Pandora media’s music genome project has been around for a while, but a recent Independent article article has only just alerted my interest. I’m glad that there’s a corporate role for 400 trained musicologists, but I can’t help thinking that the tendency of all technology like this is to homogenise, to find connections and to keep people’s music selections similar. Which is not at all how people actually listen to music – viewed from the perspective of ‘if you like this, you must like this’, most people have at least some surprises within their music collections. 21st-century technology, skills and marketing techniqes are being employed in the service of a particularly 19th-century idea: that there is one single key to unlocking the personality, tastes and style of an individual. Mapping one’s music tastes under a single set of parameters is the same as reducing Stravinsky’s oeuvre to a small collection of intervals in order to ‘explain’ how the same man could have written The Rite of Spring, Pulcinella and In memoriam Dylan Thomas. It’s arbitrariness disguised as scientific enquiry. The best way to introduce new music to people (and, perhaps, to ‘explain’ the blips and ruptures in their CD racks) is to look at the tastes of people who they trust; if someone you trust hands you a tape and says ‘you must listen to this, you’ll love it’, you will listen, and you probably will love. You’ll at least give it a fair shout, even if you’ve walked past it on the racks of HMV a hundred times.