Power and Intelligence

The world is in chaos.

Wars on every screen, and children among the dead. Markets that profit from the panic. Governments that answer to whoever pays for them. A new shock every morning, before the last one has landed.

Shock is not a side effect. It is a method. A population that is frightened, tired and overwhelmed does not organize. It survives.

And yet something has changed. More people than ever can see what is happening. The footage is on every phone. The documents leak. The pattern is no longer hidden from anyone who looks.

But seeing is not the same as acting. We can watch all of it and still feel powerless, because the people who hold power coordinate their capital, their expertise and their attention far better than the rest of us ever have.

This essay is about that gap: why it exists, why it has held for so long, and why, for the first time, it might close.

A population in survival mode does not organize.

Juvenal had a phrase for it nearly two thousand years ago: bread and circuses.[1] Keep people fed just enough and entertained just enough, and they will hand over the business of power to someone else.

The modern version is better engineered. Rent that takes more than 30% of income from nearly half of American renters, a record 22.7 million households.[2] Feeds tuned to hold attention for as long as possible. Cheap distraction, expensive necessity, and a steady supply of things to be afraid of. None of it needs a secret plan; each piece is profitable on its own. Together they produce the same result: people too busy, too tired and too numb to do the slow work of organizing.

The machinery of attention was never hidden. In 1928 Edward Bernays, the father of public relations, opened his book Propaganda with it: “The conscious and intelligent manipulation of the organized habits and opinions of the masses is an important element in democratic society.”[3]

Survival mode is the opposite of strategy. You cannot plan ten years ahead while planning for next month’s rent. The people who can plan ten years ahead are, almost by definition, the people who are not in survival mode, and they are planning.

Power is mass, times energy, times coordination.

In any contest between groups, the side that wins is usually the side with the most of three things, multiplied together.

Mass is the number of people, and how cohesive they are. Energy is what they bring: the motivation in their heads and the resources in their hands. Coordination is whether they move together, consciously by plan and unconsciously by shared instinct and timing.

The weights are not equal. Coordination counts most, then energy, then mass. A small group that moves as one beats a large group that moves as many. Military theorists have known a version of this for a century: in 1916 Frederick Lanchester showed that when forces can concentrate their fire, fighting strength grows with the square of the numbers brought to bear at once.[4] What matters is not how many you have, but how many act together, at the same moment, on the same point.

The few are not smarter. They are better organized.

In 1965 the economist Mancur Olson asked why small groups so often beat large ones in politics.[5] His answer was arithmetic: “unless the number of individuals in a group is quite small, or unless there is coercion or some other special device to make individuals act in their common interest, rational, self-interested individuals will not act to achieve their common or group interests.” When a few people share a large stake, each has a reason to show up, pay and organize. When millions share a stake, each one’s slice is small, and it is rational to let someone else do the work. The many free-ride; the few coordinate.

This is the quiet engine behind much of what feels unjust. A rule that costs every household a few dollars a year and earns one industry billions will have a lobby on one side and almost nobody on the other. A handful of families with a shared interest can hire the same lawyers, sit on the same boards, fund the same campaigns and speak with one voice. Eight billion people with a shared interest cannot yet agree on a time to meet.

And coordination compounds. Wealth buys expertise; expertise protects wealth. In the United States the top one percent of households held about 31% of the wealth at the end of 2024, and 32.5% by the middle of 2026.[6] That is not a fact about intelligence. It is a fact about organization.

The climb to power rewards the player willing to cheat.

There is a darker side to the same arithmetic. In any game among many players there is a cooperative outcome, where most do well, and a defector’s outcome, where one player takes more by breaking the rules everyone else keeps. When the referee is weak, the defector wins.

So the climb to power has a selection effect. It tends to favour people with fewer moral lines, not because everyone at the top is a monster, but because those without limits have more moves available. In one study of 203 corporate managers chosen for leadership programmes, 3.9% scored in the range researchers use to identify psychopathy; on a screening measure the rate was fifteen times that of a community sample.[7] The sample is small and the measure is debated, but the direction is the one you would expect. Those who would refuse are filtered out on the way up. What gathers at the top is overrepresented by those who play the game of power best, and their idea of justice is often simple: the winners do what they want, and write the history afterwards.

