Like any good PR review, this article finds flaws in the current approach and suggests direct improvements.
Personally, I’d prefer to do pair programming most of every day, I’ve never tried mob programming in my thirty years of writing code. I’ve had mostly really good experiences with pair programming, and a few that weren’t so good.
The proposed solution sounds vaguely nightmarish as a default way of working, though I’ve had a lot of good experiences where everybody was focused on some aspect of the same problem. That aside, just framing the question is pretty interesting. We are constantly agonizing over how to get more people to do code review, and code review is kind of an awful experience in a lot of contexts. This is going to be on my mind for a while.