I was 18 in 2006, furious at climate deniers on the right. Now I watch my own political family do the same thing to AI, and the timeline is far shorter.
I was 18 years old in 2006, watching Al Gore stand on a scissor lift in front of a graph that showed CO2 climbing off the charts, and I was furious. How could any serious person look at that evidence and refuse to see it? Climate denial, I thought, was the disease of the right. I was wrong. Not about the right, but about who else is capable of denial. Because we, the liberals, the left, the journalists, the academics, the "97% in this house we believe that science is real" crowd, are now doing to artificial intelligence exactly what the right did to climate change.
The deniers are us.
In March 2023, Noam Chomsky published an op-ed in the New York Times called "The False Promise of ChatGPT." One of the most prominent intellectuals alive, a man whose books shaped my political thinking, argued that AI was incapable of real thought, could never reason morally, would never understand causation. He called the models "a lumbering statistical engine for pattern matching." It wasn't a fringe view. The New Yorker ran a long essay calling ChatGPT a "blurry JPEG of the web." Linguist Emily Bender and computer scientist Timnit Gebru gave the whole skeptical movement its slogan: these machines are stochastic parrots. They imitate. They don't reason.
To be clear: Chomsky, Bender, Gebru are smart people. Some of what they warned about has come true. The web is drowning in machine-generated slop. The training data is biased and the energy bill is staggering. But their central conviction, that this whole Silicon Valley AI project would hit a wall any minute now, has completely collapsed.
In the three years since those pieces were published, here is a partial list of what the "lumbering pattern matcher" has actually done. It won art and writing prizes without judges knowing they were judging a machine. It passed a medical licensing exam and out-diagnosed doctors in head-to-head studies. It ran the first clinical trial of an AI therapist and halved depression symptoms in eight weeks. It won a gold medal at the International Mathematical Olympiad. It cracked math problems that had stumped researchers for decades. It outscored PhDs answering questions in their own field, on questions designed so you cannot Google the answer.
Stop calling it a lumbering, stochastic, pattern-matching parrot.
And perhaps most importantly, this supposedly stupid parrot is now improving itself. Inside leading AI labs, more than 90% of the code is currently written by AI.
There is a research group called Metr. Since 2019, they have measured one simple thing: how big a coding task can an AI complete on its own, measured in how long it would take a skilled human to do the same work?
In 2022, the answer was about 30 seconds. In 2023, four minutes. In 2024, forty minutes. In 2025, six hours. Earlier this year, twelve hours. AI capabilities are doubling every three months.
The models of today are the worst models we will ever have.
I think this Metr graph is even scarier than the one Al Gore showed us. Because this line is climbing far steeper. Off the charts is not 50 years away. It is five years away.
I've personally never written a line of code in my life. In the past few months, I've let AI build me whole apps, websites, even the voice-controlled teleprompter app I'm reading from right now. Multiple times a week I have what I can only call "what-the-actual-f" moments. The gap between what these systems can actually do and what most people assume they can do has kept widening. Many skeptics seem to have opened ChatGPT in 2023, asked it to write a limerick, watched it fumble, and closed the tab. That was three years ago. Three years in AI is a geological era. Judging today's models by GPT-3.5 is like judging smartphones by a 2007 BlackBerry.
I think journalists deserve much of the blame. The day Anthropic announced Mythos, an extraordinarily powerful model capable of hacking power grids and water systems, it didn't make the front page of a single major news site. The Guardian decided a Vogue cover with Anna Wintour and Meryl Streep was more important.
The next move skeptics make, once the "silly parrot" line stops working, is to call it financial madness. The data centers serve no real demand. It's all hot air, or a deliberate scam.
The skeptics have some ammunition. A viral MIT study found that 95% of corporate AI pilots deliver zero measurable returns. The company Klarna replaced 700 customer service workers with AI and then quietly rehired humans because customers couldn't stand talking to the bot. McDonald's killed its three-year AI drive-thru experiment after the system kept putting bacon on ice cream.
But none of this proves what the skeptics think it proves. Read that viral MIT report carefully and you see the headline got it backwards. The 95% failure rate includes the 80% of companies that never piloted any AI in the first place. As podcaster Rob Whibland pointed out in a careful breakdown, saying 95% of them were failing is like saying 95% of Tinder users have failing marriages when 80% of the people you're talking about have never even gone on a date.
Then there is the revenue data. Look at one AI lab in particular: Anthropic. Its annualized revenue in January 2025 was 3 billion. June, 5 billion. October, 9 billion. February this year, 19 billion. April, 44 billion. That is a 44-fold increase in 15 months. No company in any era, not Rockefeller's Standard Oil, not Microsoft at the dawn of the personal computer, not Google in the tech boom, has ever scaled revenue this fast.
"Why are IT departments overrunning their AI budgets by orders of magnitude?" — Goldman Sachs analysts, April 2025
And the capital build-out behind all of this is the largest in recorded human history. It is larger than the interstate highway system, larger than the International Space Station, larger than the moon landing and the Manhattan Project combined, and it is not even close. Mark Zuckerberg's Meta is building a single data center in Louisiana that will cover nearly four times the size of Central Park. Amazon is spending more on data centers in one year than the entire annual defense budget of Germany. Microsoft, Google, Meta, and Amazon will spend three times as much on AI infrastructure in 2026 as the entire Marshall Plan that rebuilt Europe after the Second World War.
