Same. I really don't care about any of this stuff.<p>I am not a luddite by any means, I constantly keep trying out LLMs in order to see if I am missing anything. I am <i>trying</i> to get some utility from them, but I just can't.<p>"But you can use them to generate code", no, not really. First of all, why would I want to generate code? Code is a liability, I want less code, not more. Also, code is very expressive, I can say exactly what I want in code much more effectively than I can try to explain in English to an LLM. The LLM always misunderstands and generates garbage, garbage that takes more time for me to read, understand and fix, compared to simply writing it in the first place.<p>"Ah, but you can generate the boring stuff, boilerplate, stuff like that". I don't write any boilerplate, any repetitive things I automate by pricipled things, like more abstract code or by using my very effective text editor skills. Trivial code is easy to get right, by definition, why would I risk getting it wrong by using an LLM?<p>I do get some utility out of LLMs by asking questions about stuff I don't know about. The LLM's answers are almost always wrong, but they can push me in the right direction by informing me of things <i>I am not aware of</i>. This is not really a feature of LLMs, it's just that Google has become garbage at searching. So yes, LLMs are useful, but only by accident.<p>Generative "art"... miss me with all that.