I am in an intense love-hate relationship with Spotify. It makes good mixes for me, I have found a lot of great bands that way. BUT IT KEEPS REGURGITATING THE SAME SONGS IN THERE. I know about Song Radios and Artist Radios, so please don’t recommend those. Smart Shuttle doesn’t cut it, either.
The best thing would be several randomness sliders. One for randomness in bands, the other for randnomness of their songs, one for genres. Please tell me some music service has implemented something like that. I am dying over here.
I’m ready for downvotes but I use Apple Music and it plays me new music all the time. And if it repeats a song I just say suggest less and it goes away. I play my station and it’s a nice mix of music I like and new stuff.
And it is free to stream!(2015)And it is only 2.99 $(2018)And it’s only 4.99!(2020) And it’s only 10.99! (2023)We went from free to 132$ in 8 years for (check notes) listening to the radio.
Is everyone cool with paying 20$ a month in 2028?
I think this article can help you https://www.newvisiontheatres.com/spotify-shuffle-plays-same-songs
One of the main reasons I stuck with YouTube Music is that I find the algorhitms superior to spotify. Maybe this is placebo, but I think between the 7 mixes + discovery + artist / song radio I always have something that does not feel the same, and I feel like any intentionally searched listening updates the mixes fairly quickly.
I disagree entirely.
I used the free trial. The algorithm was a whole lot shittier than spotify’s recommendations. It had the same annoying vibe as the regular youtube algo where if you watch one video from Jordan Peterson, you suddenly inherit the feed of an Elon Musk worshipping alt right tech bro.
It was the same for music. I listened to one song as a joke then it stopped recommending good music and just made my whole feed into joke songs. Much like youtube’s algorithm, youtube music algo clearly uses google’s machnine learning tech (which they use for ads too) where it tries so hard to predict what you like without real data, instead preferring to use solely other people’s browsing habits rather than creating a unique profile for that user without it making too many assumptions right off the bat. Perhaps, I’d describe that algorithm as “HIGHLY reductive” when compared to any other recommendatiion algorithm which seem more geared toward slowly discovering the tastes of its users.
