Recommendations for coding with/without LLMs
Hi! 1st year grad student in cmp experiment here
I'm not too keen on using LLMs for studying/learning physics due to previous attempts leaving me feeling like I didn't get much out it. However, from my understanding, they seem to be pretty decent at writing code.
For context, I did a cs minor in undergrad, but feel like most of the rinky-dink class projects don't really translate to the professional (or professional-ish) code you see out in the wild, and I'd like to learn both for my own pleasure and for practical purposes (see next paragraph), and feel like using an LLM might be useful and create bad habits at the same time.
I'm currently writing some python scripts to control instruments through rs232 and port connections to avoid having to use NI Labview, and while that is honestly going pretty well, I realize my code could be a lot more general and handle different use cases that would require me to restructure my in ways I'm not well versed.
My list of questions I guess would go something like:
- What are some general tips on how to improve my code besides just keeping the habit up
- What LLMs to use/avoid, how to best integrate them
- What are some other non-AI based resources (textbooks, webpages) that are useful for learning good coding practices and how to have better coding standards. I'm assuming this depends a lot on the language so anything for C++, C or python would be great
Thanks!