I think asbestos is a good analogy, but I think your claims that the inherent downsides of LLMs outweigh any possible upsides are as-yet unfounded. I also think it’s kind of strange that you assume anyone who thinks otherwise is being disingenuous.
Maybe they are too dangerous for broad use, and we need to regulate them like asbestos, or uranium. Maybe they shouldn’t be used outside of a laboratory setting or by anyone who doesn’t have extensive training with how to interact with them safely. It seems pretty clear that Torvalds has decided they’re worth the downside, and while I don’t know if that’s a good call, I don’t think he’s operating in bad faith to the detriment of the kernel project. That doesn’t sound like something he would do.
I feel confident that I don’t have the expertise to say for certain one way or the other, though the experience I do have with software tools makes me think there’s probably an application for them where the downsides can be mitigated to the point where they become worthwhile. I think there’s probably some single-purpose or tailored application of LLMs related to textual analysis that don’t require the theft of the whole internet, and don’t require insane amounts of energy to run. I don’t think we have discovered them yet (at least I haven’t), but saying “this kind of software is only bad and can have no ethical uses, ever” seems premature.
There are, broadly speaking, two categories of problems with LLM generated code (well, everything, but code here specifically). One is purely technical, does it bring anything to the process, does it broadly speaking saves you time or effort. And the second is everything else: environmental impact, the bubble of it all, the consolidation of power to all the worst people, erosion of skills, death of education, inflation of prices of consumer electronics, psychological impact, and so so so much more, all lumped together as non-technical downsides.
A bunch of llm-proponents, and here we sadly have to include even Linus, are only engaged with the first category of problems, throwing the whole second category aside as “irrelevant”. And a lot of those people are smart enough so I don’t believe they don’t see this.
If there was only the problem of the tool being kinda shitty but useful for some people, I wouldn’t be so against it. I still don’t believe they actually benefit from it, but if they think they do, I’m not going to argue with them about it.
But it’s fundamentally not true. You can’t ignore the second category in your evaluation. To quote Linus himself,
But the solution is not to put your head in the sand and sing “La La
La, I can’t hear you” at the top of your voice like some people seem
to do
about the fact that even one of the downsides from the second category makes the whole “but I, a person with 40 years of experience, can use llm in such a way that it sometimes produces some benefit” not worth it. No benefit of it is worth the fact that the new generation of junior developers don’t know how to write code on a fundamental level, while elon mask owns all of humanity’s personal data and can do whatever he wants with it. Even if those benefits were kinda big, it wouldn’t worth it, but they aren’t.
I think you’re missing the trees for the forest. Yes, all those downsides make the current crop of LLM implementations pretty shitty for humanity, on balance, but none of them are intrinsic to LLMs, they are intrinsic to the current business practices that are attempting to profit from LLMs. LLMs are just math. Often computationally expensive math, sure, though they needn’t be. But I don’t think it helps anyone to say “LLMs are intrinsically bad and no one should use them” any more than it makes sense to say that blockchains are intrinsically bad and no one should use them (a common sentiment about a decade ago). I think a more realistic approach is to find out what real benefit can be had by using LLMs, and see if that benefit can be realized without all the downsides you mentioned.
To take the case at hand, I think open source code review is probably a pretty good candidate domain for exploration, because there’s a large corpus of available text that could be obtained ethically, if LLM trainers bothered to put in the work. It seems like Torvalds isn’t concerned with that, and it looks like sashiko only supports LLMs that are known to have been trained on stolen data, which is shitty and I think Torvalds and the Linux maintainers should be taken to task for that. I think LLM makers should get permission from project maintainers before training on their codebases, just as natural-language rightsholders should be asked for permission before using their works for training. I think LLMs trained on ethically sourced data should be permissively licensed and, if they collect revenue, should be expected to kick some of that revenue upstream to the projects they benefit from. I also think such projects could be powered ethically: I have some plans for my own small solar powered server that will run batch jobs when production is higher than my storage capacity. Right now I’m mostly targeting Folding at Home, but I might well run my own LLM trainer if I get enough capacity, and I could see a similar distributed processing network for high-cost jobs from trusted open source LLM projects.
I don’t think you’re wrong about the scope of the downsides, but I also don’t think it makes sense to take them all as a singular block and judge all possible LLM tools by all possible downsides. I think they are problems that can be engineered around with technical and social guidelines for use. And I think it will be down to the open source community to set those guidelines, because I don’t know of any other group that has the expertise and the motivation to do so.