That is not a reason for despair. It is a design problem. Defection pays when it is invisible and unpunished. Make it visible, make the record permanent, and the payoff changes.

The game both sides lose by playing well.

In early 1950 two mathematicians at the RAND Corporation, Merrill Flood and Melvin Dresher, set two colleagues a simple game and had them play it a hundred times. That spring Albert Tucker dressed it in a story for a lecture: two prisoners, held apart, each offered a deal to betray the other.[8] Whatever the other does, each does better by betraying. Both end up worse off than if both had kept quiet. It became known as the prisoner’s dilemma.

The setting was not academic. On 29 August 1949 the Soviet Union had tested its first atomic bomb, and American aircraft collected the evidence days later.[9] Two powers now faced the dilemma at its most dangerous: each was safer with more weapons whatever the other did, and both were poorer and less safe as a result. John von Neumann, a founder of game theory, drew the coldest conclusion: “If you say why not bomb them tomorrow, I say why not today? If you say today at 5 o’clock, I say why not 1 o’clock?”[10]

The same structure runs through the questions this essay cares about. Two countries racing to build AI first, each cutting safety to keep up. Two blocs hoarding chips and minerals. Two utilities each refusing to share a transmission line. Played once, the dilemma rewards betrayal. The interesting question is what happens when it is played again.

Cooperation grows where the game repeats.

In 1980 the political scientist Robert Axelrod invited experts to submit strategies for a repeated prisoner’s dilemma and played them against each other. Fourteen entries came in for the first tournament and sixty-two for the second. Both were won by the same program, Anatol Rapoport’s TIT FOR TAT, which cooperates on the first move and then copies whatever the other side did last. “Amazingly enough,” Axelrod wrote, “the winner was the simplest of all candidates submitted.”[11]

He named four qualities behind its success: “avoidance of unnecessary conflict by cooperating as long as the other player does, provocability in the face of an uncalled-for defection by the other, forgiveness after responding to a provocation, and clarity of behavior so that the other player can recognize and adapt to your pattern of action.” And one condition above all: “For cooperation to prove stable, the future must have a sufficiently large shadow.”

The proof came from the worst place imaginable. In the trenches of the First World War, units that faced each other for months fell into what soldiers called live and let live: restraint, for as long as the other side showed restraint too. As one British officer reasoned, “if you prevent your enemy from drawing his rations, his remedy is simple: he will prevent you from drawing yours.” Breaches were punished, often two or three times over, and then the restraint resumed.[12] At Christmas 1914 tens of thousands of men left the trenches to meet in no man’s land.[13] Their commanders, playing a different game, ended it.

The lesson is the one this essay keeps returning to. “The foundation of cooperation,” Axelrod concluded, “is not really trust, but the durability of the relationship.” Make the relationship long and the record visible, and cooperation becomes the winning strategy.

Finding nothing is information.

Game theory says what the players want. Probability says what to believe while you do not yet know. The clearest lesson comes from the sea. When someone is lost, searchers do not sweep at random. They list the ways the loss could have happened, turn each into a patch of probability on a map, and search where the chance of the person being there, times the chance of spotting them, is highest.[14]

Then comes the step most of us skip. An empty sweep is evidence. It lowers the odds where you looked and raises them everywhere else. The US Navy found the submarine Scorpion this way in 1968, 260 yards from the likeliest cell on its map. Air France 447 was found in 2011, in about a week, once analysts rebuilt the map and admitted that an earlier search had marked as clear an area it could not really have cleared.[15]

Public life mostly does the opposite. We hold a view, find nothing that confirms it, and keep the view. A truth-seeking system updates on what it does not find as well as what it does, and keeps its map where others can check it. Parallel Futures applies the method to where AI, power and the economy are heading.

Seeing, doing, imagining.

Knowing that two things move together is not knowing why. The computer scientist Judea Pearl describes a ladder of causation with three rungs, association, intervention and counterfactuals, or in his words “seeing, doing, and imagining.”[16] On the first rung you notice that two things rise together. On the second you ask what happens if you change one of them. On the third you ask what would have happened had you acted differently.

