They really pushed the "infinite storage" point in their presentation. But didn't they mention something about predicting which files to cache before they are r

They really pushed the "infinite storage" point in their presentation. But didn't they mention something about predicting which files to cache before they are requested? I would have liked to see more focus on that. Otherwise, it doesn't seem much different from dropbox, which is more established.