Nobody thinks that pocket calculators are intelligent because they can do square roots much faster than average people can do. The program that beats a human at chess cannot fly a plane or even beat somebody at tic-tac-toe, you need a different specialized program for that. Nobody thinks that calculating very quickly a number of huge but finite alternatives is intelligence (takes intelligence to program it and make it "smarter" in how it chooses the alternatives and eliminates bad cases in a faster way, but not "smart" in the sense of understanding anything else than the strictly defined problem). Computing something very fast is different than intelligence. A big collection of facts (data) and speed of processing are important to intelligent people and they make them more "intelligent", but those elements are not what intelligence is. None of your examples show actual intelligence, they are either fast processing, or fast data access or combinations of that, none of them show the needed flexibility and understanding of abstract concepts that are the characteristics of the intelligence. Wilson can come up with E=mc^2 from a pun about Einstein by looking up the info in a database and matching word frequency, but that doesn't mean it understands anything about the equation and its implications. I want to see more learning and understanding of abstract concepts to be able to say that a system is intelligent. A dog can at least learn new tricks... Deep Blue can plays great chess, but it cannot learn anything else or even apply the simple concepts of chess, like "defense", "attack", "overwhelming force", "timing" to anything else. It basically plays chess without understanding it which is very much possible because it's a limited type of problem that can be brute-forced. Even though the abstract concepts in chess are limited and can probably be implemented in machine code, they are "understood" by the machine only in the context of chess, they will not translate to anything else, as far as I know learning to play a great game of chess would not help Deep Blue to play even a weak game of Go which is conceptual very close to chess. I need to start to see concepts being applied in different fields, learning, flexibility to assume that computers show some intelligence.