An Oakland, Calif. start-up is launching new search technology that lets you easily narrow Web queries and learns about your preferences based on what results you choose to view.
Surf Canyon offers up a list of Web results to queries and if you hover over a particular link or click on a target icon next to a link, more results that are related to that particular page are displayed underneath it.
As you click on links, the system begins to learn, for instance, whether you are interested in Miami football or sea mammals if you typed in "dolphins." The more you click on results, the more you train the system.
Surf Canyon is not a search engine (it doesn't crawl the Internet, create an index, order pages by relevancy and provide a query interface). Rather, it is a "queryless discovery engine," says Mark Cramer, chief executive of the company.
"It is technology that sits on top of a third-party search engine, like Google," he says. "We dig into the results and pull forward the relevant ones based on the real-time behavior of the user."