I suppose that's true, and there is a lot of anti-LLM voices that seem low-effort and/or reactionary. I however find it hard to actively be against them, because I can see how LLMs and generative AI more broadly are actively hurting communities I care about (FOSS projects being drowned in bogus PRs and issues, as well as huge waves of slop projects that are quickly abandoned; designers and artists being replaced by generated graphics [some people at my workplace have started using genAI for public facing graphics even though we have a graphic designer on staff because it's "faster/more convenient"]).

Disregarding all personal, and even the environmental factors (as I do agree that LLMs can be run in an environmentally conscientious way, even though most newly built operations very much don't seem to be, especially regarding power generation), I am still worried about devs and projects relying on LLMs to heavily, because the current level of availability and access to the models most devs seem to use is very much proped up on a very precarious financial construction, and it is very hard to judge from outside whether LLM-assisted projects, even those by experienced coders, are built in a way that would allow for manual maintenance and expansion.

So, while some anti-LLM voices can seem reactionary, I feel that there are many good reasons to be sceptical of LLM usage in a project, and I don't think that those that do don't have well-founded points, they might just be tired of reiterating them (though I feel that is probably the point where you could just write a blog post to point at).

I also see your point about bit letting perfect be the enemy of good, and I agree with that. I very much struggle with perfectionism, both in my own projects and when viewing those of others, and agree that if this leads to some funding for good protects, that is probably more important than their ai policy.