Discovery on mp3.com was horrible, I basically had to browse a list of artists by poorly defined genres. Back then not as many genre labels existed and tagging wasn't quite a thing so much as now since the #tag syntax hadn't been invented yet.

As a result I spent hours wandering around through the site finding music I liked. I don't have the time to do that anymore, so what made the site wonderful back then (being forced to dig deep) just wouldn't work for me now. :(

Discovery is a pretty interesting problem-space! It's one of those things that I'm looking forward to "earning the right" to solve (by having enough data to begin with, lol).

I like the idea of "people with your listening/purchasing habits also purchase this". Or "people in your geo purchased this", or "here's the music of people performing in your area this weekend".

Spotify/Apple Music/etc. (the "streamers") have a very different incentive model from the Bandcamps of the world, because their income stream is super concentrated on the major labels and heavily tied to plays of that music in particular. So they're biased in favor of that "kind" of music in discovery.

They actually are averse to showing people hyper-niche music, which I think is why discovery is such a tricky problem for them to "solve": their salary depends on them not fully exploring the solution space.

I think moving out of the universe of royalty-based revenue is a huge step in the right direction for somebody trying to solve that problem at scale, even if it's a smaller market.