The killer combo for us is remote+devcontainers. We use this for GPU access, so you can run a project on one or more GPUs that are on a remote server. You can spin up LLM stuff from a mac air and get the best of both worlds.

Nice! Do you have any tips or recommendations for a decent setup?