I am excited about continual reinforcement learning (RL). When I first learned about RL, I thought that it was too general for its own good. And yet, continual learning lies outside of the scope of current RL fundamental research. It’s an exciting time because very little is understood about how reinforcement learning methods work with neural networks on simple problems. Yet, many interesting problems require not only RL but continual learning (either because the environment is changing in some unknown way or the environment may include interaction with a human). We are still at the very early stages, but I expect there to be synergy with current developments like LLMs.

I would suggest just getting a laptop and nice external interface (keyboard/mouse if you prefer, nice monitor) + remote server. I bought a desktop + gpu setup back when I started my masters, but I use it shockingly little for work. The type of work that a single gpu + local machine incetivize are usually against good scientific and experimental practice. You dont really want that running jobs during the day.

As for specific cloud reccomendations, I have none. I just use what is available at my institution.