Not OP, but I have, and I'm not impressed. Perhaps it would be good if you are doing things in the training data (like churning out yet another bland website). But if you are doing things that are not well represented in publicly available code, or in unusual languages, then LLMs are not great currently.
They are great for finding the right ffmpg command line arguments though, and not having to deal with that mess must count for something.
For what it's worth, engineers at proprietary trading firms are finding that they can get substantial productivity boosts by using LLMs and it'd be hard to find a domain that's less represented in the training data.
Does definitely require more steering than if you're churning out slop apps/websites though.
They are great for finding the right ffmpg command line arguments though, and not having to deal with that mess must count for something.