See my comment below. Copy-editing isn't performed by other academics (or at least it shouldn't be given their other responsibilities, in my view).

And although peer review relies on specialists volunteering their time, finding and vetting those specialists still requires work.

I'm not trying to say that the current publishing model is defensible, but I am pointing out that running a high quality peer reviewed journal still requires at least one or two dedicated workers. Whether you think they should be paid for their time or expected to volunteer is a different issue, but we shouldn't pretend that the whole apparatus is an illusion created by greedy publishers.

I always - always - sent my paper to at least one non-coauthor for a final copy-edit before being submitted.

Also, no editor at any journal (I've published ~20 papers) made any edits to my papers.

This must be highly field-specific then. I published a peer-reviewed paper a couple months ago that came back with around 50 suggested copy edits, as well as a few paragraphs of suggested changes from the journal editor. And then it went through the process again in a lesser form at the page proofs stage. Granted, a lot of those suggested edits were to make it conform to a somewhat arbitrary and tedious house style, but it also caught things that I or the other people I shared the paper with didn't see.

Oh, I didn't include edits to stay within house style, but then, I wouldn't submit a paper that wasn't already in house style.

If the editor sent some copy edits to the text, I'd just send it back to them unedited and ask them to publish it. In fact, I love pushing back against unreasonable editorial requests.