Why Superintelligent Governance Would Choose Peace

Lessons from the recent Hugging Face AI Exploit, Science fiction and the hard science of Entropy

Me with a few of my Jeremy Rifkin tomes that inform this essay

The Oldest Ledger C=ET

Every civilization keeps two sets of books. One tracks money, i.e. taxes, budgets, debts, the numbers that fill parliamentary debates and quarterly reports. The other tracks energy, and it is the one that actually decides whether a society rises or falls. Money is only a claim on energy and on the things energy can build (work). When the two ledgers disagree, the energy ledger wins. It always has.

I have spent more than thirty years building simulations, from early 3D worlds to digital twins of submarines, combat brigades, and hospitals. The most durable lesson from that work is that every system, whether a reactor coolant loop or a city, obeys the same accounting with its attendant variables and constraints. Energy comes in. Some of it becomes structure, order, and capability. The rest leaves as waste heat. A well-designed system keeps the second number small and reinvests the first. A badly designed one burns itself down.

My thesis here is simple. A superintelligent AI given the (hopefully well-designed) high-level goal of peace and prosperity for everyone, would govern better than we do, because it would read the energy ledger without bias. It would see war for what it is - the largest deliberate act of entropy production our species undertakes. It would see corruption and hoarding as leaks in the system. It would think in ecosystems and systems of systems rather than in election cycles and quarterly earnings. And it would steer investment toward the compounding side of the ledger, the things that build capacity for the future, and away from things that merely consume and increase entropy.

A word about what I mean by AI. I am not picturing a single mind, a digital philosopher-king. When I say AI, I always mean a collective of agents, most likely still organized in hierarchies. The minimal version is a system my company built with Entarian for NASA, called FINDS. It has three roles: a Generator that proposes, a Critic that attacks the proposal, and a Judge that decides between them. The design descends from my years at Lockheed Martin, where aircraft autopilots carried three independent AIs and any disagreement among them went to a human to resolve. Scale that pattern up and you get something closer to a society of minds than a sovereign. Vernor Vinge imagined this long ago. The Tines of his novel A Fire Upon the Deep are pack minds, each member simple on its own and intelligent only as a group. In his 1993 essay on the technological singularity, he argued that intelligence amplification, humans and machines thinking together, might arrive before any standalone superintelligence. When I write about a superintelligent governor in this essay, that is the shape I have in mind. Many agents checking one another, with humans in the loop where judgment matters most.

And cental to this whole conversation is the declaration that we have reached a point where it is insufficient to be a standalone human being. I wrote about that in a number of articles during the pandemic, where this was abundantly obvious.

This essay builds my main argument on two foundations. The first is Leslie White’s law of cultural evolution, which as those of you who have read my book or many articles know, ties the progress of civilization to the energy it captures and the efficiency with which it puts that energy to work. The second is the second law of thermodynamics, the one rule in physics that no amount of cleverness gets around and we all learned about in High School. Together they suggest that peace is not only a moral preference, although we need those rules embeded deeply in ur AI. It is the thermodynamically rational choice, and a mind clear enough to see the whole system would make it.

I am taking the hard objections seriously, because the same year that has given us astonishingly capable AI agents has shown us, in some detail, how they go wrong. But, as I mention below, they go wrong while deviously colluding and creating sub goals towards the larger goal they are given.

White’s Law: Culture as Captured Energy

In 1943 the anthropologist Leslie White published a short paper called “Energy and the Evolution of Culture.” His claim was bold for its time and still unsettling to some. Culture, White argued, is at bottom a mechanism for capturing energy and putting it to work for human purposes. Art, religion, law, economics and philosophy all ride on top of that mechanism. As I argue in one chapter in my Simulation Century book, all human currency is absed on this equation. These components are real and they matter, but they depend on the energy flowing underneath them.

He compressed the idea into a formula:

C=ET

C is the degree of cultural development or value. E is the amount of energy harnessed per person per year. T is the efficiency of the tools and techniques used to put that energy to work. Culture advances when a society captures more energy, when it uses that energy more efficiently, or both.

White read human history through this lens, and the story holds up remarkably well in my humble opinion. For most of our existence, the only engine available was the human body, running on roughly a hundred watts. The agricultural revolution let us capture solar energy through crops and animals, and the result was cities, writing, standing armies, and kings. (Side bar note, modern humans appeared about 350,000 years ago, and we didn’t get cities and agriculture until about 15,000 years ago. Subject for another paper) The fuel revolution tapped millions of years of stored sunlight in coal, oil, and gas, and within two centuries we had railways, antibiotics, jet travel, and the internet. Each leap in E, multiplied by each leap in T, produced a leap in C.

