We’ve looked at the banality of problems that plague execution, the limitations of the perfect worlds of strategy, and the deep, often conflicting, hierarchies that underlie action. We’ve seen how difficult it is to create, communicate, and scale coherent action hierarchies. But these might still seem like practical challenges that can be overcome if we could just make our models accurate enough, our meta-capabilities developed enough, our goals clear enough, our language explicit enough, and our incentives powerful enough. Maybe not today, but in the future. Maybe if we could gather sufficient empirical evidence, develop direct brain-to-brain communication, build an optimal planning computer, or otherwise improve our capabilities, we could converge to solutions that are fair, stable, effective, and objectively correct.
It is common to blame lack of progress towards convergence on bad faith or incompetence. It is obvious how power, money, fame, or nepotism incentivize defense of the status quo. It is easy to find examples of hypocrisy, immorality, laziness, or unfairness. It is tempting to reach a comforting, but wrong, conclusion that perverse incentives, bad people, and flawed systems are the reason we don’t move towards obvious solutions.
There are, of course, massive practical imperfections that can be improved. But underneath all the misunderstandings, inefficiencies, bad faith, and practical challenges there are legitimate disagreements that have been engaged in good faith by smart, conscientious, and resourceful people for millennia; disagreements that have proven impossible to conclusively resolve even with the purest of intentions and even in the most idealized of thought experiments. Underneath the status quo are concessions – not just to the practical difficulty, but – to the theoretical impossibility of complete agreement.