There’s one other angle to the social network story that I wasn’t able to mention in my Wired essay. Right now, retail companies are investing a pretty penny in consumer preference algorithms, that AI software which suggests books to buy on Amazon, and DVD’s to rent on Netflix, and songs to purchase on iTunes. The assumption is that human preferences are predictable – if you like Bob Dylan, you’ll probably like Astral Weeks; if you like How We Decide, you might like Predictably Irrational; if you like The Wire, you might like Homicide, or The Shield.
These algorithms have their limitations, but they can be useful prompts for browsing. (When I’m forced to shop in an actual physical store, which is as rarely as possible, I feel rather lost and adrift. Where is the helpful “Customers Also Bought” window?) But these software programs are missing one crucial source of data – your social network. After all, the best way to figure out what kind of music you like, or whether or not you’ll enjoy Napoleon Dynamite, is to study your friends. If they’re looking forward to the new Monsters of Folk album, then so are you; if they enjoy old kung-fu movies, then you probably do as well.
How would this work? Before long, I bet we’ll have simple preference tools installed on our phones. These tools will track our social network – that’s easy to do on a phone, since our network is stored in the contacts folder, thus bypassing the mess of Facebook – and then give us advice based on the feedback of our friends. I imagine one day we’ll simply be able to walk into a mall and ask the phone where we should go. It will think for a second before suggesting a particular clothing store selling vintage shirts, or a quaint cafe serving a good tuna sandwich, or that poster store on the ground floor. It will be like word-of-mouth, only transmitted indirectly – the hearsay of an algorithm. Your preferences will be calculated as a function of your network, and not as an abstract human island, which strikes me as much closer to the social reality of preference formation.