That's one of those non-sequiturs so bizarre that it's hard to reply. I think it's because it's misleading in two dimensions: obviously rape is different from (all kinds of) consensual sex because it's, well, non-consensual; lets call this the dimension of coercivity.

It's also different from other (consensual or non-consensual) activities because it's sex; the same way that it's different from playing frisbee or (relating to the sibling comment) fixing a car; let's call this the dimension of intimacy.

So in binary terms, we have four quadrants, with prototypical examples of playing frisbee (non-coerced, non-intimate), forcing someone to fix a car (coerced, non-intimate), prostitution (non-coerced, intimate), rape (coerced, intimate).

I hasten to add that prostitution is often, in fact, very clearly coerced (by threat of violence), and of course there's an argument to be made that many other instances are also coerced (by threat of poverty).

Edit: Re-reading my comment, calling it misleading and bizzare seems unkind, I apologize. Thinking in terms of these two dimension is (I think) helpful, but it's also an arbitrary choice. When you choose other (or fewer) "variables", the point made in the article follows.