I've greatly benefited from investing just a few hours then utilizing LLMs. I was very unfamiliar to a lot of the syntax from the nix language, so I spent most of my time getting an overview. It was less than 8 hours total.
Another thing that helps a lot is to browse other people's codes on Github.
This is anecdotal but I've been using bluetuith for a year on my nix machine with several adapters then switched to bluetui recently, and my experience with bluetui has been amazing. I just ran into issues all the time with bluetuith and I assumed it was due to the OS or something.
An example I can think of is trusted connections- my devices just never automatically connected once switched on, but now they do.
I linked "Negative Parallelisms" because it's relevant here, but the article in general covers a lot of AI writing styles