> You should look at the math in the article I linked. It's proportional to the aggregate of everybody - which basically completely disconnects it from what individual users listen to.
I have actually read the article, and that's wholly compatible with my take on things (that before and after Spotify artists are paid roughly proportionally to how much their music is "used"). There are two interesting points that can create the apparent problem/paradox you're seeing:
(1) It's interesting that a typical power user before Spotify would pay more money for more music, and now there's a flat rate (or counting ads, two flat rates).
(2) The definition of "usage" has shifted a bit. Skipping a lot of details, before recordings people paid per performance, with the introduction of CDs and whatnot people are able to pay per unique song listened to (with power users typically though not necessarily buying more CDs), and Spotify shifts things back so that artists are compensated in terms of total listen time.
It's probably true that power users should pay more (if not, and if artists are appropriately compensated, then everyone else is subsidizing their use), and there can be an interesting discussion around how listeners actually derive value from music and if unique songs matter more than total minutes (or if a more complicated but less explainable metric ought to be used), but the big guys dominating everything is not a new phenomenon.
As an aside, the advertising, distribution channels, and network effects enabling the big guys to dominate everything to the extent they do _is_ moderately new on a human timescale, but that would be true independent of the payment scheme and doesn't seem super relevant to the discussion.