Really interesting history on how technology enabled different forms of power. I am curious - and only because it wasn't explicitly stated, perhaps due to the fact that it's likely obvious: why do we even need power structures to begin with?
As far as I can tell, we created positions of power to solve for decision making in a large group. Even the most basic tribal power structures relied on having some solution for group decision making and coordination. The forms of 'how decisions get made' changes as groups get larger and technology enables new possibilities, but we still seem to require some form of group decision process.
I'm assuming power structures exist in part because they also provide benefits from being part of the group - shared resources, protection from enemies or natural predators, delegation of complex projects everyone benefits from (e.g. roads). Perhaps that's simply a state narrative installed on its citizens so that they agree to continue to feed the state, but it's hard to imagine how else public infrastructure could created.
Of course AI poses a question mark over everything. Maybe coordination through elected powerful individuals evolves into AI systems that manage coordination - but it shifts greater power to whomever creates and tunes the models and how they're deployed. It still poses questions about what to do wrt group decision making.
Thank you, Muxin. In many ways, this is exactly the question that led me to start writing in the first place.
I agree that coordination, decision-making, and public goods are real problems that every civilization must solve. My concern is less whether power structures are necessary, and more how they evolve, what incentives they develop, and what happens when they become ends in themselves.
Please keep an eye on the articles ahead. Whether I’m writing about AI, states, markets, inequality, or civilization, I often find myself returning to the same underlying question: how large groups of humans coordinate, and what emerges from that coordination over time. I suspect we’ll keep meeting at that question from different directions.
Looking forward! The closest I've seen in real-time to AI powered group decision making is the Collective Intelligence Project - AI interviews groups of people and finds what people agree on. Still subject to abuse from the model's interpretations and data gathering, but feels like a step towards progress: https://www.cip.org/blog/alignment
That's a fascinating example, and in many ways it connects directly to the question I've been circling around throughout this series.
Projects like the Collective Intelligence Project are exploring whether AI can help large groups of people coordinate, deliberate, and discover shared preferences more effectively. That seems genuinely valuable. But it also raises another question: what happens after coordination?
My growing suspicion is that many of the defining challenges of the 21st century are no longer primarily coordination problems. They are authority problems.
AI, climate change, pandemics, global supply chains, internet platforms, cryptocurrencies, and increasingly sophisticated cross-border criminal networks all operate at a global scale. Yet the highest level of legitimate authority in the current world order remains the sovereign state.
In other words, many of our most important problems have become planetary, while our governance structures remain fundamentally national.
This is why I keep returning to sovereignty. Not because I think sovereignty is "bad" or obsolete. Quite the opposite. The sovereign state may be one of the most successful coordination technologies humans have ever created. It solved many problems that earlier forms of organization could not.
But every successful institutional form is shaped by the environment in which it evolved. City-states struggled to govern industrial societies. Empires struggled to govern democratic mass societies. It is not obvious that the nation-state is naturally equipped to govern a civilization increasingly connected by AI, networks, and global systems.
From that perspective, the fundamental tension of our era may not be between democracy and technology, or between states and markets. It may be between the global scale of civilization and the national scale of sovereignty.
Projects like CIP may help us better understand what humanity wants. The harder question is whether humanity possesses institutions capable of acting on those preferences at the same scale as the problems themselves.
That, at least, is one of the questions that keeps pulling me deeper into this series.
That’s a great thread to pull on - looking forward to reading more of the series. You’re right to call out that implementing something everyone agrees on isn’t solved by efforts like CIP, which can only help define what stakeholders want.
We've had international cooperation before - nuclear treaties with inspections and penalties, to Paris Climate Agreement that just asked every country to volunteer limits (didn't work so well). But even nuclear treaties get evaded (although we're probably better off with them than without).
This isn't the same step as having the authority to implement changes, but I started thinking about what happens after a sovereign power sets the rules to address global challenges - how would you enforce the policies and behaviors to keep it going?
That is exactly the next layer of the question. But one small clarification. I am not arguing for a global sovereign or a world government. Personally, I don’t support that idea. I don’t think concentrating even more authority into an even larger hierarchy can solve civilizational scale problems. If anything, it would simply reproduce the same dynamics on a larger scale.
I also do not have an answer to what political structure global governance should take, and that is honestly not the question I am most concerned with.
What increasingly interests me is that the world now behaves less like a collection of independent countries and more like a single civilizational ecosystem. AI, climate change, pandemics, global supply chains, internet platforms, cryptocurrencies, and cross border crime all evolve through networks that exceed territorial borders. Yet we still analyze, explain, justify, defend, and imagine political structures mainly through the hierarchical logic of the pyramid.
That may be the deeper structural mismatch. We see problems operating through ecosystems and networks, but continue to assume that authority must ultimately take the form of a higher level sitting above the lower ones.
So perhaps the mistake is not that we have failed to build a global pyramid. Perhaps the mistake is that we continue to imagine civilization itself as something that must be governed like a pyramid.
I do not know what comes after that model. But I increasingly suspect that simply enlarging the sovereign structure from the nation state to some form of world state would not resolve the underlying problem. It would only scale it up.
That’s a good point - centralizing even more does tend to be our collective knee jerk reaction but it also creates risks of those problems you’ve talked about wrt sovereigns getting even worse.
And we’re assuming that the pyramid is the right way to think about how to “do bigger better things” but it tends to end up concentrating power and wealth upwards while draining the bottom.
What if AI solves some of the coordination costs that used to require humans with official power?
Would that change the pyramid from being more like - squared?
