The Past Is Not in the Option Set
About that little house.
Robert Reich ends his case against artificial intelligence with a wonderfully direct question: why can’t we “stop the whole damn thing?”
One of the most popular comments underneath, liked by Reich, quotes Rachel Maddow from a live event: “The future is analog, I’m telling you.” Build real relationships. Know people offline.
Someone replies that they are watching Little House on the Prairie. They were born in the twentieth century, live in the twenty-first, and increasingly feel that their soul belongs in the nineteenth.
It would be easy to laugh at this.
I don’t want to.
Modern life really does require too many screens, accounts, passwords, subscriptions, rankings, feeds, notifications, forms, and machines watching other machines watch us. A neighborhood where people know one another is safer in ways no predictive system can reproduce. A child should be able to disappear into an afternoon without becoming data. A person should be able to buy something, repair something, read something, or sit in a room without a server farm learning from the occasion.
So build that world.
Know the people next door. Put the phone down. Keep paper records where paper is safer. Learn how to fix things. Grow food. Strengthen local institutions. Build systems that can survive a dead network connection.
Then look past the end of the street.
The mistake begins when a good rule for living is promoted into a plan for civilization.
There were about 8.2 billion human beings alive in 2024. Our food system now depends upon technological capacities that did not exist in the nineteenth century.
- Ammonia-derived nitrogen fertilizers alone are estimated to produce food feeding roughly 3.8 billion people, about half the global population.
- Current ammonia production remains heavily dependent on fossil fuels, while excessive nitrogen use creates serious environmental damage.
That is one of the ugliest facts of modernity because every comforting answer breaks against it.
Industrial agriculture has damaged soil, water, climate, habitat, and the nitrogen cycle. We need different farming systems, far less waste, better nutrient recovery, lower-carbon fertilizer production, healhier soils, and much more room for ecological repair.
We also have billions of people currently eating through the system we have.
Turn it off before the replacement can carry them and the old world does not grow back underneath us. The fertilizer does not become manure. The global population does not politely shrink into nineteenth-century settlement patterns.
Food production loses one of the central inputs around which a century of agriculture has been organized.
The existing system deserves radical transformation.
Transformation is a transition through a world full of bodies.
The planet outside that system has changed too.
The World Meteorological Organization reports that 2015 through 2025 were the eleven hottest years on record. In 2025, global temperature was about 1.43°C above the 1850–1900 average, while Earth’s measured energy imbalance reached the highest level in a 65-year record. The oceans keep accumulating heat.
Some consequences are already committed over much longer timescales. The Intergovernmental Panel on Climate Change concludes that some future changes are unavoidable or irreversible, even though rapid emissions cuts can sharply limit how much worse they become. Sea-level rise will continue for centuries to millennia because oceans and ice sheets respond slowly. Continued glacier loss and other long-duration changes remain in motion.
So imagine that we achieve the most spectacular act of technological refusal in history tomorrow morning.
Every data center goes dark. Every artificial-intelligence company closes. Every billionaire founder is sent home to garden.
The ocean does not give the heat back.
The coastline does not return.
The glacier does not climb uphill.
The city built for yesterday’s temperature does not cool itself.
The farm experiencing a changed rainfall pattern does not receive the old climate because we recovered our souls.
Stopping additional damage is indispensable. It is no longer sufficient.
We have entered a period in which civilization has to repair, adapt, relocate, redesign, restore, maintain, and sometimes retreat while keeping people alive through the work. Developing countries alone are estimated to need roughly $310 billion to $365 billion every year for climate adaptation by 2035. Current international public adaptation finance is a fraction of that.
There is no untouched natural world waiting for us to stop interfering with it.
There are many places where the wisest intervention will be to stop building, stop extracting, stop poisoning, remove what we installed, and allow living systems to recover. There are also cities that need cooling, coastlines that need decisions, water systems that need rebuilding, forests that need protection, electrical grids that need changing, people who will need to move, and damaged ecosystems whose restoration itself requires knowledge, labor, equipment, money, and coordination.
