Write out what you want to do in a single line. Then make a list of items needed to achieve that. Then split each of those items into 3s. Then split each of those items into 3s, again. Do this enough, and you will have the code you want to write. Then it just becomes a pseudo > real code transition, and LLMs are pretty decent at that.
Debugging is great. If I get block because the code is too complex, debugging is a good way to interact and understand the code. It starts to make more sense.
Console logging is still coding too! Debugging can be console logging or stepping through with a debugger. Or even run the program with different inputs.