstrable merits. Decentralized development will continue, if nowhere else, among the unceasingly horny randos doing image and video stuff.

We will in fact have to live with programs that can program, but thankfully, using one won't give you a lobotomy.

This exactly what I was thinking. This doesn't even mention the ramification of AI outside the digital space such the environmental costs, the replacement of labor and the complete disregard of copyright or even preservation. I understand people can use many things as a "tool" but when the tools cause some much untold harms as a side effect, is it truly just a tool? It's something used at the expense of another. I know there is no ethical consumption under capitalism but AI is one of those things you could personally avoid if you wanted to. One of the things you technically shouldn't be forced to use. The fact that it is pushed everywhere is so you devalue everything else along with it. For me, it started as being anti-AI in terms of art and academics alone but it has evolved to being anti-AI in general.

Which is a shame considering the Linux distro has never ever once ever ever ever contained a bug