What Was I Thinking?
Computers may not be able to make decisions for you (yet), but they can sharpen your judgment. By James Fallows. (June 2007)
Online music-matching sites have tried for years to suggest new songs or artists you’ll like based on ones you’ve previously chosen.
One of the earliest was MusicMatch.com, founded by a San Diego entrepreneur named Dennis Mudd. It familiarized users with the idea that if you told it what music you like—Scarlatti, the Beach Boys, Ludacris—it could create a stream of related songs and artists, often ones you didn’t know.
Three years ago, Mudd sold his company to Yahoo, which made it part of Yahoo Music Jukebox. Now two of the best-known services for music matching are Pandora and Last.fm. Pandora’s recommendations are based on its “Music Genome Project,” which analyzes and compares the properties of each work—bass line, instruments used, and on down a list said to include 400 items. Last.fm uses “social intelligence” to create recommendations—people who like the Arcade Fire will probably also like Radiohead.
Mudd has now founded a new company, Slacker, which is supposed to make personalized music available anywhere. A satellite service will broadcast it to a system you can set up in your car; the company will soon sell an iPod-like portable device that can cache songs while it’s connected to the Internet. The service will be free if you’re willing to listen to some ads—or, for $7.50 a month, it will be ad-free and allow users to store songs they particularly like, and skip the ones they don’t.