I believe we need to find more interesting and useful ways of providing usable, enlgihtening, enjoyable interfaces which provide real people with real insight and real agency when it comes to sorting through an increasingly large music collection.
Either way, learning from listeners’ personal context is one way of providing agency and ease-of-use, and it would be worth thinking about the basic technical aggregators of that contextual information (to infer meaning from time, location, mood, backgrounded/foregrounded etc).
Devices and services which learn from user behaviour could provide a richly variegated middle-ground between the very different models of how most people listen to music and how digital music service users listen to music.
← Guardian Mobility - Greenland Expedition | DE-ACCM application note: Hacking an alarm clock →