The Future of Search

Robert Scoble writes about attention information. is one step ahead of you there, Robert. The rankings you see on the home page and in your WP dashboard are computed from attention information. How do we get it? We’re the host. We can aggregate more attention information than you might think, while maintaining a very friendly privacy policy because we anonymize the data. Why scrape when you can host? Flickr recently shipped a new API method by which you can order a photostream by interestingness. That’s getting to the root of attention. Whatever is interesting is a magnet for attention. Whatever got attention before is likely to get attention again. Relevance, authority and influence, curtsey to the new king of the hill. I predict that by 2008 we will be calling interestingness old-hat because we will have mastered it by combing our server logs for every grain of useful data, maybe even implementing some new HTTP headers to fatten the stock. By 2009 there will be so much attention spam that the search for the next big search factor will be as hot as it is today. While search engines are computing past interestingness from weeks and months of aggregated attention information, the real pioneers will be focused on potential interestingness. When we know what is interesting, we can begin to conceive of what might become interesting. If you really get the war on comment spam, you know that identifying spam boils down to one thing: potential uninterestingness. The computer must look at a new blog comment and determine, without any prior human interaction with that exact comment, whether it will be interesting to a human in that context. It doesn’t have weeks and months in which to do this. It has milliseconds. If we can determine programatically what will be uninteresting in a given context, as with spam, we can do the same for potential interestingness. And it can be done seconds after publishing. That’s live.

2 thoughts on “The Future of Search”

Comments are closed.