It certainly means I get to review a lot of code that the author doesn’t understand. That alone makes me wish we could put the cat back in the bag.
It did. I spend magnitudes less time browsing docs and can concentrate on actual programming. LLMs are very nice at searching docs for something you are not sure that even exists.
Except when the bot tells you it exists, but it doesn’t. This can be managed with some prompt-fu, but it’s still annoying sometimes.
I use quite a few niche libraries and the hallucination rate is unacceptably high.
I find it works for popular libraries though.Still better than looking for it manually. Checking results is easier than doing the search manually. Manual search is especially fun if you don’t know the terminology.
