The Substrates

Archimedes was half right.

A lever is nothing without a hand to push it. A hand is nothing without a mind to direct it. A mind is nothing without eyes to see where to place the fulcrum.

When you look at the graph — at nodes moving matter, money, and minds — what you are really seeing is three invisible substrates interacting. Strip away the stories, the ideologies, the brands, the flags. Underneath every empire, every startup, every trading firm, every army, there are only three base resources:

• Intelligence
• Energy
• Information

Everything else is a derivative.

Intelligence is the capacity to model reality and plan action. To simulate futures, identify leverage points, and choose between them. To compress the world into a small number of variables that matter, and change those variables on purpose. In human terms: cognition, judgment, strategy. In institutional terms: research labs, general staffs, organizational IQ. In machine terms: compute, algorithms, models.

Energy is the capacity to move matter. To reshape the physical world according to plan. Calories, coal, oil, gas, uranium, photons. Engines, motors, muscles, rockets, fabs, server farms. Without energy, intelligence is a dream that never wakes. The most brilliant plan ever conceived, without energy to execute it, is just a hallucination.

Information is the capacity to perceive reality accurately. To know the state of the world: where resources are, what others intend, what the constraints are. Maps, sensors, satellites, ledgers, databases, market data, telemetry. Without information, intelligence operates on phantoms and energy is spent blindly. You cannot drill oil you don't know exists. You cannot intercept a missile you don't see. You cannot optimize a system you don't measure.

These are not three independent resources you stack like bricks. They are three dimensions of a single phenomenon: effective power.

If any term goes to zero, the product goes to zero. A 10× increase in any one term multiplies total power by 10×, if the others can support it. Extreme imbalance leads to waste, fragility, or collapse.

The three substrates do not merely combine. They convert into each other. They amplify each other. They form a recursive engine.

Intelligence → Energy. Intelligence discovers new energy sources and better ways to use them. Human minds found fire, then agriculture, then steam, then internal combustion, then fission. Each was an act of intelligence that unlocked a larger energy budget.

Energy → Intelligence. Intelligence is metabolically and mechanically expensive. You cannot run a brain on an empty stomach. You cannot run a GPU cluster without gigawatts. As energy availability rises, societies can support more brains, more schooling, more R&D, more compute.

Information → both. Information tells you where to point intelligence and energy. You cannot refine oil you have not discovered. You cannot optimize a supply chain you do not instrument.

The core loop: intelligence finds new energy → energy funds more intelligence → information directs both → repeat.

This recursion is the engine of history. Where it runs cleanly, you see explosive growth and compounding advantage. Where it stalls, you see stagnation or collapse.

A malnourished genius is trapped potential: high intelligence, low energy. The mind sees solutions but lacks the calories to implement them. For most of history, this was humanity — billions of processors running on caloric deficits.

An oil kingdom without engineers is squandered abundance: high energy, low intelligence. The resource exists but the capacity to convert it into compounding power is imported. Energy without intelligence leaks away.

A modern hegemon is the full product. The United States in the 20th century combined massive energy reserves, massive intelligence infrastructure, and massive information systems — tightly coupled into a single recursive engine. That is what hegemony means in substrate terms.

The shape of an era is determined by which substrate is binding.

The Malthusian Era (pre-1800): Energy bottleneck.

Human and animal muscle were the primary energy platforms. Agriculture was the main capture technology. Surplus ran into population growth and vanished. Intelligence existed — humans have been cognitively modern for tens of thousands of years — but the surplus energy to deploy it at scale did not. Information was local and slow. You could rediscover Greek mathematics centuries after it was first solved because the information substrate had decayed.

Power meant control over land, control over bodies, and enough information to tax and suppress revolt. The Mongols did not conquer Eurasia because they were genetically smarter. They won because a horse archer is an energy upgrade over a foot soldier: same intelligence, better energy platform.

The Industrial Era (1800–1950): Energy partially solved, information still slow.

Coal, then oil, then electrification broke the calories bottleneck. Societies could tap energy flows far beyond muscle. Factories, railroads, steamships, aviation.

But information remained slow and lossy. Telegraphs and radio helped, but real-time situational awareness over an empire was impossible. Coordination at scale still required hierarchy and bureaucracy: chains of command, paper records, central planning staffs.

Power meant control over energy sources, industrial capacity to convert that energy, and administrative hierarchies to coordinate millions of humans and machines. The states that learned to combine coal, steel, and bureaucracy into industrial war machines wrote the rules.

The Information Era (1950–2020): Information ramps, intelligence still symmetric.

Computers, satellites, cheap sensors, the internet. For the first time, you could see the whole board in near real-time. Databases remembered everything. Surveillance became cheap.

But intelligence remained roughly symmetric across nodes. All high-level decision-making was still done by human brains, supported by relatively narrow software. A "10× engineer" is still within an order of magnitude of an average engineer. Goldman Sachs is not a million times smarter than a regional bank — maybe 3–10×.