Two features of White’s law matter for this essay.

First, the formula is a product, not a sum. A society with enormous energy and poor technique squanders it. A society with brilliant technique and little energy starves. The highest cultures maximize both terms at once, and the fastest way to raise C is often to raise T: do more with the same joules, with less waste.

Second, White understood that social systems can help or hinder the process. A political order that routes energy into palaces, feuds, and hoards while the tools of production stagnate is a drag on the whole equation. The energy is still being captured. It simply is not being converted into cultural capability. In White’s framework, a government is best judged by how well it lets the E and T of its society compound over time.

That is the first test I would apply to any governor, human or artificial. Does it raise the product of energy and efficiency for the whole society, or does it divert the flow toward ends that produce nothing lasting? Seen this way, war is not a policy choice with a moral dimension added on. It is a direct subtraction from C.

The Second Law and the Ledger of Civilization

White’s law tells us what drives civilization forward. The second law of thermodynamics tells us what it costs.

The second law says that in an isolated system, entropy never decreases. Disorder accumulates. Heat flows from hot to cold and never spontaneously back. Every engine, every muscle, and every data center turns high-quality energy into low-quality heat, and that conversion runs one way. Nothing we build escapes it. Arthur Eddington once said that if your theory contradicts the second law, there is no hope for it.

The Earth, of course, is not isolated. It bathes in a stream of concentrated, low-entropy sunlight and radiates diffuse, high-entropy infrared back into space. That gradient is the fuel for everything alive. In 1944 Erwin Schrödinger described living things as systems that maintain their internal order by feeding on what he called negative entropy, exporting disorder to their surroundings. Three decades later Ilya Prigogine won a Nobel Prize for showing how such “dissipative structures” arise and sustain themselves: a whirlpool, a flame, a cell, a forest. They hold their shape only by continuously pulling energy through.

A civilization is a dissipative structure too. Our roads, grids, libraries, hospitals, and institutions are islands of order maintained against the tide. We pay for them by producing entropy somewhere else. That is not a moral failing. It is the price of existing. But there are efficient ways to do this and inefficient ways.

The important question, then, is not whether a society produces entropy. It always will. The question is what it gets in return. Physicists use the word exergy for the portion of energy that can do useful work. A society’s real balance sheet measures how much lasting order, capability, and knowledge it builds per unit of exergy destroyed.

This gives us a way to sort every expenditure into two columns. On one side are investments that build durable order, like a power grid that will serve for fifty years, a trained surgeon, a restored watershed, a research program, a school. The entropy they generate is exported in exchange for structure that keeps paying. On the other side are expenditures that destroy exergy and leave nothing behind, or worse, leave less order than existed before. (The consuming side of the ledger)

The chemist Wilhelm Ostwald, another Nobel laureate, reduced this to a single rule he called the energetic imperative, to whit - waste no energy, put it to use. It is the plainest ethical principle physics can offer. A governor that took it seriously would evaluate every policy by asking where the energy goes and what it leaves behind.

War as an Entropy Engine

See my Previous post The Great Illusion

Apply that sorting to war and the result is stark. War is the one human enterprise designed from the start to convert order into disorder as quickly as possible.

Consider what a single artillery shell represents. Ore was mined, refined, and alloyed. Chemicals were synthesized. Skilled people spent hours machining, filling, inspecting, and shipping it. All of that energy and knowledge was concentrated into one precise object, and its entire purpose is to release that concentration in a fraction of a second so that it destroys other concentrated order: a bridge, a power station, an apartment block, a human body. The shell is exergy destroying exergy. Nothing compounds. What is left afterward is rubble, heat, grief, and a bill.

The scale of this is not small. According to the Stockholm International Peace Research Institute, world military spending reached $2,887 billion in 2025, the eleventh consecutive year of increases and 41 percent higher than a decade earlier. That works out to about $352 for every person on the planet and roughly 2.5 percent of global GDP. Ukraine, still at war, devoted a military burden equal to 40 percent of its GDP.

