George_E_Hale
insufferable blowhard
The things you lean on / are things that don't last
User ID: 107
George_E_Hale
insufferable blowhard
22d ago
Personally--and ai may be alone here--I would prefer a poorly-worded, convoluted, even slightly ungrammatical-yet-readable post to anything made "with the assistance of LLM." I am not anti-LLM and don't think they're by default evil, but that's not why I come to theMotte.
We can all converse with LLMs. There they are, we need only ask them a question (worded properly.) And they can, I agree, give reasonable, even "insightful" answers, if that is the word to use. I myself have gotten good advice from LLMs on issues as diverse as how to word an email in Japanese to what kind of fertilizer to use on the verbena (though ChatGPT abetted my murder of my olive tree.) The blossoming use of LLMs everywhere outside of mathematical or computational applications (in things like cooking, cleaning, shopping, stain removal, gardening, etc.) suggests considerable utility in their use. They are not evil.
My point is simply that when I open the Motte I prefer the warts-and-all version of humanity. (And often get it.) I should say, since I'm spouting off my druthers, that for my part I do not mind if people use LLMs for direct translation, though the non-English-native-speakers here tend to have far better written prose than most native speakers. My reasoning is that LLM direct translation is often very good, especially if calibrated for tone, and in such cases the original writing was, well, written originally by the writer.
More options
Context Copy link