All in all, once an organization gets big enough, power does what power wants, and power wants what is good for them in the short term, regardless of what is good for the organization. That's how most large companies end up spending very large amounts of money on things that wouldn't actually pass muster to anyone aiming for the organization's best interest and with actual knowledge of what is being accomplished.
You see new, wide eyed PMs approaching budgeting processes as if the goal really was profitability, or customer satisfaction, or something reasonable. But if they are going to stay as PMs for long, they better realize quick that the vast majority of project proposals have only a passing interest in what will be accomplished, and are mainly about making sure every sub-organization gets fed sufficient money to not lose people, or possibly even grow if the manager is well liked. All the efforts in documentation and justification are just theater.
This has the ring of truth.
Has anyone solved this problem?
Is anyone trying to solve this problem? (Or is everyone in a position to work on the problem just playing the game?)
People could self-organize, on-demand, for a task, and structure whatever hierarchy was appropriate, based on somewhat optimal resource allocations for that task.
(Example: Person A might normally be the most experienced at facilitating the group's coordination, but A is providing key technical expertise for this task. Person B isn't critical path on this task, and has facilitating skills and interest in that role, so B volunteers for that role for the duration of the task.)