The Thing Underneath Everything
Many narrative violations are easily legible. We refuse to see.
In 1909, a German chemist named Fritz Haber found a way to make ammonia from atmospheric nitrogen and hydrogen, under pressure, using catalysts and heat. Four years later Carl Bosch had scaled the process at BASF into an industrial system. The chemistry is hard. The result is simple. The nitrogen that crops need to grow can now be pulled out of the air at industrial scale, on demand, in whatever quantity the global food system requires.
In 2020, roughly four of the eight billion humans alive on this planet owed their existence to that process. They eat the wheat and the rice and the maize grown with the fertilizer it produces. The reactor that produces the fertilizer runs on natural gas. There is no commercial non-carbon alternative deployable at the scale the species currently requires. None.
This is settled accounting. The numbers come from the energy and materials scientist who has spent forty years counting what modern civilization actually runs on. Most of those four billion people will live their entire lives without knowing the fact. So will most of the other four.
Which raises a question I want to walk through carefully, because I think it sits underneath almost every other civic problem we have. What kind of citizen does a democracy need when the systems it depends on become this technical?
The bar is lower than the technocrats imagine. A democracy works when a citizen can trust the peers and institutions handling the technical detail, when they want the future more than they fear it, and when they hold enough basic numeracy to recognize compounding when it shows up in the data they encounter, and enough historical literacy to read the record of human progress without the distortion of nostalgia.
This essay is about the citizen we still need, and my attempt to help you be one.
What civilization actually runs on
Cement. Steel. Plastics. Ammonia. These are the four pillars of modern civilization. In 2019, the world consumed about 4.5 billion tons of cement, 1.8 billion tons of steel, 370 million tons of plastics, 150 million tons of ammonia. Making them eats roughly seventeen percent of global primary energy and produces a quarter of all fossil-fuel CO₂ emissions. None of the four is readily replaceable at scale. There is no other industrial material with steel's strength-to-cost ratio, no other malleable polymer family that does what plastics do, no other binder that produces reinforced concrete, no other way to fix atmospheric nitrogen at the rate ammonia synthesis does.
The four pillars is Vaclav Smil's phrase. He has spent forty years counting the biophysical substrate of modern life and has written more than forty books about it. His 2022 popularizing summary How the World Really Works is where the figures above come from. Smil is not making a moral argument. He is making a material one. The morality is what the reader has to do afterward.
Underneath the four pillars sits a second layer Smil insists is even less visible than the first. The diesel engine, fed with refined crude, moves nearly everything humans move at scale. Container ships, freight trains, long-haul trucks, ocean-going bulk carriers, the equipment that plants and harvests almost all industrial agriculture. The popular story of globalization is digital. The actual story is mechanical. Most of what is in your kitchen and most of what is on your back spent thirty days on a ship whose 109,000-horsepower diesel engine burned 4,650 tons of low-grade residual oil to push twenty-three thousand other steel boxes across the ocean. The internet coordinates the chain. Diesel and the box are the chain.
This is what Smil means by substrate. Every other domain we talk about (food, housing, transport, communication, medicine, sanitation) rests on a thicket of bulk materials and energy flows whose mass and timescale most public conversation has never engaged.
The substrate also explains why decarbonization arguments that ignore the four pillars are missing most of the problem. The electricity grid is only about a fifth of final global energy use. Roughly eighty percent of what humanity actually uses energy for is industrial heat, transportation, and processes the grid does not touch. A debate about electric vehicles is a debate about something less than a fifth of the problem. A debate about renewable electricity is a debate about another fraction of the same fifth. The other four-fifths sit under steel mills, cement kilns, fertilizer plants, smelters, refineries, container ships, and the kilometer-deep mines that supply them. The grid does not reach any of them.
Energy is the layer everything else sits on.
