Hit Machines

Two articles worth reading together:

Rob Walker on Pandora in the New York Times and an older piece from Malcolm Gladwell in the New Yorker. Both are about tech companies working to “predict” artistic tastes based on the formal characteristics of the art in question. The first is about music, the second is about film; but both deal with the problem of trying to use objective measures to make sense of subjective judgements.

Pandora has come a long way since its inception and I genuinely enjoy it as a way of finding out about new music (though I’ve used it less since there was a crackdown on Canadian user a little while back). Still, the connections it makes between songs and artists is at least worth using as part of a wider strategy of finding out about new music. The film story line prediction service Gladwell talks about seems a little less scientific. I’m not sure why music strikes me as easier to codify than film, but I guess where ever there’s money to be made from making the subjective more objective, then companies like Pandora or Epagogix will be trying to figure out the formula.

As Walker points out though, and as anyone who’s read Carl Wilson’s Let’s Talk About Love knows, trying to get rid of the cultural and social baggage that comes with art is ultimately a futile process. Pandora’s model rests on the belief that people’s music tastes should be based on purely musical attributes. Forget what your friends like, what the latest mp3 blogs recommend, or what Pitchfork said. Pandora thinks this shouldn’t matter when making musical decisions. But we’re social creatures at heart and we express our sociality through art. Stripping music, or film, or books of all the cultural infrastructure that gets built up around them might lead us to interesting musical discoveries, but there’s no art to it. It’s pure science.