Terrific analysis by Ben Lovejoy, 9to5Mac. Apple’s motivations aside, I think music curation is a tough problem to solve.

My musical tastes are all over the place. People who know me very well, musicians who share much of my musical tastes, have a difficult time making recommendations for me. It’s just not easy to figure out what music will press that button for someone.

That said, I’d like to see more experimental interfaces that make it possible for me to wander my musical tree on the fly. I’d like to play some music that I love, then have an interactive training session, via Siri, purely dedicated to exploring “more like this” kind of music, where I can do more than just love or hate the presented choices. Something more conversational that allows me to express specifics of what I enjoy or dislike.

For example, I love soulful music that contains technical artistry (e.g., great singing, complex harmonies, or technical guitar, piano skill, etc.) If you and I were listening to music together, I could point all these things out. But if all I get is “love” or “don’t love”, it becomes much harder to convey the subtlety of my personal tastes.

All this said, I think this ability is coming. With advancements in processing complexity and machine learning, I believe a new window to modeling musical tastes is opening. I have no doubt that someone, somewhere will crack this problem.