How Unsupervised Learning Is Progressing?

There is a lot about unsupervised learning which is about the assumption that the data that you want is sort of already out there, and all you need to do is find it. I think my concern is that we should be looking hard at trying to grow the bandwidth of what should the system actually do? How do we verify if the system is doing the right thing? Why are we even doing unsi abite o in the first place? There's so many more of those questionss to be answered, especially as we make these more i think that's right.