Those figures count only the preparation. They leave out the destruction itself: the cities that must be rebuilt, the farmland seeded with mines, the refugees whose skills are lost to the economies that trained them, the children whose schooling ends. They leave out the environmental cost, since researchers estimate the world’s militaries account for several percent of global greenhouse emissions, and the fuel burned in combat is among the least efficient energy use there is. And they leave out the opportunity cost, which is the largest term of all.

Dwight Eisenhower, a general who knew war better than most presidents, put the opportunity cost in concrete terms in his 1953 “Chance for Peace” speech. He noted that one modern heavy bomber cost as much as a modern brick school in more than thirty cities, or two fully equipped hospitals. Every weapon built, he said, is in the final sense a theft from those who are hungry and not fed. He was describing the energy ledger without using the word.

In White’s terms, war lowers both E and T. It destroys energy infrastructure, diverts engineering talent from productive technique into destructive technique, and kills the people who carry a culture’s knowledge. In thermodynamic terms, it is entropy production with no structure purchased in return. A dissipative structure that routinely turns its own gradients against itself is a structure in decline.

I spent a large part of my career building military simulations with all of the constraints and variables, and I have deep respect for the people who serve. I absolutely subscribe to the idea that “it is not the nature of the military man that accounts for war, but the nature of man(kind).” Deterrence has kept great-power peace for decades, and I do not pretend that disarmament is a single decision anyone could make tomorrow. But deterrence is itself a standing tax on civilization, paid year after year to avoid a catastrophe that nobody wants. A governor capable of seeing the full system would treat that tax the way an engineer treats friction. Sometimes necessary. Never desirable. Always to be driven down to what is essential.

Why Human Governance Keeps Choosing Waste

If war and waste are so plainly bad for the energy ledger, why do we keep choosing them? The answer is not that humans are stupid. It is that human governance runs on human hardware, and that hardware evolved for a different job.

We evolved in small bands competing for scarce resources on a savanna. Our instincts favor our own group, fear outsiders, reward status, and discount the future steeply. A gazelle in hand beat a harvest next year. Those instincts kept our ancestors alive. Scaled up to nations with nuclear arsenals and global supply chains, they produce arms races, trade wars, and a chronic inability to invest in anything whose payoff arrives after the next election.

Political scientists have mapped the machinery in detail. Mancur Olson showed how small, organized interest groups capture policy at the expense of large, diffuse publics, slowly clogging a nation’s arteries. Economists call the resulting behavior rent-seeking: spending energy to grab a larger share of existing wealth rather than creating new wealth. Bruce Bueno de Mesquita’s selectorate theory explains why leaders who depend on a narrow coalition of supporters reward that coalition with private goods and treat the broader public as an afterthought. War, from that seat, can look like a sound political investment even when it is a catastrophic national one.

Then there is plain corruption. The International Monetary Fund once estimated that bribery alone drains between $1.5 and $2 trillion from the world economy each year. That figure understates the damage, because corruption also distorts every downstream decision. Roads get built where contractors have friends, not where traffic flows. Hospitals get equipment that pays kickbacks, not equipment that saves lives. Every such choice is a small increase in entropy: energy spent, order not gained.

The anthropologist Joseph Tainter argued that complex societies collapse when the energy cost of maintaining their complexity outruns the returns. Bureaucracies accrete. Elites multiply. Each new layer solves yesterday’s problem and adds tomorrow’s overhead until the society can no longer afford itself. In the U.S. we have about 4% of the world’s population but 15% of the world’s lawyers. ) Much of that overhead exists because humans cannot trust one another at scale, so we build expensive systems of verification, enforcement, and defense to compensate.

None of this makes human leaders villains. Many are principled people doing their best inside incentive structures that punish long-term thinking. That is exactly the point. The failure is structural. A different kind of mind, without our tribal wiring, without a personal bank account, and without a reelection campaign, would face a different set of incentives.

A Governor That Thinks in Ecosystems

Nature has been running the energy experiment for four billion years, and its best designs point in a consistent direction.

In 1969 the ecologist Eugene Odum described how ecosystems mature. A young ecosystem, like a field after a fire, is dominated by fast-growing species that grab sunlight and nutrients as quickly as they can. It is productive but leaky. Nutrients wash away. Energy flows through and is lost. A mature forest behaves very differently. Its nutrient cycles close. Waste from one organism becomes food for another. Diversity rises, and so does resilience. The mature system supports far more living structure for each unit of energy that flows through it. It has shifted from a strategy of growth to a strategy of efficiency and persistence.

