They're just people running the peer review and production systems (think the online submission system, solving technical issues, writing emails, etc). It's in the name: assistant. They're not the people looking for peer reviewers or doing anything scientific.
As I point out in https://news.ycombinator.com/item?id=35657036 the costs of having employees that "write emails" scales linearly with the number of submissions that are eligible for peer review, not with the number of papers that are successfully published.
I'm sorry, have you ever been involved in the process of publishing a scientific paper as an author or editor? Because it clearly doesn't sound like it from your description in that linked comment. The numbers you write are just nonsense. The assistant editors also don't interact with submitters - that's typically handled by the online platform. They interact with authors when it's time for reviewing proofs, signing documents, etc. But all that only occurs after papers are already accepted.
I've been an author, and just once; never a reviewer or editor. It was with PLOSOne which is different and, though a lot cheaper, is still $2000.
As an author, I exchanged multiple emails with employees of PLOS even before it was submitted to peer review (or reached the editor desk) because the automated process of submitting my paper to the system crashed and required intervention. And mine was N=1, and it still broke. It's not "Google-quality engineering", if you know what I mean. So, all this "oh, that's just an automated system" just does not hold water.
Now, my turn to insinuate. Have you ever written a multi-step workflow management software system that coordinates thirty email threads? If yes, how much human involvement from the operators of that multi-step workflow is needed? And if not a whole lot, do you work at a "cool" VC-funded modern firm, or at a company that got started with servers in a datacenter in the 1990s?
I'm not even sure why my experience as an author matters. As I said in the very beginning of this giant thread, I work in the industry. Most of the systems I've encountered or heard about are not the beautiful automated platforms critics imagine them to be; and I am not even talking about technical intervention.
