Hetare King

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.

This is really more a case of possibility #2. Specifically, why are you rolling your own in the first place? The whole reason the output of the AI is “probably flawless”, is because this operation already has a million implementations. What’s the value in conjuring up another one and adding it to the code that needs to be maintained? Yeah, binary trees aren’t terribly complicated, but neither is struct Vector3 { float x, y, z; }; or any of its operations, until you realise you’ll be running those operations an awfully lot, so it would be better if it had the right alignment and SIMD instructions were used in its operations, and of course, if you have Vector3 you’ll probably also need a Matrix4x3 and a Quaternion etc., so in the end it’s just better to use an existing solution.

This is what I mean with turning good devs into mediocre ones: the worst solution becomes the easiest solution, so that’s what people will lean towards.

Of course, there can also be situations like you have a very specific use-case and using something off-the-shelf would be sub-optimal, or you’re working with old code and making their own was the decision the previous person or your dumbass past-self made, so that’s just what you have to work with. But the AI is going to struggle more with those cases too.

Lastly, I also mentioned “energy”. If you’re trying to put in words things that aren’t easily conveyed in English (or any other natural language) all the time, you burn through your cognitive endurance like crazy. This is why you don’t want to be in a highly technical meeting in the morning, so being in one all day with a moron you can’t really come to common understanding with (beyond the size of its context window) is even worse.

Local models are still controlled by a small group of people, in the sense of deciding how and with what to train the thing. And given the cost of doing so, you can’t just easily fork it and make adjustments yourself.

If it takes less time and energy to describe what you want in natural language than to just write the code, there are two possibilities:

1. You’re not a good dev.
2. There’s a better, more reliable way of doing it that is less likely to needlessly increase your maintenance overhead.

In the end, only shitty devs benefit, while being prevented from becoming good devs, while good devs turn into mediocre devs.

It only appears dogmatic to someone under the mistaken impression that technologies are value-neutral. Make no mistake: even if every generative AI model were open-source and you could run on your own hardware, it would still be a technology that empowers the already advantaged far more than anyone else.

It doesn’t matter what the reason for its creation is, you can make similar arguments about laws against assault and murder or the right to vote, what matters is what it’s doing in the current legal situation.

If you think that the end of copyright will suddenly render corporations powerless, you are beyond naive. Some will fail to adapt, undoubtedly, but many will not just adapt, they’ll be able to take full advantage of this new situation. Copyright isn’t the only mechanism they can dig a moat with, and as long as the economic incentives remain what they are, they will find a way and be no less exploitative. Meanwhile, the individual creator or indeed projects like Linux will lose what little protections they have and be completely at the mercy of these monsters. Forget exploitative contracts, they can just exploit you without one!

A better world without copyright is possible, it just doesn’t look exactly like this world, except without copyright. The abolishment of copyright is not an ideal, it’s a means to an end. But for those means to lead to a good end, how and when it happens matters. To cheer on a development because it nominally achieves what you believe is needed to lead to your ideal outcome, even when the way it happens would lead to an outcome that is in direct opposition of the ideals that lead to that belief in the first place, is a sign you’ve stopped thinking.

You do realise that even if copyright were abolished, large, powerful corporations would continue to exist and even manage to use the absence of copyright to their advantage?

Without the GPL being enforced, a company like Microsoft could finally start digging a moat around Linux without constraints.

The thing is, if all other things remain the same, an absence of copyright will also be used by the wealthy and powerful to further enrich and empower themselves and as a tool of oppression.

From the Ferengi Rules of Acquisition:

#34 War is good for business.

#35 Peace is good for business.

You cannot have internally consistent, principled view on copyright without also having a compatible view on the economy. And even then, the economy needs to come first; you cannot start destroying load-bearing pillars, no matter how ugly, without first getting everyone out of the house.

The tool in itself can really be usefull when well used. Many people use it and get nice results.

The way I see it, there are two kinds of people: those to whom AI is utterly useless because it just ends up taking more time and have them burn through their cognitive endurance faster, and those to whom it would be irresponsible to let them use it, because it would prevent them from finding better solutions for the busywork they’re letting the AI do, and they just end up creating a lot of needless maintenance overhead.

But more importantly, being “useful” to certain individuals is not the same as being good for society. Which brings us back to it undermining licenses like the GPL, which would further enrich and empower the already advantaged.

Also, I can’t find any example in history of a technology that was successfully banned despite being usefull.

CFCs? Asbestos? The bans on them aren’t perfectly enforced, but nevertheless, the hole in the ozone layer is shrinking and the number of people dying of asbestos-related lung cancer is reducing, so they do have value.

The strict “no-AI” movement has already lost its battle

Even if it were true, that would be no reason to stop trying or to stop saying the truth: No matter useful you experience it as being, you are causing harm. To yourself, to the field, to society.

It won’t exactly be able to accomplish anything like that if it’s been undermined no matter what the mechanism of its enforcement is.

Analogies are comparisons of structure, not intensity. There are a lot of things that are “not as bad” as child labour, that doesn’t make them OK.

But i don’t believe in copyright laws either. Do you believe in the goals of the GPL? Because you have to admit, the way AI works, regardless of whether you frame it as “theft” or “copyright infringement” or not, is at odds with those goals.

GPL may call itself “copyleft”, but legally it is dependent on copyright to be enforceable. In the first place, the whole point of licenses like GPL is to ensure that when people take from the commons, they also give back to the commons. The way AI gains its ability to generate shitty code undermines this, and so is ideologically very much at odds with Linux and the wider FOSS world.