I think asbestos is a good analogy, but I think your claims that the inherent downsides of LLMs outweigh any possible upsides are as-yet unfounded. I also think it’s kind of strange that you assume anyone who thinks otherwise is being disingenuous.
Maybe they are too dangerous for broad use, and we need to regulate them like asbestos, or uranium. Maybe they shouldn’t be used outside of a laboratory setting or by anyone who doesn’t have extensive training with how to interact with them safely. It seems pretty clear that Torvalds has decided they’re worth the downside, and while I don’t know if that’s a good call, I don’t think he’s operating in bad faith to the detriment of the kernel project. That doesn’t sound like something he would do.
I feel confident that I don’t have the expertise to say for certain one way or the other, though the experience I do have with software tools makes me think there’s probably an application for them where the downsides can be mitigated to the point where they become worthwhile. I think there’s probably some single-purpose or tailored application of LLMs related to textual analysis that don’t require the theft of the whole internet, and don’t require insane amounts of energy to run. I don’t think we have discovered them yet (at least I haven’t), but saying “this kind of software is only bad and can have no ethical uses, ever” seems premature.
There are, broadly speaking, two categories of problems with LLM generated code (well, everything, but code here specifically). One is purely technical, does it bring anything to the process, does it broadly speaking saves you time or effort. And the second is everything else: environmental impact, the bubble of it all, the consolidation of power to all the worst people, erosion of skills, death of education, inflation of prices of consumer electronics, psychological impact, and so so so much more, all lumped together as non-technical downsides.
A bunch of llm-proponents, and here we sadly have to include even Linus, are only engaged with the first category of problems, throwing the whole second category aside as “irrelevant”. And a lot of those people are smart enough so I don’t believe they don’t see this.
If there was only the problem of the tool being kinda shitty but useful for some people, I wouldn’t be so against it. I still don’t believe they actually benefit from it, but if they think they do, I’m not going to argue with them about it.
But it’s fundamentally not true. You can’t ignore the second category in your evaluation. To quote Linus himself,
about the fact that even one of the downsides from the second category makes the whole “but I, a person with 40 years of experience, can use llm in such a way that it sometimes produces some benefit” not worth it. No benefit of it is worth the fact that the new generation of junior developers don’t know how to write code on a fundamental level, while elon mask owns all of humanity’s personal data and can do whatever he wants with it. Even if those benefits were kinda big, it wouldn’t worth it, but they aren’t.
I think you’re missing the trees for the forest. Yes, all those downsides make the current crop of LLM implementations pretty shitty for humanity, on balance, but none of them are intrinsic to LLMs, they are intrinsic to the current business practices that are attempting to profit from LLMs. LLMs are just math. Often computationally expensive math, sure, though they needn’t be. But I don’t think it helps anyone to say “LLMs are intrinsically bad and no one should use them” any more than it makes sense to say that blockchains are intrinsically bad and no one should use them (a common sentiment about a decade ago). I think a more realistic approach is to find out what real benefit can be had by using LLMs, and see if that benefit can be realized without all the downsides you mentioned.
To take the case at hand, I think open source code review is probably a pretty good candidate domain for exploration, because there’s a large corpus of available text that could be obtained ethically, if LLM trainers bothered to put in the work. It seems like Torvalds isn’t concerned with that, and it looks like sashiko only supports LLMs that are known to have been trained on stolen data, which is shitty and I think Torvalds and the Linux maintainers should be taken to task for that. I think LLM makers should get permission from project maintainers before training on their codebases, just as natural-language rightsholders should be asked for permission before using their works for training. I think LLMs trained on ethically sourced data should be permissively licensed and, if they collect revenue, should be expected to kick some of that revenue upstream to the projects they benefit from. I also think such projects could be powered ethically: I have some plans for my own small solar powered server that will run batch jobs when production is higher than my storage capacity. Right now I’m mostly targeting Folding at Home, but I might well run my own LLM trainer if I get enough capacity, and I could see a similar distributed processing network for high-cost jobs from trusted open source LLM projects.
I don’t think you’re wrong about the scope of the downsides, but I also don’t think it makes sense to take them all as a singular block and judge all possible LLM tools by all possible downsides. I think they are problems that can be engineered around with technical and social guidelines for use. And I think it will be down to the open source community to set those guidelines, because I don’t know of any other group that has the expertise and the motivation to do so.