Will Machine Learning Outperform Human Curation in Apple Music?
Steven Levy

I actually appreciate Trent Reznor and his music. A friend introduced me to him in the nineties, and although I thought a lot of it was just angsty noise (the Quake soundtrack in particular —worst missed opportunity in game music history if you ask me). But sometimes I just had to respect his intensity, sincerity and craftsmansship, even when the lyrics were over the top or disagreeable. “Hurt” in particular, really impressed me.

But thing is, I have zero confidence Trent Reznor would understand my music taste. You think he would be able to tell good vintage gospel from bad — either “objectively”, or as I see it? Renaissance music? Norwegian folk-pop?

But that isn’t why Apple has hired him either, is it. Apple has always had the attitude that they know best. Listening to customers isn’t necessary, all they ask for is “faster horses”, don’t they? They wouldn’t like to understand my taste, they would like to shape it. Trent Reznor might actually not be a terrible person for that job, as these things go. But I’m not interested in joining the Apple Fashion club.

I’m not interested in the hippest homo sapiens making me a playlist, because like most people, I’m not all that hip. Personalized recommendation is infinitely more valuable to me. Like many, I’ve found that Spotify Discover Weekly “gets me” — at least, it gets me more than anything or anyone else. I still only save 4–5 songs per week, and it has some bad misses. But for me, that is a great result. I’m used to a much worse track record from would-be tastemakers.

Like what you read? Give Harald K. a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.