Smil's discipline through the book is to refuse moral framing. He is impatient with apocalyptic environmentalism and equally impatient with Silicon Valley solutionism. His standing line is that "we are a fossil-fueled civilization whose technical and scientific advances, quality of life, and prosperity rest on the combustion of huge quantities of fossil carbon." Stripping that combustion out by 2050, he writes elsewhere in the same book, is conceivable only "at the cost of unthinkable global economic retreat, or as a result of extraordinarily rapid transformations relying on near-miraculous technical advances." He endorses neither. He refuses the both-poles framing in which one has to be chosen.
What the book does instead is count. Cement in tons. Ammonia in tons. Energy in joules. CO₂ in gigatons. Transitions in decades. The numbers do not by themselves say what should be done. They do say what is. And what is, in Smil's accounting, is a civilization whose material dependencies cannot be rearranged on a presidential election cycle or a quarterly earnings call.
The secret silent miracle
The substrate exists. The question that should sit next to it is harder than it sounds. What has that substrate produced?
For thirty years, the same experiment kept producing the same result. Audiences around the world, including medical students, journalists, hedge-fund managers, and Nobel laureates, were given thirteen multiple-choice questions about the basic state of the world. What share of children are vaccinated. How extreme poverty has changed. How life expectancy has moved. Almost everyone failed, and they failed in the same direction. They guessed at a darker, more frightening, more polarized world than the one the data described. A chimpanzee picking randomly would have scored thirty-three percent. The human average was closer to sixteen. The error was in their biased thought, not random.
The experimenter was Hans Rosling, a Swedish physician and global-health professor who spent the last decade of his life turning that finding into Factfulness. What the data actually show, in his reading, is what he called the secret silent miracle of human progress.
The share of humanity living in extreme poverty was eighty-five percent in 1800, twenty-nine percent twenty years before he wrote, nine percent in 2017. Life expectancy at birth was thirty years in 1800. It is seventy-two now. Eighty-eight percent of one-year-olds globally are vaccinated against something. Sixty percent of girls in low-income countries finish primary school: most audiences guess twenty. Deaths from natural disasters have fallen to about six percent per capita of what they were a century ago.
Rosling was careful, and the carefulness is what saves the book from the cheerful-progress cartoon it's enemies might try to paint it to be. He refused the optimist label. He coined the word possibilist for the stance he wanted:
"I'm not an optimist. That makes me sound naïve. I'm a very serious 'possibilist.' That's something I made up. It means someone who neither hopes without reason, nor fears without reason, someone who constantly resists the overdramatic worldview."
The possibilist holds two thoughts at once. The world is bad. The world is better. A premature baby in an incubator is both improving and critical. The trend can be sharply upward and the situation can still be unacceptable. The discipline is to keep both in mind without flinching toward consolation or despair.
What Rosling diagnoses across the book is the cognitive equipment that makes the directional error so reliable. Ten dramatic instincts — gap, negativity, straight line, fear, size, generalization, destiny, single perspective, blame, urgency — each adaptive in the small-group environments where human cognition evolved, each catastrophically miscalibrated for the modern information environment. The fear instinct selects for what is frightening over what is dangerous. Plane crashes and shark attacks pass through the filter. Diarrhea and traffic accidents, which actually kill far more people, do not.
Heat kills more Europeans each year than guns kill Americans. Heat deaths in Europe ran above forty-seven thousand last year. Gun deaths in the United States ran below it.
The fear instinct selects for what is frightening over what is dangerous. Plane crashes pass through the filter. Diarrhea does not.
The lift Rosling counts and the substrate Smil counts are the same civilization seen from two angles. Rosling shows what has actually happened to human lives across two centuries. Smil shows what it physically takes to keep that happening. Both stories are true. Read together, they describe a species that has accomplished one of the most dramatic improvements in human welfare in recorded history while embedding itself ever more deeply in material flows that are not obviously sustainable.
I want to take a smaller scale now, because the global numbers are easy to nod at and forget. Let me put myself in the picture.
In Smil's book there is a kitchen-table demonstration. Take a tomato from a heated greenhouse in Almería, the kind shipped to a Scandinavian supermarket in February. Slice it. Pour five tablespoons of dark oil over the slices. That is what you are looking at when you put a winter tomato on a salad in a Northern European supermarket. About 650 milliliters of embedded diesel fuel per kilogram of fruit. Most of it for the greenhouse heat, some for the transport, the rest for the fertilizer that grew the plant. Smil's point is the consumer surface that hides the substrate, not the tomato itself.
