I crossed over from Project Management into Engineering, and my experience was that as a PM my job was to be bothered from Sunup til Sundown and periodically in between. I was expected to know everything about anything that I was asked about, and I was responsible for a large number of people's output without having any authority over them.
It was horrible.
I think the issue with the SO question is simply the view that what PMs and BAs do isn't 'real' work. They might not write code, but sometimes they're the only ones making it so the engineers can.
Or a large, international organization with employees and clients all around the world, each of whom believe that their problem is the most important thing in the world.
I've worked for some of the largest corporations in the world, 3 of them in the fortune 100 one of which is a fortune 20 company.
I've seen it done well and done poorly at that scale. What the person I replied to was describing is a dysfunctional organization, not a functional large organization. Creating a large functional organization is just like creating good abstractions in software. You have to figure out how to structure things so that teams/business units can work together with minimal communication across teams.
I disagree. A PM is an excellent communicator between engineers and the rest of the organization. It's not a matter of dysfunction, but a way of filtering what is important and what isn't.
I've been in both situations where a) I am in direct contact with the rest of the organization and b) where an intermediary is. I prefer B and that is what a PM is for.
Some large organizations may be able to eliminate the middle man, but I wager they've dug a ditch so deep they can't tell the difference.
It was horrible.
I think the issue with the SO question is simply the view that what PMs and BAs do isn't 'real' work. They might not write code, but sometimes they're the only ones making it so the engineers can.