“Intervention ranks higher than Association,” Pearl writes, “because it involves not just seeing what is, but changing what is.” Most arguments about power are stuck on the first rung: this group rose while that one fell, so one must have caused the other. Most propaganda lives there too. The questions that lead to better decisions are higher up: if we changed this rule, who would gain? Had that law passed, where would we be now?

Pearl adds a warning for the decade ahead: “Most animals as well as present-day learning machines are on the first rung, learning from association.” Machines that only see patterns will reproduce the world’s patterns, its injustices included. Systems that can reason about interventions can help change them.

A system is perfectly designed for the results it gets.

The systems thinker David Peter Stroh likes to say that every system is perfectly designed to achieve what it is achieving right now.[17] The cybernetician Stafford Beer put it more sharply: the purpose of a system is what it does. If the outcome is concentrated power and a confused public, that is what the arrangement is for, whatever its mission statements say.

So where do you push? Donella Meadows ranked the places to intervene in a system, from weakest to strongest.[18] Near the bottom are the numbers: a tax rate, a subsidy, a budget line, the things politics argues about most. Near the top are the structure of information flows (who does and does not have access to information), the rules, the goals of the system, and the paradigm it grows from. “Missing feedback is one of the most common causes of system malfunction,” she wrote.

That is the lever this essay keeps returning to. Change who can see what, and when, and the behaviour of the swarm changes without anyone being shamed into it.

Power decides which questions are safe to ask.

For most of history, the explanation of the world was whatever served the people who ruled it. Myths in which the gods looked after the king. An earth at the centre of everything, with the church at the centre of the earth. In 1633 Galileo was tried by the Inquisition for arguing that the earth moves around the sun, and he spent the rest of his life under house arrest.[19] The trouble with his astronomy was never the astronomy. It was that the question itself threatened the order.

The physicist David Deutsch calls what changed this the beginning of infinity: the moment a culture stops protecting its explanations and starts testing them, preferring good explanations, hard to vary, to stories that can explain anything.[20] Everything we call progress followed from that one habit.

The habit is not secure. Power still decides which questions are safe. It took two decades, a lenient plea deal, a death in a federal jail cell and finally an act of Congress before the Epstein files began to come out,[21] and what came out was only what was forced out. Wherever exposure would cost someone powerful, there is an incentive to hide the truth, and a market for the misinformation that hides it.

From inside, a bubble is simply the world.

Looking back, it is easy to judge. We read about people who went along with a lie, a mania or a cruelty, and we are sure we would have acted differently.

Most of us would not have. In Solomon Asch’s experiments in the 1950s, people asked to compare the lengths of lines gave an obviously wrong answer to match a room of actors about a third of the time, and three in four went along with the room at least once.[22] The right answer was in front of them.

Every era has its bubbles: of price, of belief, of status. Inside one, the incentives are local and immediate, the cost of breaking with the group is personal, and the reward for being right is distant. Behaviour inside a bubble is not mainly a failure of character. It is emergent: the predictable result of many people each making the locally sensible move.

That is why shaming does not work. You cannot scold a swarm out of its shape. You can only change what each member sees.

Ideas spread by status, not by truth.

The old fight was between propaganda and science: a central authority pushing a story, and a method pushing back. The new terrain is stranger. There is no single author. Ideas move through a swarm of people, carried by culture and status, and they take hold whether or not they are true.

When researchers traced eleven years of rumour cascades on Twitter, they found that false stories were seventy percent more likely to be shared than true ones, and that the truth took about six times as long to reach the same number of people.[23] Novelty and outrage travel. Corrections walk.

In that terrain, power does not need to censor. It only needs to flood. When everything is contested and everyone is exhausted, the best-funded story wins by default.

Or it only needs to frame. “The smart way to keep people passive and obedient,” Noam Chomsky observed, “is to strictly limit the spectrum of acceptable opinion, but allow very lively debate within that spectrum.”[24] A decade earlier he and Edward Herman had described the filters news passes through before it reaches us: who owns the outlet, who advertises in it, who it relies on for sources, and who punishes it for straying.

The feed does not only distract. It predicts.

Herbert Simon saw the economics coming in 1971: “a wealth of information creates a poverty of attention.”[25] What he could not have seen was the business built on that poverty.

