(1) Start in google colab if you haven’t run local python files before or you don’t know what a local python environment is. This avoids dealing with setting up the environment on your computer (python people may tell you this is not a big deal but it’s a deferrable learning curve)

(2) Ask the llm the smallest possible next useful question you don’t know the answer to. eg “how can I access a csv in my colab file”, “where should I put my python code in google colab”, “I want to try writing python in google colab, what’s the first, simplest thing I can do”, “create a function that takes a pair of locations (stored as lat longs) and returns the distance in km” etc. I think the key is to make your question as precise as possible, and usually this is iterative as you realize your prior question is not precise enough.

(3) If you don’t understand something in the code the llm gives you, ask the llm about it. Copy and paste code only when you understand what each line does, and keep your code organized— use functions liberally, and put all your functions near each other in your file (if you don’t understand what this means, ask the llm).

Sounds like the LGO person had similar recommendations, but the devil is in the details I suspect. Iteratively doing (2) and (3) are essentially how I use llms myself if I’m trying to write unfamiliar code.

Fair warning, I have no idea what you’re supposed to be doing for this OS.

—

Copilot seems fine from when I use it, but I think having the ability to ask the llm conceptual questions is important learners (imo).