I wonder how much of this effect has to do with the mixture of pseudo-neutral & centralized media we have.
The long tail didn’t really address the rich-get-richer trend in even random distributions. (If you have a bunch of evenly connected nodes, all of which randomly decide whether or not to share a signal they recieve with their peers, you get a hockey stick distribution. If some signal starts off with a positive bias, that bias grows enormously. And, if connections are allocated the same way, the number of connections between nodes has the same asymmetry.) If we have a mix of curated, centralized media that promotes some particular figures (and particularly if those figures are promoted toward highly connected clusters of people), we’ll see a much larger growth in the popularity of those figures than we would in a purely centralized context.
On the other hand, the greater the influence of neutral carriers, the easier it is for whole groups to become less affected by some popular figures. I have never heard a song by Drake, Beyonce, Rihanna, or Adele; in 1993, I would have definitely heard any artists of similar popularity, because the only way to discover music would have been radio and television — in other words, curated broadcast media. Because I don’t listen to the radio or watch television, my taste is music is mostly formed by the tastes of my peer group — and presumably many of them have heard those artists but haven’t found them impressive enough to convince their peer group to engage with them (the way that they would for Bablicon, or Skinny Puppy, or other groups that are much more obscure on the global scale but are much more popular in particular groups).
When we look at the head in isolation, we sort of miss the point, which is that favoring organic communication of tastes over broadcast both raises all ships & slightly flattens the distribution. The long tail is longer and fatter, but the head is also higher, because the communication between preferences & the things that fulfill them is more efficient & tastes can evolve more quickly in a resource-rich and diverse environment. Someone who really only likes free jazz or ambient harshnoise would have, in 1993, thought they “didn’t like music”, but now has a much greater likelihood of being exposed to these much less globally popular genres.
(This is not to propose either preference-essentialism or some kind of free market optimism. Preferences evolve in response to experience, and this kind of evolution only happens when the marginal cost of exposure approaches zero. The long tail is made profitable only because piracy has expanded the domain before it, creating a market of savvy and dedicated people willing to spend money on obscure things where before there was only a hostile and alien environment.)