Comment on The sole purpose of language models is to lower the market value of human skills.
Artisian@lemmy.world 4 days agoI do those too! That’s where the ideas for new architectures, datasets, and training tweaks come from! Math is fun, and it’s fascinating that math can talk sometimes.
Lemminary@lemmy.world 4 days ago
Yeah, I also talk to ChatGPT sometimes in extremely short conversations of no more than six messages, fully knowing that it’s a flawed machine and that it’s for my amusement. It’s incredible that it can do that! Just like with wireless communication, which I thought was impossible growing up. Like, holy shit, we’re living in the future, why not enjoy it a little? I don’t even think a conversation spans more than six messages tops. It’s amusing, but not that amusing if you can see right through it.
I don’t know what OP is on about but they seem to be on a crusade. They’re citing articles about getting advice from it and having it think for you, totally missing the point of having fun. Like, if you’re against AI, at least give better arguments that address why, instead of throwing things at the wall and seeing what sticks.>
lazynooblet@lazysoci.al 3 days ago
Looking at traffic analytics pretty much all our developer staff use chatgpt for 3+ hours a day. I’m not a big fan of using llm for my own development work. I’m proficient in they languages I write in so I don’t need it as much.
I feel like using llm can get you a quick fix but for programming a lot of the results are nonsense. It’s really really well formatted but nonsense still the same. Maybe I can’t use it right. Or I’m asking the wrong questions.
I find it hilarious how when you call it out for being BS it responds with “yes of course, you are right!..” then gives a possibly working or nonsense answer, who knows.