Monday morning just keeps arriving.
Insulin still needs refrigeration. Sewage still has to stay out of drinking water. A dialysis machine needs electricity. Wheat needs nitrogen from somewhere. A city needs food brought inward and waste carried outward. Transformers fail. Medicines expire. Bridges rust. Wildfire smoke crosses county lines. Records have to survive the people who created them. A hospital needs somebody awake when everyone else is asleep.
“Look after your neighbors” is excellent moral advice.
It becomes political fantasy when everything your neighbor depends upon outside the neighborhood is treated as fictional.
Your neighbor may be a diabetic whose medicine arrived through a global pharmaceutical supply chain. Or a nurse working in a hospital attached to a continental electrical system. Or a parent buying food produced with machinery, fertilizer, refrigeration, roads, warehouses, inspection, payment systems, and thousands of strangers. Or a renter living in a building that becomes dangerous during heat its designers did not expect. Or a farmer whose growing season is moving underneath decades of accumulated practice.
Mutual aid can save people when large systems fail. It can expose what those systems forgot. It can carry forms of care that no bureaucracy will ever reproduce.
It cannot become an excuse for letting the systems fail.
There is a much more serious version of the analog argument. Deliberately simplify. Relocalize production where local production can carry the work. Shrink environmentally ruinous industries. Destroy surveillance advertising. Build products that last. Walk more. Drive less. Restore local agriculture. Reduce energy and material throughput. Stop manufacturing demand for disposable garbage. Give communities more control over the systems they actually inhabit.
Good.
Now do it.
- Which industries close first?
- Which workers keep income while they close?
- Which farms receive replacement inputs?
- Which supply chains become local,
- and which become dangerously brittle if they do?
- How does the electrical grid remain stable while generation changes?
- Which buildings are retrofitted first?
- Which communities receive retreat money?
- Who owns the replacement infrastructure?
- Which materials are scarce?
- Which ecological damage gets priority?
- How does a hospital keep functioning through the transition?
Degrowth serious enough to work would be one of the largest coordination projects civilization has ever attempted.
A smaller destination does not imply a smaller transition.
This is where Reich’s argument reaches something real and then takes one step too far. He is right to see extraordinary private power forming around artificial intelligence. He is right that data centers consume electricity, water, land, capital, and political attention. He is right that a handful of executives do not acquire authority over everyone else’s future because they built useful technology.
Those conclusions do not establish that civilization can dispense with high technological capacity.
Post-capital cannot mean post-capacity.
Some systems should be abolished. Some capacities should never be built. Some should become local. Some should become public infrastructure. Some should remain available through markets. Some belong in commons. Some require specialists. Some need hard limits that no purchaser, ministry, founder, or machine may buy away.
Artificial intelligence enters that already difficult field as another capacity. It may help compare enormous bodies of evidence, coordinate infrastructure, search scientific possibilities, reduce administrative labor, expose failures across institutions, and help smaller organizations do work once reserved for enormous ones. It also carries substantial material costs and can make existing concentrations of power much harder to escape.
The important political question therefore cannot be whether the machine feels warm, human, natural, or analog.
Ask who owns it. Ask what function it carries. Ask which resources it consumes. Ask who bears the damage. Ask what happens when it is wrong. Ask whether another institution can inspect it, replace it, refuse it, or continue the work after it disappears.
Capacity does not create title.
And refusal of title does not make the work disappear.
Know your neighbors. Build a life that can survive a dead phone. Keep rooms no model can enter. Learn the names of the people who live beside you.
Then look past the end of the street.
Somebody is keeping the water clean. Somebody is balancing the grid. Somebody is growing food for a city that cannot grow its own. Somebody has to cool the buildings, restore the soil, move the medicine, rebuild infrastructure, and carry billions of people through a climate we have already changed.
- The future cannot be analog.
- It also does not have to belong to whoever owns the biggest machine.
Once that much is admitted, the harder problem begins: how do we make all these systems capable of working together without appointing one of them king?
Comments ()