I'm just some dev.
I feel exhausted, stuck between two sides about LLM usage that I feel are both going too extreme - open source people and corporate world. So this is my rant, take it as you wish.
I am not happy LLMs exist. The same way I'm not happy that banks, or the fractional banking system exists, or that CFDs or spread-betting exists, or that coal power plants exist. But they do. They're tools created by humans and for better or worse are here.
But I'm tired of being met with two extreme views - one from certain parts of the FOSS community and the other from corporate higher ups with zero sense.
For FOSS I'm going to reference this document: https://codeberg.org/ethical-foss/open-slopware/src/branch/main/why_not_llms.md
First off some of their points have almost nothing to do with writing code, and I don't even know why they're there. For example: "On one hand, many things that you think are 'AI' are actually humans in another country pretending to be an AI chatbot for you for either extremely low wages or in some cases, no wages e.g. prison labor." None of the tools I use for coding are humans, there's no way a human is doing this. That sentence doesn't even have an "on the other hand" - so I'm not even sure what their point is.
On the ethics side, "Ties to the War Industrial Complex" and "Environmental Impact" are real concerns - as they have been for banks or trading. But this repo calls for a total ban on LLMs. They very freely list anything that uses an LLM, not just "slop," despite the name. They're anti-LLM, full stop.
But some of the articles they reference themselves say differently:
-
https://newsroom.ucla.edu/stories/opinion-ai-is-destroying-our-planet-we-must-act says: "There is a better way to build technology that protects our data and benefits the earth, but it requires reigning in some of the most powerful and profitable companies in the history of the world. It's high time for us to collectively act. Technology doesn't need to destroy our planet and empty our pocketbooks. Rather it can be designed to enhance the protection of both." That's not "using LLMs is evil," that's "regulations must exist and things like data centers must be controlled" - which I agree with.
-
https://arxiv.org/abs/2508.21634 explains: "These findings highlight the distinct defect profiles of AI- and human-authored code and underscore the need for specialized quality assurance practices in AI-assisted programming." Again, the conclusion isn't "don't use LLMs," it's "LLMs need to be treated with care and more attention."
They have other points about FOSS maintainers being overwhelmed with low-effort PRs, which I fully agree with.
It's become nearly impossible in the modern age to express an opinion of "I partially agree and partially disagree" with something, even when that something is genuinely complex. That's the case here - I disagree with their conclusion to avoid LLMs entirely, but I agree that many of the things they raise are real issues.
I have been struggling in vain to convince friends and colleagues to stop buying from Nestle or Amazon - it's nearly impossible. I do it myself, but I don't put people who don't on a mental list of "people who use evil things," because frankly any big corporation you've used - Apple, Microsoft, Visa, Mastercard, hell if you don't know even CocaCola had actually hired hitmen to assassinate pro-worker activists in third world countries. And I'm going to mention Nestle again because on the rank of companies that are evil, Nestle overshadows almost everyone. Forget stealing data, they use child labor, destroy entire towns, steal water from even parts of the US for pennies and resell for dollars. And they lobby against other things.
Their point about "deskilling" is also something I've thought about and been concerned with. But it's always the case with progress. Most of us don't know how to write bytecode by hand anymore - we rely on compilers to do it. We've been using optimizing algorithms and neural networks for decades to do things most of us don't fully understand. There are fields where only a handful of people remain true experts. The rest of us rely on the technology. I'm not saying "don't be concerned about LLMs because of that" - I'm saying deskilling is an inevitable side effect of progress, one that can and should be mitigated by keeping real experts around and building in safeguards. Some people will use LLMs to get a better understanding of code and to learn. Others will write a vague prompt and push a PR without reading it. That second thing is the real "slop" - not that all LLM output is slop by nature. And a good chunk of that can also be caught by automated checks on repos.
Part 2: the other side of the coin is absolutely insane. Corporations are pushing so hard for LLMs that it's mind-boggling. My company had a higher-ups go on stage at the start of the year with some really strong "we must use LLMs" messages in ways that were obvious to most developers were bad. And I was told there's a push from leadership for us to be "10x more productive" in general. I have flat out been vocal in my company that's never going to happen for me, or most other devs, and if that becomes the actual requirement, I might as well start looking for a new job now.
LLMs aren't the productivity boost for developers that CEOs seem to think they are - not unless you throw all caution to the wind, which unfortunately plenty of people are willing to do.
I can't express how much more upset I am at the corporate world than at anyone trying to push back LLMs - because while I completely understand the anti-LLM pushback, I do not for a second agree with the corporate profit-driven agenda that drives higher ups to make us all more productive - all the while paying us the same and working same hours.
That's f*cking absurd and if you ask me to make an absolute choice of "Are you anti-LLM or pro-coporate agenda" I will 100% be anti-LLM. But we don't live in a world of absolutes. We live in a world full of nuances and often the things that make life easier have major downsides. Developers (myself included) see the effects of LLMs directly though, so its much easier. We've been for the most part ignoring (or never had major pushback) against non-developer things - corporations that use child labor or steal water (Nestle) before LLMs existed.
They ARE, however, still incredibly useful tools that can make development easier when used correctly.
LLMs aren't fully ethical - but that's true of most things manufactured or run by large companies. Your best bet for pushing back on that isn't "ban it completely," - because you won't even make a dent. They're here to stay, evne if the future might involve less of them -they're not going away. The solution is activism for pushing for stronger regulations on these companies, since regardless if the open source community rejects LLMs, the corporate world isn't - and neither are random internet non-developers who use LLMs for far less productive things (image gen, video gen, clickbait videos, etc).
I'm going to quote Linus (source):
AI is a tool, just like other tools we use. And it's clearly a useful one.
It may not have been that "clearly" even just a year ago, but it's no longer in question today.
There are other questions around AI (like what the economy of it will actually look like in the end), but "is it useful" is no longer one of those questions. Anybody who doubts that clearly hasn't actually used it.
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.
The solution is to make sure those LLM tools help maintainers instead of just causing them pain. There's no question on that side.
If you are concerned with ethics, I honestly implore you to not to buy anything from Nestle or Amazon at least (if you happen to).