I once had a role in management at a mid size software company. The company was quite stodgy, with a lot of regulations and processes. If I'd wanted to I could have worked a solid 60 hours a week keeping up with all the process and procedure changes, and then following them. After attempting just this, I realized in a few weeks I did not want to, so I simply stopped. And guess what? Nothing happened. Of course I heard about a few things that needed to be done, and quickly learned exactly which ten of the hundreds of processes were important; or rather had someone who's job it was to track them.
I continued in a this purely reactive manner, saving myself a lot of work and trouble, and had a really high performing team in the organization. There was an article some weeks ago about the merits of doing nothing, indeed this is why I don't believe this article expresses a valid concern.
A generalization of this key insight also appears, illustrated with many historical examples, in The Collapse of Complex Societies by Joseph Tainter:
https://en.wikipedia.org/wiki/Joseph_Tainter
In his book, he outlines how bureaucracy increases in complex societies up to the point they are collapsing under it. The key reason for this collapse is the diminishing marginal return of bureaucracy.
It is my hope that we find a way to prevent societal collapse due to bureaucracy by automating arising tasks cleverly. However, most things I see at the moment, even most computer projects, tend to further increase bureaucracy and even also add new failure modes.