I think we’re applying computing concepts to human problems here as a not-analogy instead of aiming for the heart: that both are governed by queuing theory.
I think you are right. It will be fun to create a queueing model for this problem. I really enjoy operations research but I have only done combinatorial modelling.
Another factor which is related to synchronisation is the need for external I/O, if teams depend too much on external or "system teams" to perform then they are bottlenecked by those interactions
Many large companies are still organised this way, where there are teams that are the gatekeepers or only ones allowed to do a certain task
I think one thing posts like this miss out is teams are made up of humans It may be a fun thought exercise to compare teams to parallel computing but one key thing is team members are not necessarily interchangeable or have the same expertise. They also don't produce at the same rate everyday. I would be very skeptical of applying these principles to teams in real life.
Does this post claim that they're interchangeable or actually machines? It's talking about the impact of coordination overhead, through a computing lens. This is a well-studied area already if you want to drop the computing lens and go to operations research.
And yet, we know that work stealing functions both in parallel computing and Kanban or other LWIP systems.