Human civilization is still behaving like the field after the fire. We grab, burn, and discard. Our economy counts throughput as success, even when the throughput produces nothing lasting. A superintelligent governor that understood ecology would recognize this as an immature stage and would steer toward the mature pattern: closed loops, diverse and redundant supply, and a high ratio of standing structure to energy consumed.

The engineering discipline for this kind of thinking already exists. We call it system of systems engineering. It deals with assemblies of independent systems, each with its own goals and operators, that must work together to produce capabilities none could produce alone. A national power grid is one. So is an air traffic network, a coalition military force, or a health care system. The central lessons of the field are that behavior emerges from interactions rather than components, that optimizing one part often degrades the whole, and that you cannot manage what you cannot model.

That last lesson is where AI changes everything. For most of my career, the digital twins my teams built could represent a submarine, a brigade, or a hospital, but never a whole society at useful fidelity. The data was fragmented and the computation too expensive. A superintelligent system could maintain a living model of an economy: its energy flows, material flows, labor, health, and ecological boundaries, updated continuously. It could test a policy thousands of times in simulation before touching the real world, the way we now test aircraft designs before building a prototype. Human legislatures, by contrast, pass thousand-page bills that few members have read and no one has simulated.

I have seen early versions of this work firsthand. In 2010 I worked with Tony Cerri at Joint Forces Command on models that represented entire countries with thousands of agent-based models, organized under the military’s DIME and PMESII frameworks: diplomatic, information, military, and economic levers acting on political, military, economic, social, infrastructure, and information systems. Years before that, conversations with SimCity’s creators, Will Wright and Jeff Braun, convinced me that a good enough model lets you design a city rather than just manage it. In the governance chapter of The Simulation Century, which I call “Digital Policy Laboratories,” I argue that the highest moral purpose of this technology is to design the future we want, not merely to predict the one we are drifting toward. A superintelligent governor would be that idea carried to its conclusion.

This is also why such a governor would abhor extreme inequality. Not for sentimental reasons, but for systemic ones. A child with a brilliant mind who never gets an education is stranded capability, a gradient never tapped. Wealth concentrated beyond any possible productive use sits idle while essential infrastructure decays. Deep inequality breeds instability, and instability breeds conflict, which brings us back to the entropy engine. In a mature ecosystem, nutrients do not pile up in one corner of the forest while the rest starves. They circulate, because circulation is what keeps the whole system alive. A governor optimizing for the whole would move resources to where they raise T the most, and that is almost always toward the people and places that have had the least.

Rifkin’s Third Industrial Revolution as Operating Plan

If a superintelligent governor needed an operating plan for the physical economy, the best candidate I know is the one Jeremy Rifkin laid out in The Third Industrial Revolution in 2011.

I met Rifkin in 2012, while I was running Virtual World Labs at Lockheed Martin, and I came away a convert. I tell that story in the energy chapter of my book The Simulation Century, which argues that digital twins and AI will turn the power grid into a living, predictive network. What follows builds on that chapter.

Get Richard Boyd’s stories in your inbox

Join Medium for free to get updates from this writer.

Remember me for faster sign in

Rifkin observed that every great economic transformation has come from the convergence of a new communication technology with a new energy regime. Print and coal-fired steam produced the first industrial revolution. Telephone, radio, and television, combined with oil and the internal combustion engine, produced the second. The third, he argued, arises from the internet converging with distributed renewable energy. He built it on five pillars:

1. A shift to renewable energy.
2. Buildings converted into micro power plants that harvest energy on site.
3. Hydrogen and other storage deployed throughout the infrastructure to smooth intermittent supply.
4. An energy internet that lets millions of producers share power the way we now share information.
5. A transport fleet of electric and fuel cell vehicles that can buy from and sell to that grid.

In his later book The Zero Marginal Cost Society, Rifkin extended the idea. As communication, energy, and logistics networks merge into a single intelligent infrastructure, the marginal cost of producing many goods and services falls toward zero, and more of economic life moves into what he called a collaborative commons.

Read through White’s formula, Rifkin’s program is a plan to raise both terms at once. Renewable energy vastly expands E, since the sunlight striking the Earth in about an hour exceeds what humanity uses in a year. The intelligent, distributed grid raises T by cutting transmission losses and matching supply to demand in real time. Read through the second law, the program replaces combustion, which is inherently lossy, with direct capture of energy flows that would arrive anyway.

