November 25, 2014
Recommendations—for products, content and experiences—have been getting more and more finely tuned as brands get access to new data streams and learn how to use data more predictively. The latest futuristic idea in predictive technology: speakers that play what they think the listener wants to hear at a given time.
Cone, introduced earlier this year, “learns what, when, and how you like to listen to music. Turn it on and it automatically plays what you’ll love.” The first product from the startup Aether, the stylish Wi-Fi-connected speaker plays music from the streaming service Rdio and from the user’s Apple devices. Pressing a button on Cone lets you verbally request a track or artist, but the idea is that it requires little user intervention, even self-adjusting volume based on learned preferences. “This is the direction our always-on technology is heading: It’s about learning habits as much as predicting taste,” observes a Wall Street Journal reviewer—while also noting that this debut version still has a lot of kinks.
Prizm, which is wrapping up a successful Kickstarter campaign this week, is a similar product from a French company, but it works with multiple music services. The speaker can also adapt based on who is in the room (as long as people have the companion app on their phones), finding music that it predicts everyone will like if several people are present. And Amazon’s new Echo device is a self-learning digital assistant in the form of a speaker for the home that can play music or take instruction from the user (e.g., add items to a to-do list).