I'm actually fairly impressed with this demonstration. It's showing actual spontaneous learning. Admittedly, of the most trivial sort. Admittedly, guided by another being. But that's how all human beings start learning too.
I'm also impressed by how human-like Asimo's head movements are. And I think it's more than just a case of the clever Japanese making something that fakes it well. Asimo seems to really react like a being that can actually see. He notices important objects in his environment and head-tracks them just like a person does.
I was also surprised to hear him say "toy car" when presented with the toy mini cooper that he'd never seen before. He obviously "knows" the general category of toy car, and can classify new things accordingly. The chair and table demonstration pretty much nails this point home.
I'm actually very curious about the mechanics of how he knew the stool was a chair and not a table. We human beings know a stool is a chair because they both have the same purpose. We can imagine a person sitting on both of them comfortably, by using our visual imagination. I don't believe Asimo is advanced enough to have a visual imagination (and I know he doesn't have mirror neurons) so he must be using some other criteria. I suspect it's size, or height, or a combination of both. The stool is much more similar in size and height to the other chair than to the table. I would be very interested to present him with small, low coffee table with a surface about the size of the stool's, and see what he makes of it. I believe that he will call it a chair.