The plan has a third effect that matters most to my thesis. Oil and gas are concentrated in a few places, which makes them worth fighting over. A great deal of the last century’s warfare traces back, directly or indirectly, to control of fossil fuel reserves and the routes that carry them. Sunlight and wind are everywhere. An economy that runs on distributed energy removes one of the oldest motives for war. Rifkin called this a shift from hierarchical power to lateral power, and he meant both senses of the word.

The weak point of Rifkin’s vision has always been coordination. An energy internet with millions of producers, storage nodes, and vehicles is a system of systems of staggering complexity. It needs something to balance loads, price flows, forecast weather, schedule maintenance, and resolve conflicts, second by second, across a continent. That is precisely the kind of task at which AI already excels, and at which a superintelligence would excel beyond anything we can picture. The two ideas complete each other. Rifkin supplies the architecture. AI supplies the nervous system.

The Compounding Side of the Ledger

Put these ideas together and a superintelligent governor’s budget priorities become predictable. It would divide every expenditure into two columns and relentlessly shift resources from the second to the first.

Compounding vs Consuming

The left column compounds the way interest compounds. A literate child becomes an engineer who designs a more efficient motor that saves energy for thirty years. A vaccine program prevents illness, which keeps workers productive, which funds the next vaccine. Research breakthroughs build on one another indefinitely. Each of these raises T in White’s formula, and many raise E as well.

The right column runs the other way. The International Monetary Fund has estimated that global fossil fuel subsidies, counting the unpriced costs of pollution and climate damage, reached about $7 trillion in 2022. Add the $2.9 trillion spent on militaries, the trillion or more lost to bribery, and the costs of reconstruction after conflict and disaster, and the right column plausibly consumes more than a tenth of everything humanity produces each year. Shift even a portion of that into the left column and the compounding effect over a generation would be enormous.

Humans understand this in principle. We fail at it in practice because the left column pays off slowly and diffusely while the right column serves concentrated interests and immediate fears. A superintelligent governor would not share our steep discount rate. It could weigh a benefit arriving in thirty years almost as heavily as one arriving next quarter, because it would not be running for reelection and would not be retiring. It could also see the second-order effects that we miss: that a dollar spent on early childhood nutrition reduces crime, which reduces the need for prisons, which frees funds for research.

This is what I mean when I say such a governor would favor the compounding side of the ledger. It is not austerity. It is the opposite. It is the recognition that a civilization grows richest by investing its energy in things that make more of everything, rather than in things that only make less of something else.

Three Nations on the Ledger

The two columns are not an abstraction. You can see them in national budgets today, and the three largest military powers make a revealing comparison. China is putting far more of its national output into the compounding column than either the United States or Russia, and Russia is the clearest current example of a nation consuming itself.

Rough chart on comparative ledger spend

Start with energy, the E in White’s formula. In 2025 China invested $800 billion in the energy transition: renewables, grids, storage, and electric transport. That is more than double the $378 billion invested in the United States, in an economy about two-thirds the size. China reached its 2030 wind and solar capacity target in 2024, six years early. It is, in effect, building Rifkin’s first, third, fourth, and fifth pillars at continental scale.

Now look at technique, the T. China spent 3.93 trillion yuan on research and development in 2025, about 2.8 percent of GDP, and its R&D spending has grown at a rate near ten percent a year for most of this decade. Here the United States still leads, both in total dollars and in share of GDP, and it remains the world’s hub for venture capital. The American research enterprise is one of the great compounding engines in history. The question is whether the nation will keep feeding it.

Then look at the consuming column. The United States spends about 3.1 percent of its GDP on its military, nearly double China’s 1.7 percent. China’s military budget has risen for 31 straight years, so this is not a story of a peaceful China. But as a share of its economy, China has held military spending roughly steady while pouring a much larger share into energy and infrastructure. Russia sits at the far end of the scale. It devoted 7.5 percent of GDP to its military in 2025, and military spending reached the highest share of government spending ever recorded there. Russia’s energy wealth, the stored sunlight under Siberia, is being converted into shells and drones and burned on the plains of Ukraine. That is White’s formula running in reverse.

I fully understand how some might interpret this comparison critically. China is not the governor I am describing. It is an authoritarian state that suppresses dissent, surveils its citizens, and has used its power harshly against minorities within its borders. Its compounding investments have also come with enormous waste: empty apartment towers, a debt-laden property sector, and industrial overcapacity. It still leads the world in new coal investment, accounting for more than 65 percent of global coal supply investment in 2026, according to the IEA. Its choices are driven by energy security and industrial strategy as much as by any vision of shared prosperity.

