Why isn’t Apple Music better at predictions

https://lemmy.world/post/1140281

Why isn’t Apple Music better at predictions - Lemmy.world

So Apple Music knows all kinds of data about my listening habits. So like I genuinely don’t understand why when I opened the app this morning it wasn’t like hit ok to stream the new Taylor Swift album. I had to like click around for a second. No big deal In the grand scheme of it. Like I feel like I have pretty predictable tastes I listen to mostly albums by a small number of song writers in a few styles but for some reason it’s very hit and miss and single driven even if that’s not a match for the way I listen to music. IT’s algorithmic radio isn’t bad but for having 20 years of data it seems pretty mediocre.

I used to have the same issue, and still do to a certain degree. But after listening to the auto play for a while now and love or disliking songs, the auto play has seemed to have gotten better. From what I read Apple’s algorithm is just super slow to learn you.

I gave up on Apple Music after 3 years and moved back to Spotify a month ago.

Apple Music pros:

  • really good UI (including light/dark mode)
  • spatial & lossless audio
  • included in Apple One subscription

Apple Music cons:

  • no crossfade (still! In 2023!) - useless for parties
  • thinks I want to listen to my Christmas playlist in June
  • generally terrible music suggestion compared to Spotify
  • no ‘playlist radio’ (again, compared to Spotify)
  • music discovery massively sub-par compared to Spotify
They actually introduced Crossfade in IOS 17, still amazed it took this long and the feature already being available on android. I’ve learned that I really have to train the algorithm and it gets better if you tell it what you liked