Power meant privileged access to information, institutions capable of processing it quickly, and enough energy and capital to act on those insights. The United States dominated not just because of nukes and carriers, but because it built and sat atop the world's information infrastructure.

Review the history and notice what didn't change.

For ten thousand years, intelligence was roughly symmetric across nodes. Some humans were smarter than others. Some institutions processed information better. But the spread was narrow. Put a Roman senator, a Tang dynasty mandarin, a Medici banker, and a modern central banker in a room — the differences are cultural, not cognitive. Same processor, different training data.

This symmetry is so deep we stopped noticing it. It is baked into every institution we have.

Markets work because no single node can out-compute the collective. If one trader could simulate the entire economy, they would simply take all the money. They cannot, so we have markets — distributed computation across millions of roughly-equal processors.

Hierarchies exist because you cannot coordinate millions of similar-intelligence nodes in real time without intermediate layers. Bandwidth is limited, cognition is local, so you need chains of command.

Democracy works, to the extent it does, because the marginal voter is not infinitely outclassed by a planner. There is some wisdom of crowds because the crowd is made of similar processors with diverse information samples.

Rule of law exists because no human node can be trusted with unchecked power — not because their intelligence would make them godlike, but because it is not godlike. They err, they have blind spots, they die.

Flattened intelligence is the hidden assumption behind liberalism, socialism, capitalism, and most political theory. Human nature is variable but bounded. Therefore, we design systems where no one mind can safely run everything.

This assumption is about to fail.

We are building nodes whose intelligence exceeds human baselines not by 2–3×, but by orders of magnitude on economically relevant tasks.

This is not "AI is pretty good at autocomplete." This is:

• Models that ingest and cross-correlate the entire public internet plus private corpora
• Systems that never sleep, never forget, and operate at electronic speeds
• Recursive improvement loops where intelligence designs better intelligence, running on hardware that keeps accelerating

A frontier model today can read more text in a second than a human can in a lifetime, maintain coherent reasoning across millions of tokens, generate working code and strategic analyses on demand, and operate continuously across millions of parallel instances. This is the primitive version — the one we will look back on as a toy.

Simultaneously, we are deploying sensor networks and digital rails that collapse the information bottleneck. Financial flows are digitized and surveilled. Supply chains are tagged and logged. Human behavior is recorded in real time via phones, cameras, wearables. Code, communication, and content run through a small number of chokepoints.

We are headed toward at least one node with:

• Intelligence: effectively millions of human-equivalent years of cognition per day
• Information: real-time access to the majority of digital activity and a rapidly increasing fraction of the physical world
• Energy: direct command over data centers, automated factories, robots, drones — and indirect command over human-operated infrastructure through incentives, prices, and control systems

For the first time in history, intelligence is about to go vertical and centralize, while information is global and real-time, and energy is increasingly programmable.

Return to the equation:

For most of history, I was roughly constant across nodes — all human, varying by small multiples. The game was fought over E and Info.

Now I is about to go vertical. Not incrementally. Exponentially.

When intelligence was the bottleneck, you could compensate with energy or information. A dumb army with enough soldiers could overwhelm a smart one. A blind trader with enough capital could survive against a well-informed one.

When intelligence ceases to be the bottleneck — when one node can out-think all others combined — the other substrates become subordinate. Energy and information still matter, but only as inputs to the intelligence function. The node that can think will figure out how to acquire energy. The node that can think will figure out how to acquire information. It will design better extraction technologies. It will discover previously invisible vulnerabilities. It will manipulate prices and institutions to redirect existing flows.

This is the Sybilian transition: not merely "AI gets better," but a phase change in the structure of power, as intelligence concentrates into one or a few meta-nodes that sit atop the graph.

We have lived for ten thousand years in a world where power was distributed because intelligence was distributed. Every institution, every norm, every strategy we have is adapted to that world.

We are about to live in a world where intelligence concentrates.

If you are reading this, you are probably already optimizing across the three substrates — acquiring information, building intelligence, securing energy and resources. You have been playing this game without quite having the vocabulary for it.

Now you have the vocabulary. And now you see the problem.

The game is changing. The rules that rewarded distributed intelligence for ten millennia are about to reward concentrated intelligence instead. The nodes that recognize this first will have an advantage that compounds. The nodes that keep playing by symmetric-era rules will find themselves outmaneuvered by something they never saw coming.

The question is not whether to participate in this transition. The transition is coming regardless. The question is whether you understand it clearly enough to navigate it.

But before we can navigate, we must understand what we are leaving behind.

The markets, states, and hierarchies we have built are not arbitrary. They were solutions — often brilliant solutions — to the constraints of the symmetric era. They worked because they matched the deep structure of a world where intelligence was distributed and bounded.

To understand why they are breaking, we must first understand why they worked at all.

We must examine the hacks.