Really interesting history on how technology enabled different forms of power. I am curious - and only because it wasn't explicitly stated, perhaps due to the fact that it's likely obvious: why do we even need power structures to begin with?
As far as I can tell, we created positions of power to solve for decision making in a large group. Even the most basic tribal power structures relied on having some solution for group decision making and coordination. The forms of 'how decisions get made' changes as groups get larger and technology enables new possibilities, but we still seem to require some form of group decision process.
I'm assuming power structures exist in part because they also provide benefits from being part of the group - shared resources, protection from enemies or natural predators, delegation of complex projects everyone benefits from (e.g. roads). Perhaps that's simply a state narrative installed on its citizens so that they agree to continue to feed the state, but it's hard to imagine how else public infrastructure could created.
Of course AI poses a question mark over everything. Maybe coordination through elected powerful individuals evolves into AI systems that manage coordination - but it shifts greater power to whomever creates and tunes the models and how they're deployed. It still poses questions about what to do wrt group decision making.
Thank you, Muxin. In many ways, this is exactly the question that led me to start writing in the first place.
I agree that coordination, decision-making, and public goods are real problems that every civilization must solve. My concern is less whether power structures are necessary, and more how they evolve, what incentives they develop, and what happens when they become ends in themselves.
Please keep an eye on the articles ahead. Whether I’m writing about AI, states, markets, inequality, or civilization, I often find myself returning to the same underlying question: how large groups of humans coordinate, and what emerges from that coordination over time. I suspect we’ll keep meeting at that question from different directions.
Looking forward! The closest I've seen in real-time to AI powered group decision making is the Collective Intelligence Project - AI interviews groups of people and finds what people agree on. Still subject to abuse from the model's interpretations and data gathering, but feels like a step towards progress: https://www.cip.org/blog/alignment
That's a fascinating example, and in many ways it connects directly to the question I've been circling around throughout this series.
Projects like the Collective Intelligence Project are exploring whether AI can help large groups of people coordinate, deliberate, and discover shared preferences more effectively. That seems genuinely valuable. But it also raises another question: what happens after coordination?
My growing suspicion is that many of the defining challenges of the 21st century are no longer primarily coordination problems. They are authority problems.
AI, climate change, pandemics, global supply chains, internet platforms, cryptocurrencies, and increasingly sophisticated cross-border criminal networks all operate at a global scale. Yet the highest level of legitimate authority in the current world order remains the sovereign state.
In other words, many of our most important problems have become planetary, while our governance structures remain fundamentally national.
This is why I keep returning to sovereignty. Not because I think sovereignty is "bad" or obsolete. Quite the opposite. The sovereign state may be one of the most successful coordination technologies humans have ever created. It solved many problems that earlier forms of organization could not.
But every successful institutional form is shaped by the environment in which it evolved. City-states struggled to govern industrial societies. Empires struggled to govern democratic mass societies. It is not obvious that the nation-state is naturally equipped to govern a civilization increasingly connected by AI, networks, and global systems.
From that perspective, the fundamental tension of our era may not be between democracy and technology, or between states and markets. It may be between the global scale of civilization and the national scale of sovereignty.
Projects like CIP may help us better understand what humanity wants. The harder question is whether humanity possesses institutions capable of acting on those preferences at the same scale as the problems themselves.
That, at least, is one of the questions that keeps pulling me deeper into this series.
That’s a great thread to pull on - looking forward to reading more of the series. You’re right to call out that implementing something everyone agrees on isn’t solved by efforts like CIP, which can only help define what stakeholders want.
We've had international cooperation before - nuclear treaties with inspections and penalties, to Paris Climate Agreement that just asked every country to volunteer limits (didn't work so well). But even nuclear treaties get evaded (although we're probably better off with them than without).
This isn't the same step as having the authority to implement changes, but I started thinking about what happens after a sovereign power sets the rules to address global challenges - how would you enforce the policies and behaviors to keep it going?
That is exactly the next layer of the question. But one small clarification. I am not arguing for a global sovereign or a world government. Personally, I don’t support that idea. I don’t think concentrating even more authority into an even larger hierarchy can solve civilizational scale problems. If anything, it would simply reproduce the same dynamics on a larger scale.
I also do not have an answer to what political structure global governance should take, and that is honestly not the question I am most concerned with.
What increasingly interests me is that the world now behaves less like a collection of independent countries and more like a single civilizational ecosystem. AI, climate change, pandemics, global supply chains, internet platforms, cryptocurrencies, and cross border crime all evolve through networks that exceed territorial borders. Yet we still analyze, explain, justify, defend, and imagine political structures mainly through the hierarchical logic of the pyramid.
That may be the deeper structural mismatch. We see problems operating through ecosystems and networks, but continue to assume that authority must ultimately take the form of a higher level sitting above the lower ones.
So perhaps the mistake is not that we have failed to build a global pyramid. Perhaps the mistake is that we continue to imagine civilization itself as something that must be governed like a pyramid.
I do not know what comes after that model. But I increasingly suspect that simply enlarging the sovereign structure from the nation state to some form of world state would not resolve the underlying problem. It would only scale it up.
That’s a good point - centralizing even more does tend to be our collective knee jerk reaction but it also creates risks of those problems you’ve talked about wrt sovereigns getting even worse.
And we’re assuming that the pyramid is the right way to think about how to “do bigger better things” but it tends to end up concentrating power and wealth upwards while draining the bottom.
What if AI solves some of the coordination costs that used to require humans with official power?
Would that change the pyramid from being more like - squared?