Shoshana Zuboff calls it surveillance capitalism: an order that “claims human experience as free raw material for translation into behavioral data,” sold as predictions of what we will do next.[26] The most valuable thing a platform makes is not the content. It is the forecast of your behaviour, and the ability to nudge it. “It is no longer enough to automate information flows about us,” she writes; “the goal now is to automate us.”

This is the coordination gap in its newest form. The platforms see the whole swarm, in real time, and can steer it. The swarm sees only itself.

The story of one hidden cabal is a trap.

There is a trap waiting for anyone who sees all of this clearly. It is the story that a single secret circle runs everything: that every war, every crash and every crime is one plan, drawn up by one set of families.

The story feels like clarity. It is the opposite. It makes the enemy all-powerful and impossible to disprove, so nothing we do could matter. It buries the documented crimes, which are real, named and prosecutable, under claims no one can check. And for more than a century it has been used to point the anger of the many at a scapegoat instead of a structure.

The truth is less cinematic and more useful. Most of what the powerful do is done in the open: legal, coordinated and recorded. Even the “invisible government” was named in public, by its own practitioner: Bernays wrote in 1928 that those who manipulate opinion “constitute an invisible government which is the true ruling power of our country,” and sold the book in shops.[3] The arrangement does not need a secret meeting. It needs aligned incentives, better organization and our exhaustion. That is good news. A plot can only be stopped by its plotters. A structure can be changed by anyone who understands it.

Anger is fuel. It needs accuracy to become power.

Intelligence is the cost of coordination, and it is falling.

Go back to the formula. The many have always had the mass. More and more, they have the energy. What they have lacked is coordination, because coordination was expensive. It took organizers, lawyers, analysts, institutions and time. The few could afford it. The many could not.

Intelligence is the cost of coordination, and intelligence is becoming nearly free.

That cuts both ways. Concentrated in a few labs and a few platforms, cheap intelligence makes the few better coordinated still: every lobby with a thousand tireless analysts, every feed tuned to every person. Kai-Fu Lee, who built AI businesses in both superpowers, put the default plainly: “The AI world order will combine winner-take-all economics with an unprecedented concentration of wealth in the hands of a few companies in China and the United States.”[27] Spread across the many, it does what unions, co-operatives and open source did at their best, only faster. It lets people who have never met find each other, pool what they know and move together.

Which of those happens is not a technical question. It is a question of who owns the models, the data and the record.

Make the record the referee.

Truth has a coordination problem too. Friedrich Hayek saw it in 1945: the knowledge a society needs “never exists in concentrated or integrated form, but solely as the dispersed bits of incomplete and frequently contradictory knowledge which all the separate individuals possess.”[28] The people holding the pieces rarely meet.

Elinor Ostrom won a Nobel prize for showing that communities can govern a shared resource well without a ruler or a market, when the rules are clear, the monitoring is shared and the members help write the rules.[29] In Philip Tetlock’s forecasting tournaments, volunteers who kept score of their predictions reportedly did about 30% better than intelligence analysts with access to classified data.[30] In 2015 Taiwan used an open tool, Pol.is, to let thousands of citizens map where they agreed on regulating Uber; the statements most of them could accept became the basis of the rules.[31]

Put those together and you have the outline of something new: data-driven governance that belongs to the governed. Open models of the systems we live inside, energy, food, money, housing. Claims that carry their sources. Predictions that are scored when they resolve. A permanent public record of who said what, and whether it turned out to be true.

In that world the defector loses the oldest advantage there is, the fact that nobody can check. Misinformation has to compete with a record. Being right becomes the credential, and it is one anyone can earn.

Mass and energy we have always had.

This is the turn. For the first time, the tools that made the few powerful, expertise on demand, analysis at scale, a memory that never forgets, can belong to everyone.

It will not happen by itself. It happens if the people who understand a system build its model in the open. If the data that matters is published with its source. If those who contribute are credited and paid, so that cooperating is not a sacrifice. If every important claim carries a record, and every record can be checked.

Mass, energy, coordination. The first two were always ours. The third is now within reach.

The tools are on the table: the games that explain who wins, the probabilities that say what to believe, the causes that say where to push. Parallel Futures puts them to work on the question that matters most now: which future we are heading toward, and how to move the odds.