@alatiera@mastodon.social > I think that the point of these statements is shaping *social norms* and not micro-managing developer workflows. They should be about signalling what kind of behavior we want, and what kind we reject. I think this is such a salient point that I haven't seen explicitly stated before (but implicitly debated).  There's a million dollar (at least) marketing machine behind LLMs and plenty of anecdotes about the "times they've saved me a lot of time"* that tend to bog down the discussion and it inevitably gets closed as "too heated".  Not to mention the bad faith arguments, claims without citations ("I don't think there's any environmental impact from LLMs at all"), etc. But it really is about longer term behavior than just a single PR or two, isn't it?  Especially since it seems like a lot of the problems of LLMs don't manifest immediately (either in psychological impact or project code impact). * it is deeply ironic and frustrating that, when one points out that the existing statistics on long term use of the "statistical inference machine" do not support any claims of saved time or effort, none of the users of said large statistical inference machine think about it for even a millisecond and immediately dismiss it, if they even bother to respond.