According to Zena Iovino, writing in New Scientist, in the future the news will find you – at least according to Futureful, a Finnish startup building a predictive discovery iPad app that will deliver personalised information. The company’s mission is to “give you what you want before you knew you wanted it.”
Hmm, this all sounds very Zeitgeist-y, no doubt. But isn’t Futureful just another wannabe start-up aiming to hoover up as much venture capitalist funding before it predictably sinks without a trace?
Its methods seem sound enough. Futureful’s algorithms glean information from social feeds, such as Facebook, LinkedIn, Twitter, Delicious, Tumblr and Flickr, to locate trending topics. Algorithms crunch users’ interests, behaviour and posts with those of friends and other users to suggest subjects for further exploration. Tapping one or more of these subjects on the touch-screen interface will generate a collection of web pages that can be swiped like a magazine.
But perhaps Futureful might be onto something here. For it seems unlikely that future web exploration will always involve typing queries into Google’s search box. Neither will it be purely social. Last autumn, former Google CEO Eric Schmidt said, “the ability to tell me things I didn’t know but am probably very interested in is the next great stage of search.”
“It’s unclear what exactly the next stage of search will be. Nobody has nailed it yet,” says Anderson.
Futureful says it is bringing serendipity into the game by proposing sites beyond users’ pre-defined interests, an idea intended to counter suggestions that a high degree of personalised information will inevitably narrow people’s worldview.
“What is the point of personalisation if it only reaffirms what you already know,” says Marko Anderson of Futureful.
Defining serendipity can be tricky. Robin Burke, a computer scientist at DePaul University in Chicago, points out that it doesn’t mean showing users random web pages.
“We’re not just suggesting things that people would be interested in, but we’re also helping them discover new stuff,” says Anderson, whose app relies on a combination of natural language processing and network mathematics. The more data the system has on a large number of people, the easier it becomes to find commonalities between individuals.
“Having a system be capable of understanding what’s a reasonable recommendation, that’s a challenge for sure,” says Burke, adding that Google is also increasingly acting as a recommender system, rather than just a web search engine.