I have eaten foods like this in the winter myself. I have rarely thought about it. I certainly have never broken it down.
The same blindness runs through my electric bill. Once a month a dollar amount arrives in my email. I pay it. If asked what fraction of that dollar amount paid for heat, what fraction for hot water, what fraction for the refrigerator, what fraction for the lights, I could not tell you with much confidence. I have lived in this house for years. Even the writer who reads about energy for fun cannot decompose his own monthly bill.
The substrate underneath the dollar amount is invisible to me. I am someone who reads about energy for fun.
The blindness extends upward. The civic story that has run through 2025 and into 2026 in my part of the algorithm is data center water use. Counties holding up permits. Citizen groups filing public-records requests. Op-eds measuring the cost of artificial intelligence in gallons. Denver City Council voted unanimously to halt new data centers for twelve months, effective the week this essay was finalized.
The data is public. In 2023, all the data centers in the United States combined consumed about seventeen billion gallons of water for direct cooling, plus roughly two hundred billion gallons more indirectly through the electricity that powered them. Call it two hundred twenty-eight billion gallons all in, on the most generous accounting.
In the same year, residential lawn irrigation in the United States consumed about three and a third trillion gallons. California almonds alone consumed roughly one and a half trillion more. The lawns the country waters and the almonds it grows together used something like twenty times more water than every AI query, every CRM database, every video stream, text message, and Instagram Reel the country generated combined.
No one seems to be protesting almonds.
What is doing the work is the dramatic instinct. Data centers are new, technological, corporate, and run by companies whose names trigger something. Lawns and almonds are familiar, agricultural, residential, and run by no one whose name I know besides my neighbors' 2500sqft of turf. The fear filter passes one and not the other, exactly as Rosling diagnosed. The inflamed group does not care that the math runs the other way.
These three blindnesses cut across partisan lines and across the question of good faith. They are the routine condition of a human being trying to make sense of a technical civilization through interfaces (the supermarket shelf, the monthly bill, the news headline) that were designed to hide the substrate rather than expose it. The civilization functions because the substrate is hidden. The democracy functions, when it functions, because someone bothers to look.
The dramatic-instinct economy Rosling diagnosed runs on these blindnesses. Every successive incentive in the chain that connects the substrate to the citizen (advertising, branding, retail, journalism, social media, political campaigning) selects for vividness over accuracy and for emotional valence over orders of magnitude or analytical thought. The miscalibration is the system's product, not its malfunction.
The pencil on your desk required loggers in Oregon and graphite miners in Sri Lanka, and not one of them knows how to make a pencil alone. Two strangers, neither related, swap two different objects at the same time. Both walk away better off. Multiply that across a population, repeat for a hundred thousand years, and the species' intelligence stops being held in any one head and starts being held in the exchange itself.
The word for this is Hayek's: catallaxy, the spontaneous order generated by exchange and specialization. Hayek named it. The third voice in this dialogue carried it from political economy into deep human history. Matt Ridley is a zoologist who, in mid-career, turned into a popularizer of evolutionary thinking. His 2010 book The Rational Optimist extends Darwin and Adam Smith into a single explanation of how the modern world got so rich.
Ridley's deeper claim is that the catallaxy compounds when ideas can find each other, and that the rate at which ideas can find each other has accelerated faster than any other variable in human life since 1800. Population multiplied six times. Real income rose nine times. Life expectancy more than doubled. Vietnamese poverty fell from ninety percent to thirty in twenty years. The hockey stick is real.
But Ridley does something else that matters for the substrate question. He grounds the catallaxy in physics. The hockey stick rests on cheap dense energy in sufficient supply. Coal made every modern person, in Ridley's image, a little Louis XIV. By his arithmetic, it would take roughly a hundred and fifty human beings pedaling exercise bicycles in eight-hour shifts to deliver the energy the average global citizen consumes. Six hundred and sixty for the average American. Modern prosperity rests on cheap dense energy: it is the substitute for the form of slavery cheap dense energy made uneconomic.
