It will come. In fact, GPT is really good at correcting its code given the compiler output. We just need to automate that. I think that this is not done yet b

It will come. In fact, GPT is really good at correcting its code given the compiler output. We just need to automate that. I think that this is not done yet because LLMs costs are O(n+1) so you really don’t want it to be stuck in some loop.