Those are some examples of AI, it's a fairly broad field. I think almost everyone who works in AI or works in some sub-field and really identifies pretty close with that sub-field. What are the current challenges in AI? I'll give you some examples. So there is a current challenge of going from what's called signals to symbols, or syntax to semantics. So think of those images, the images come in as signals. They come in as the amount of light that falls on a sensor in particular places. We do a lot of basically, very fancy signal processing to get more singles out. Is this a dog? Is this a cat? Is this a person? Whenever else. We think, and we may be wrong about this, we think the human brain at some point, talks about symbols. I talk about those abstract level of symbols. I talk about people and I talk about how language is practically symbols right there, these discrete symbols we put together. Where do those things connect? It's not that no-one has any ideas about this but this is a current interesting topic. There's a notion of safe exploration. So that is if my robot, or whatever my artificial gadgets, has to go gather it's own data like small children do to some degree, how do I gather data and get examples of lots of interesting things without essentially, killing myself off in the process? If you have some notion of what I shouldn't be going and collecting data about before I even know what the effect is. So how do I do that generalization? I think there are issues about being social. It's not quite clear what you want in artificial intelligence in terms of social. Not enabling humans to be social, but the artificial intelligence itself being social. Should it be social in the same way as human beings? Should it be social in a different way? Then finally, there are clearly a lot of issues right now about the oversight of artificial intelligence as a technology. My personal view is that, that's actually not very different than many other technologies. It's just perhaps currently a little more scary, a little less understood. So for instance, a table saw is a pretty dangerous tool, and when I use a table saw I'm very careful about how to use it and I understand how the table saw works to some degree. I don't understand exactly how that blade is coming up and throwing off pieces of wood, but I do have an operational understanding of when I use the table saw and I do this, it's going to kick the wood that way or if I do this, my thumb could go through the table saw, things like that. I think the average person doesn't have the same kind of understanding about artificial intelligence. Artificial Intelligence is, I think, more complex than the average table saw. But that needs to be built up with any technology, and there are risks. There are risks to table saws: you should not let children on table saws, you should not use a table saw unattended, these sorts of thing. The same thing has to be developed for artificial intelligence. So that's a broad overview of AI as a field, its history, maybe some ideas of how it tries to think about things, and how it has evolved in a bunch of different sub-problems. I hope you enjoyed it.