How about a computer program that can compete with human beings in a game of Jeopardy:

For decades, humans have struggled to create machines that can extract meaning from human language, with all its messiness, subtle context, humor, and irony. Traditional approaches require a great deal of manual work up front to render material understandable to computer algorithms. The ultimate goal is to make this step unnecessary.

IBM hopes to advance toward this objective with Watson, a computer system that will play Jeopardy!, the popular TV trivia game show, against human contestants. Demonstrations of the system are expected this year, with a final televised matchup--complete with hosting by the show's Alex Trebek--sometime next year. Questions will be spoken aloud by Trebek but fed into the machine in text format during the show.

Very cool.

A computer program that can parse Jeopardy questions (I know: answers) would be the perfect front-end for something like Wolfram Alpha. Take the question and translate it into a query that the Wolfram program can understand, and you're off and running. But while Wolfram Alpha includes a huge knowledge base, I doubt it contains more than a tiny fraction of all the information that can be found on the Web.

To work beyond a pre-built knowledge base, we would need to hook that parser up to a good search engine and let it scan pages for answers that aren't built in. Perhaps it would bring back answers with an estimated reliability. Of course, if we have a parser capable of reading the web for meaning, and capable of estimating the reliability of the information it's receiving, we probably wouldn't want it just to wait around for questions.

Let's get that sucker fired up and see what kind of stuff it can figure out that we didn't even think to ask. I'm sure it would provide us with some amazing insights on the way to, you know, taking over the world.

Via Geekpress.