> corporations knowing about my music taste is not cracking the top 1000

They don't care much about your music taste, that's not the valuable data they're collecting and reselling. They're tracking your location, habits, address, email, passwords, billing info, extensions/other apps, locations, etc etc and storing it insecurely to be easily acquired by people with arguably even less scruples than the corporation they're taking it from because the cost to store this shit securely is higher than the fines/consequences from a data breach, so why bother? All of that (meta)data is for more valuable to them and the numerous "bad actors" that data-harvest them on the regular.

> as opposed to trying to find and rip audio files from the internet and put it on my server and deal with metadata and cover art manually.

This can be automated, which can also be a form of curated music discovery.

> so even if the platform is missing stuff you can fill it in yourself.

With what? All that music you yourself assert you aren't acquiring "manually"? Sounds like self-hosting with a couple of extra steps to me.

> I'm sure it works fine if you've basically settled on what music you like and never listen to anything new, but if you do like to discover new music, self-hosting just isn't an option.

I can't tell if this is ignorance or arrogance, but it's laughably out of touch either way, especially in this day and age. You can just say you're too lazy to fuck with it, don't know how to do it, or don't socialize with people that share your music tastes. It's fine.

There's numerous examples in this comment section of how to do music discovery without subsidizing a company that takes advantage of both the artists that users listen to, and the users themselves, and self-hosting is in no way a barrier to that.