All are inherently rivalrous
Even discounting any financial returns, howling into the void / winking in the dark fails to accomplish anything if there is no receiver or audience.
Ms. Typeset might be engaging in a form of personal journaling or meditation, but in a world that has so many competing distractions her curation is never shared, or even where perhaps it is but the potential community of interest, and the discovery tools possibly reaching it, are fractured and distracted to the point of dissolution ... the effort to communicate and gains of doing so are lost.
Mind that large audiences really aren't communities -- I'm talking Facebook or national broadcast scale. As Dave Winer observed, conversation doesn't scale very well. Dialogue between two people is its most intense form, and past a handfull of participants or so, what remains is mostly a set of serial monologues. Above 15 or so, the graph starts trending increasingly to a star, with one broadcaster and numerous recievers, collectively an audience. At sufficient scale, selling that audience to those who'd hope to advert its attention to their own message becomes a principle commercialisation model. The baiting message tends to the minimum viable common basis.
At extreme levels of specialisation, such as, say, aa maths PhD, the relevant community might easily fit in a small classroom. Possibly in a single car. My cursory analysis of Google+ communities by size and activity levels suggests diminishing returns above about 10,000 registered members, of whom perhaps 1--10% were highly active, so 100--1,000 participants, roughly within Dunbar's Number range. Extremely large groups are visible but not especially useful --- most seemed overrun with spam and memes.
Pursuit of that audience tends strongly to sacrifice or ignore nuanced, specialist, or niche interests, to the point of actively trying to steer potential members of such groups to the larger, and more monetisable, vulgar pool.
My own seeking of expressive outlets is an increasingly frustrating trade-off between the expressivity of the tools themselves (mostly formatting and media capabilities), technical hosting issues, audience discovery and cultivation, search, and related concerns. I'm hoping to learn, share, and discover. The sense that this used to be more viable back in the days when I used Usenet and mailing lists may be a Golden Age illusion, but its a damnedably persistent one.