A key skill when coding with AI is code review and feedback. Reviewing the output to guide the LLM to get it right.

I find the best prompt pattern is “will this work? Won’t that actually do…?” If I declare it won’t work, the LLM will change it, even when it was right.