I've tried Nim with ChatGPT, which uses GPT3.5 I think. Nim is one language I 
wouldn't use because there's not enough published code to train on to be good 
enough. I have also seen this tech make up APIs that don't exist, and for Nim 
it seemed worse. The situation should improve with time, but it's really 
dependent on more Nim code being published, which is tied to the size of the 
community.

Reply via email to