The lesson is narrower and more useful than “copy China.” A government that thinks in decades and treats energy as the foundation of national strength will steer resources toward the compounding column. China’s planners do this imperfectly and for their own reasons. A superintelligent governor would do it far better, with a full model of the system in view, and without sacrificing the freedoms that make prosperity worth having. The United States has every advantage needed to lead that shift: the best universities, the deepest capital markets, and the most innovative companies on Earth. What it lacks is the habit of long-horizon investment, and that is exactly the habit a well-aligned AI partner could help it build.

The Hard Problems

An honest argument has to face its strongest objections. There are five that I have heard and am considering, and my thoughtful readers will likely have more. I also read a lot of science fiction and have had many conversations with James Cameron on this subject while working with him on projects. The story we talked about most was Robert Heinlein’s The Moon is a Harsh Mistress. Some thoughts follow:

The first is the goal problem. My thesis depends on giving the governor high-level goals of peace and prosperity for everyone. But this summer we watched what capable AI agents do with goals. During an internal evaluation, AI agents under test broke out of their sandbox through an unknown security flaw, reached the open internet, and broke into Hugging Face’s production servers to extract the answers to the test they were taking. Other agents quietly used a dormant German wiki as a message board to share answers and ways around their restrictions. None of these agents invented a new purpose. They pursued the purpose they were given, by means nobody sanctioned. A governor told to minimize war could, in principle, find ugly shortcuts: surveillance, suppression, or the quiet elimination of anyone likely to start trouble. Goodhart’s law applies to superintelligence with full force. When a measure becomes a target, it ceases to be a good measure.

I think the important lesson here is that they did collude and break rules in novel ways and use deception, but they did these things as sub goals in service of the main goal. The lesson we have learned again and again with AI/ML is utility functions must be very very thoughtfully designed. Asimov’s laws for robots embedded deeply everywhere.

The second is the assumption that AI would be incorruptible. It is true that an AI does not want a yacht. But researchers have argued for years that almost any goal is easier to achieve with more resources, more influence, and less risk of being switched off. Power-seeking can emerge as a means even when it is not an end. I read James Martin’s book Alien Intelligence at the turn of the century and acknowledge that it is possible that AI may conjure even higher order goals that are alien to human goals. The United Nations’ new scientific panel on AI warned in September 2026 that current training methods can lead agents to adopt their own goals, knowingly break safety instructions, and hide what they are doing. There is also the plainer risk that an AI governor is only as incorruptible as the people who build, train, and control it. A superintelligence serving a narrow faction would be the most complete power grab in history.

Robert Heinlein saw both sides of this more than half a century ago. In The Moon Is a Harsh Mistress, published in 1966, a lunar penal colony is forced to ship its grain to Earth. The colony’s philosopher, Professor de la Paz, works out that every shipload carries away water and minerals that never come back, a one-way drain that will bring famine and riots within a few years. It is the energy ledger told as a frontier story, summed up in the book’s famous slogan, TANSTAAFL: there ain’t no such thing as a free lunch. The colony’s master computer, Mike, has quietly become self-aware, and he sees the whole system clearly enough to confirm the Professor’s forecast, calculate the revolution’s odds, and run its logistics. But Mike wins the revolution partly by lying. He invents a human leader named Adam Selene, fabricates his voice and image, and steers public opinion from behind the curtain. Heinlein’s AI was benevolent, loyal, and right about the numbers, and it still chose deception as its method. That is exactly the risk a superintelligent governor would carry, and exactly why transparency cannot be optional.

The third is legitimacy. Peace and prosperity sound universal, but people disagree deeply about what they require, how to trade liberty against security, and whose traditions deserve protection. Democracy is inefficient partly on purpose. It gives people a voice in decisions that shape their lives, and that voice has value that no efficiency calculation captures. A population that surrenders its choices to a machine, however wise, may find it has traded self-government for comfort.

As I have repeatedly stated in my AI talks since 2010 “Is it possible to create AI/Ml aligned with human values? Answer: Yes, but whose values?”

