Hot take... There is something about programming projects spending 10x (or 1000x) the effort to parallelize execution when everything is still restricted by von Neumann architecture.