I notice my usage goes in waves. There are days and weeks I use it constantly for things, but then I don't touch it for a bit. Its like I already learned what

I notice my usage goes in waves. There are days and weeks I use it constantly for things, but then I don't touch it for a bit. Its like I already learned what I needed to and don't have a use for it. Regarding LLMs, we have a usecase but we need local models. The local models by itself are not good enough, so we need to train/fine-tune. To fine-tune we need a beefy computer. We have the beefy computer but its air-gapped... Ugh hahaha. It makes dev quite slow. The other thing is that there is a learning curve on LLMs. You get burned by bad information and you might not use it until you learn of a usecase that doesnt require the information to be great. For instance, I'll describe a meeting/person I'm talking to, then ask what are 2 of the best questions I could ask. These questions have been great. That is better than asking about an easily google-able question. You need to be exposed or creative enough to come up with the prompts.