Great write-up, and as always, really impressive. Handling GPU-intensive visualizations remotely can be really tough. I've dealt with similar in the EDA space.

Great write-up, and as always, really impressive. Handling GPU-intensive visualizations remotely can be really tough. I've dealt with similar in the EDA space. I realize this is new and in still in beta, but do currently have any performance characteristics? Is the application latency/responsiveness tolerable?