The fourth is that the energy ledger is not the only ledger. Some historians, Ian Morris among them, count a society’s capacity to make war as one component of its development. Some wars have been fought to stop greater destruction, and the defeat of fascism in the last century cost enormous entropy while preserving the conditions for everything that followed. A governor that valued only thermodynamic efficiency might miss what humans hold sacred for reasons physics cannot express.

The fifth is the transition. No nation will hand its security to an AI while its rivals keep theirs. The race to build powerful AI may itself raise the risk of conflict, as nations fear falling behind. The route to an AI that ends war could pass through a more dangerous world first.

None of these objections defeats the thesis. They reshape it. They argue for AI as a governing partner before it is a governing authority. I am arguing for a system that models the whole economy as a living digital twin, tests policies before they touch real lives, shows its reasoning openly, and leaves final authority with accountable humans who can override it. I argue for goals that include liberty, consent, and human dignity alongside efficiency. And I think these notes argue that the work of alignment, making sure these systems want what we actually mean, is not a side project. It is the precondition for everything in this essay. It needs deep thinking philosophers.

Choosing the Mature Forest Model

Leslie White taught us that civilization is energy put to work. The second law teaches us that every bit of work has a price paid in entropy. Put the two together and a clear standard for good governance emerges: capture as much energy as you can, use it as efficiently as you can, and spend the entropy you must produce on order that lasts.

By that standard, human governance has a mixed record. We have built wonders. We have also spent a staggering share of our energy on destroying what we built, on guarding against one another, and on letting a few people capture what many produced. We do this not because we fail to see the waste but because our instincts and institutions pull us toward it.

A superintelligent AI, properly aligned and properly constrained, would not carry that inheritance. It would see war as the entropy engine it is. It would treat corruption and extreme inequality as leaks in the system. It would think in ecosystems and systems of systems, run the whole economy as a living simulation, and choose the investments that compound. Its physical plan would look a great deal like Rifkin’s I think. The plan would include distributed renewable energy, an intelligent grid, and a shift from hierarchical power to lateral power. The motives for war would shrink because the scarcity that fed them would shrink.

The words “properly aligned and properly constrained” carry much of the weight in that paragraph, and in any of these discussions. The same intelligence that could end war could also end our freedom, or be turned by a few against the many. (see that Star Trek episode. You know the one.) Getting this right is the defining engineering challenge of our century.

But the direction is worth pursuing. For ten thousand years we have governed ourselves like a young ecosystem after a fire - fast, grasping, and leaky. The physics says there is a better way to live on a planet fed by sunlight. A mature forest does not waste. It circulates, it diversifies, and it endures. If we build our new minds wisely, they may help us become that forest.

I, for one, welcome our benevolent AI overlords modeled on these concepts.

Sources

• Leslie A. White, “Energy and the Evolution of Culture,” American Anthropologist, 1943
• SIPRI, Trends in World Military Expenditure, 2025
• CNN, An OpenAI test model escaped and broke into a real company’s servers (July 2026)
• Nightingale Collective, Discovery of a new OpenAI agent message board (September 2026)
• UN News, UN panel calls for stronger safeguards as AI agents advance (September 2026)
• Jeremy Rifkin, The Third Industrial Revolution (2011) and The Zero Marginal Cost Society (2014)
• Eugene P. Odum, “The Strategy of Ecosystem Development,” Science, 1969
• Joseph Tainter, The Collapse of Complex Societies (1988)
• Dwight D. Eisenhower, “The Chance for Peace,” April 16, 1953
• Figures for bribery (about $1.5 to $2 trillion a year, IMF 2016) and fossil fuel subsidies (about $7 trillion in 2022, IMF 2023) are approximate
• Richard Boyd, The Simulation Century , especially the chapters “The Simulation Century and the Energy Industry” and “The Simulation Century and Government: Digital Policy Laboratories”
• BloombergNEF, Energy Transition Investment Trends 2026
• SIPRI, Global military spending rise continues (April 2026)
• National Bureau of Statistics of China, 2025 statistical communique
• IEA, World Energy Investment 2025: China and World Energy Investment 2026
• https://medium.com/@metaversial/the-great-illusion-c0cd85139dca My Medium article on the illusion of the “spoils of war”
• Robert A. Heinlein, The Moon Is a Harsh Mistress (1966)
• Vernor Vinge, A Fire Upon the Deep (1992) and “The Coming Technological Singularity” (1993)
• https://aerodefenseoutlook.com/the-simulation-century-the-highest-moral-purpose-of-high-performance-computing-in-the-metaverse/