Elliot Van Buskirk had a recent post on Caterina Fake's (pictured on left) recent talk at the Wired Business Conference titled, "Mapping Desire, Exploring the Science of Online Recommendations."
Her observations on personal data and new web functionality are interesting: “One of the things that we saw with the efflorescence of Web 2.0 was that there are now exobytes and exobytes of data online. The future of the internet goes to whosoever is able to make all of this information work for the benefit of people out there who are trying to find things and are standing on the corner of 44th street and 5th avenue and it’s their mother’s birthday in two weeks, and they know it, and their computer knows it, and also knows that her taste is such that if there’s three scarves that are actually in inventory at Sax 5th Avenue, and if she just walks up the five blocks, she can get that for her. If you are able to solve problems like that for people, that will be immensely useful to people.”
Her firm Hunch gathers such a wide variety of data from its users that it may in fact be able to fill in gaps left unfilled by vertical recommendation engines such as Amazon’s and Netflix’s. For example, it has ascertained that people who prefer their sandwiches cut diagonally also prefer Ray-Ban sunglasses.