Rosling counts the lift. Smil counts the substrate. Ridley names the mechanism that ties them together. The catallaxy needed the substrate to produce the lift. The substrate exists. The lift is profound. The three voices converge.
And here is where the dialogue produces something none of them quite says alone.
The three accountings are public knowledge. They are in mainstream books written for general audiences. They are taught nowhere in particular. The conversation that should follow from them, about what kind of civilization we want, what we are willing to spend to keep it, what we are willing to lose to change it, is conducted instead through the interfaces I named in the last section. Through headlines that confuse stocks with flows. Through bills whose components are illegible. Through the dramatic-instinct economy that selects for the photogenic over the load-bearing.
The catallaxy needed the substrate to produce the lift. The conversation that should follow from those three accountings is conducted through headlines, bills, and slogans.
Vaclav Smil wrote, near the start of his book, a sentence I want to give in full because it is the civic argument the whole substrate project is making in one line:
"The gap between wishful thinking and reality is vast, but in a democratic society no contest of ideas and proposals can proceed in rational ways without all sides sharing at least a modicum of relevant information about the real world, rather than trotting out their biases and advancing claims disconnected from physical possibilities."
This is the turn. The substrate of consent is a different thing than the conclusions consent should reach. The substrate of consent is the minimum civic numeracy that lets the conclusions be reached at all. A citizen who cannot tell a megawatt from a megawatt-hour cannot consent to the energy decisions made in their name. A citizen who has never looked at the four pillars cannot meaningfully weigh in on what they are worth and what should replace them. Self-rule in a technical civilization requires self-acquaintance with the systems being ruled.
The point lands harder because of when we are. The two decades behind us were a window in which the binding constraints were attention, code, and cheap money. The substrate could be hidden behind the interface. That window is closing. The energy needed to train and run artificial intelligence, the chips required to build it, the manufacturing required to build the chips, the materials and mines and diesel substrate the manufacturing rests on, are all reasserting themselves as the binding constraint on the next decade. The physical world is coming back. The numeracy that did not matter when the constraint was attention now matters because the constraint is joules.
The substrate of consent
I have written before about the politics of energy. One essay diagnosed historicism, the conviction that history has a knowable destination, as the failure mode driving signal-first energy policies across Western democracies in our own decade. Another, future piece under construction now, lays out specific proposals (a carbon price, nuclear deployment at scale, transmission infrastructure, the institutional posture I think the environmental movement built in the 1970s and now needs to revise) and will argue for them under my own name. Both of those essays argued. This essay does not. It establishes the substrate underneath argument. The numeracy without which no argument about energy can begin.
Numeracy is a virtue you practice into. Start with six numbers and three rules.
The six numbers: four billion people alive because of synthetic ammonia. The four pillars in tons in 2019: 4.5 billion of cement, 1.8 billion of steel, 370 million of plastics, 150 million of ammonia. Eighty percent of human energy use the grid does not touch. Global life expectancy thirty years in 1800, seventy-two now. The share of humanity in extreme poverty, eighty-five percent in 1800 and nine percent in 2017. A hundred and fifty humans on exercise bicycles to deliver the average global citizen's energy, six hundred and sixty for the American.
The three rules: stock versus flow, watt versus watt-hour, percentage of new capacity versus percentage of total use. The discipline of pausing on a headline long enough to ask the unit, the kind, the denominator. A thirty-second move. It catches most of the innumeracy modernity throws at you.
The willingness to be the person at the dinner table who asks. Not because numbers settle arguments. They do not. But because the argument cannot even begin without them.
The chain runs Individual → Society → Civilization. We say it often at the Dispatch. In a technical age, the first link does the work the rest depends on. Knowing what you do not know about the systems you vote on is the smallest version of self-rule.
Sources and Inspiration
How the World Really Works - Vaclav Smil
The Rational Optimist - Matt Ridley
Numbers Don't Lie - Vaclav Smil