As @svantana mentioned this unfortunately does not remove any personalization in Song Radio, etc. As long as you're logged in, these auto-generated playlists will contain exactly the same (personalized!) list of songs, regardless of whether you open them with or without the `si` and `utm` parameters. Removing `si` and `utm` just removes marketing / referral tracking information.

That's a completely different thing. Playlists will be personalised as long as you're logged in, which you must be to listen. Even their public playlists get reordered automatically based on your account history.

As a Spotify user, I get the sense that there's less interest these days in refining the music discovery experience at Spotify. Hopefully someone can prove me wrong. The only feature I've used in recent years has been the "Blend" feature where I can combine my most-listened-to songs with a friend's.

I think Spotify sees their "power users" as people who getting the most out of their platform, but probably only provide a small chunk of their music spending on Spotify in particular. It seems that they want to make their money chasing the common denominator, as well as content that can have ads baked in easily (i.e. podcasts).

I think you can empathize with my frustration as a user here—the largest music streaming service with the largest catalog has seemingly made music discovery a secondary priority. That being said, there's more than just their own recommender system that you can use for discovering new music on Spotify:

- "Track IDs" Playlists are usually pretty good. It's the closest thing you'll get to some the fruits of some good crate digging by someone who really knows their stuff.

- The Playlist search feature is a killer app.

- Some record labels are actually listed as artists on Spotify, making it easier to search through their releases.

The core of Spotify's recommendation system is collaborative filtering[1], which means popular stuff often gets filtered to the top. When a breakout artist from a niche genre gets signed to a large label, I often have issues with the artist radio, because now most of that artist's audience isn't coming from the scene that shaped them.

1: https://developers.google.com/machine-learning/recommendatio...