Yeah, I mostly use emacs, which is pretty terminal friendly, or ssh-piped Jupyter, which I love and hate. I used to try to make those remote mounting solutions work.

One thing you can do is run a Linux VM locally with VMWare’s shared folder stuff (annoying to get working but solid once working). Develop using your local IDE, test locally on your VM, and then check in/check out on the heavy weight machine once you have something ready to run something big. Depends on what you’re doing, though. For machine learning, Jupyter is niceish.