I expect so as well, and my prediction is that we’ll have LLMs that are roughly as capable as the current frontier that can be run locally within a year or two. At that point, it’s just going to be good enough for vast majority of tasks most people need to do.