Is There an Alternative to Generalization in Reenforcement Learning?

We don't even know what the applications of reinforcement learning will be, says ritike. The actual applications of ould will not be in our having a robot stack a blue block on top of a breat block. That something the trobats have been doing for 50 years but doe's not actually attask that anybody cares about it. People care about something that's much more practical. But maybe they'v lern the biggest application of reinforcement learning is robots that are more general. They're more general than any kind of robots we have right now. And presumably the specific way in which that will happen, if it will happen, would be by having more general robots.