These are two completely different statistics: 1) last year, new releases sold poorly, perhaps unexpectedly (NYT article's claim); 2) last year, a relatively small portion of
all currently released titles sold a lot of books, and most titles sold few books (linked article's data).
Pulling from the central table, last year: one title sold over 1 million copies; ten titles sold a collective 5-10 million copies; 267 titles sold a collective 26-133 million copies; 7,294 titles sold a collective 70-700 million copies; and 2.6 million titles sold between 0 and 2.6 BILLION copies. My guess is that last number is way closer to zero than 2.6 billion, so I'll exclude it when I say: the table the author cites shows that last year sold 100-840 million copies. Digital copies! Nearly double that when you include print books. So there's no support for the author's claim, "Books as a medium just don’t have an audience—or rather, they have a very niche audience."
Was last year bad for new release sales? Seemingly, according to NYT. Was last year relatively bad for all book sales? Maybe, I don't know. Does the data support the implication that 0.0001% of new authors will make a living? No. It says that 0.0001% of all currently selling books will single-handedly earn a living for their author in that year.
My wild guess is that authors make most of their money from new releases, so we'd really need to see the data underlying the NYT article on new releases, not this article on all online book sales.