Comment by chenzhekl

maybe you are not bottlnecked by coding. but there is high probability that you will be bottlenecked by verifying the correctness of LLM-generated code.