Watson is probably a capability overreach thanks to some marketing gurus at IBM. If you think about complex games like Chess, or Mario, or Find Waldo, and then

Watson is probably a capability overreach thanks to some marketing gurus at IBM. If you think about complex games like Chess, or Mario, or Find Waldo, and then think about the abstractions we use to reason about very complex topics, they aren't terribly different. Games like Cities:Skylines are being looked at seriously for modeling civic infrastructure. If we can train a system to reason about things using the abstractions we also use, and they're as successful as a human is using those abstractions (or better), then we've "won" in a sense. Find Waldo isn't too far removed from "find the right screw in the parts bin" at a factory.