My friends often tell me: "Even if a topic has been covered, write it anyway! People come for your unique perspective." I think this is wrong for anybody without their own parasocial cult of personality.

When I think about the audience, the main economic model for my external writing is monopolistic competition within the marketplace of ideas.

In other words, for a specific sub-sub niche (e.g. Ted Chiang book reviews, or a summary article on bee welfare, or an introduction to anthropic reasoning), it is very important for me that, at the point of publication, my writing is plausibly a contender for the very best post in that genre. This isn’t insanely difficult, but it’s fairly hard!

My main model is that assuming search and recommender systems are good, people want to read the best article for their sub-sub topic of interest first. So I primarily provide value to readers if my articles are better than the existing contenders. 

This isn’t completely true: the “best articles” in a given genre don’t provide 100% of the value, even with perfect search. For example, if a concept is difficult to understand, readers might benefit from reading not just the best explanation, but the second and third best explanations as well, to approach a fuller understanding. Similarly, die-hard fans of an author, or very careful book choosers, might wish to consult 3-5 book reviews rather than just one.

However, I think my model is correct enough. The second best post in a sub-sub-niche might provide a significant fraction of the value. The tenth best post is effectively useless.

(At least, my model applies when the writing is intended primarily for a global audience to read, as opposed to e.g. public writing as a form of taking notes for my own edification, or as part of a work trial, or to fill my 500+ word blog daily goals)

Incidentally, this is why I’m also skeptical of my own ability to write interesting fiction. This is because fiction competition is much steeper than nonfiction competition, since a broader basin of ideas and implementations can be substitutes for each other. An explainer post on confirmation bias is not an adequate substitute for an explainer post on anthropic bias. However, a short story that relies on confirmation bias can easily substitute well for a short story on anthropic bias. Further, readers have a poorer model for what they’re interested in with regard to fiction than nonfiction. Thus, fiction has a much higher absolute quality bar to clear before it’s worthwhile for readers.

Sometimes when I express that I worry I don’t have original thoughts, interesting angles, or a plan for unusually excellent execution on X topic, my friends or subscribers try to comfort me by saying that even if the best angle on X has already been covered, they’re interested in my unique voice and angle on X. 

This might be literally true for a small number of people, but as a small fish in the marketplace of ideas without my own parasocial cult of personality, I don’t think this is really applicable. Say I have 750 subscribers. Optimistically 250 of them are unusually interested in my specific opinions (as opposed to following me for unique insights, or because their hands slipped on the “subscribe” button). And maybe 100 of them will actually open an email.

In contrast, if I write a post for a general audience, and I have novel insights with solid writing execution, I can hope for 2,000-30,000 views. Views are of course not everything, but it’s a proxy for whether my ideas actually have the potential to make a difference, especially given the lack of very precise audience targeting and audience segmentation (e.g. it's not like Congresspeople and billionaires are reliably reading my writing). 

So it seems like the simple algorithmic economics and sociology of ideas point heavily towards trying to say novel and insightful things packaged well, as opposed to just getting my “own takes” out there, even when what I say is obvious or redundant.