Honestly if there's one topic you can trust these language models to be relatively expert in — just based on the amount of internal testing they have undergone at the hands of myopic autodidact techies — it's basic coding / stats / data science stuff. (Obviously caveat emptor, all LLMs will fabulate when pushed hard enough in a so-called "out-of-distribution" direction... but IMO it will take you a while to get there.) While it may make sense to supplement with some kind of structured training/course — especially if you want to eventually make larger-scale sites/software — I wouldn't immediately spend $1000s on it like one might have a couple years ago simply because you can ask soooo many basic and goofy technical questions to a $20/month (and/or free!) model (available 24/7) that you'd be embarrassed to ask a human, even if you were paying them.