With a bold move, Amazon has taken another step towards the future. They released an Amazon API, which will allow people to write programs that can obtain and manipulate data from the Amazon product database. Google recently did the same thing. Having recently come from academia, I think this is great. These are two huge groups of information that we all have access to. My mom runs a used bookstore. She could use an application that combines the two to find author information for customers, find out what the latest paperback books are and search google for used copies, etc etc.
One of the things I'm dying to try is to combine the power of AIML with the Google API. AIML is the Artificial Intelligence Markup Language developed by Dr. Richard Wallace. You can speak to an Alice bot at alicefoundation.org to test her out. AIML is based on XML and basically breaks language down into a series of words. These words are then matched against known words in Alice's database and then appropriate answers are spit out. Now, don't get the wrong impression: Alice can't learn from conversation (per se), she can only spit out what has been stored for her. However, Dr. Wallace based her speech patterns on a principle that claims that all of human speech is comprised of a small amount of words used repeatedly. Based on this, Alice can have some decently intelligent conversations and has done well in numerous AI contests.
My thoughts go like this: What if an Alice interpretter was written i such a way so as to use the google index as her knowledge base? I.E., what if she could take your question, break it down, and then search google, grab the most relevant info, and spit it back at you? What about the same with Amazon? All of a suddent, Alice knows what the top selling books are. Alice can find the wishlist for your computer wiz nephew. Alice becomes a rather intelligent agent.
This line of thought interests me a lot. Easire navigation of quality information pools. I'm down with that.