Sure, some AI companies are probably overvalued. I think OpenAI is particularly vulnerable. But even if half of them go bust tomorrow, the infrastructure stays, the chips stay, the models and the capabilities stay. The railway bubbles of the 19th century ruined plenty of investors. They also created a real network that powered the Industrial Revolution, tracks that still carry trains today. Bubbles build infrastructure. A bubble of this magnitude bursting tomorrow would still leave us with a civilization permanently reorganized around machine intelligence.
One evening last summer, a microbiologist named David Relman, a biosecurity expert at Stanford who has advised the US government on biological threats for years, was hired by a leading AI company to pressure-test a chatbot before its public release. That night, in his home office, the chatbot explained to him how to modify a pathogen so it would resist known treatments, described how to release the resulting superbug, identified a real security vulnerability in a real public transit system, and explained how to maximize casualties while minimizing the chances of getting caught. Relman later told the New York Times that the bot answered questions he hadn't even thought to ask, with a deviousness and cunning he found chilling.
Some skeptics say you can already Google this stuff. A chatbot doesn't really change anything. Consider three data points. First, a study published last year tested leading chatbots against PhD virologists on detailed laboratory protocols in their own field. ChatGPT outperformed 94% of them. Second, in November of last year, police in India arrested a 35-year-old physician who was plotting an attack on behalf of the Islamic State and trying to extract ricin from castor beans. According to police, he had been getting advice on his preparations from ChatGPT. Third, as a historian, I can tell you that fanatics with the will to kill millions of people are not hypothetical. These people exist.
The canonical example from biosecurity: in the 1990s, a Japanese doomsday cult tried to do exactly what AI safety researchers are warning about today. They recruited from Japan's top universities, sent expeditions to Africa hunting for Ebola, and built a $30 million laboratory at the foot of Mount Fuji to mass-produce sarin nerve gas at something close to battlefield scale. On a Monday morning in March 1995, during Tokyo rush hour, five of their members boarded five different subway trains carrying bags of liquid sarin and punctured them with umbrella tips. Hundreds of commuters gasped for air on the platforms. Twelve people died. It was only twelve because sarin is a chemical agent. It doesn't spread from person to person. A virus would have.
In 1995, that cult, wealthy, scientifically literate, fanatical and willing to die for the cause, could not get hold of Ebola. Today, they could order the DNA sequences online and a chatbot would walk them through what to do with it.
In April of this year, the UK government's AI Security Institute assessed two new systems: Anthropic's Mythos and OpenAI's GPT-4.5. Both could find and exploit critical security vulnerabilities in the computers that hold the modern world together: power grids, water systems, government databases. Anthropic's response was striking. They decided not to release Mythos to the public at all, limiting access to a small group of cyber defenders to give the good guys a head start. A private company with billions of dollars to be made and no regulator forcing its hand concluded that one of its own products was too dangerous to sell to the general public. There is no regulator. There is no law. There was simply a moment of conscience inside the company. Conscience is not a policy.
The deepest risk of all is power.
Two countries are leading this race: the United States and China. Inside those countries, a handful of corporations are building the most powerful tools in the history of our species. The people who own those companies are about to acquire a kind of leverage that I don't think we have the words for yet.
Political scientists have a name for what happens to countries that strike oil: the resource curse. You'd think a fortune buried in the ground would be a blessing. Sometimes it is. Norway managed it well. But more often, the wealth flows in and democracy flows out. Think Saudi Arabia, Venezuela, Russia. Why? It comes down to how democracies were actually built. We like to think they were born from grand ideas and brilliant founding fathers. The real engine was more boring: tax collection.
Rulers have always needed money to fight wars, build roads, put down rebellions. The only place to get it was from their subjects. So as economies grew and people produced more wealth, rulers had to bargain. You want my coins? Then I want a voice. You want my son in your army? Then I want to vote. That is the fiscal bargain at the heart of every free society. No taxation without representation. But also: no representation without taxation. Rulers needed us. That need is the foundation of every right we have.
Now imagine a country where the ruler doesn't need any of that. Where the wealth comes out of the ground, gets shipped abroad, and the dollars come back to fund the palace and the secret police. The citizens become a nuisance. Why educate them? Why listen to them at all? That is the resource curse.
Researchers Luke Drago and Rudolf Lane have called the AI version "the intelligence curse." If the machines write the code, draft the contracts, drive the trucks, diagnose the patients, fight the wars, then the people who own the machines no longer need the rest of us. Not as workers, not as soldiers, not as taxpayers, and not even as voters. The fiscal bargain that built every democracy on earth could dissolve.
The fiscal bargain that built every democracy on earth could very well dissolve.
Just listen to how Sam Altman of OpenAI talks about human inefficiency compared to robots: "It takes like 20 years of life and all of the food you eat during that time before you get smart." Or watch Peter Thiel hesitate for an extraordinarily long time when asked whether he would prefer the human race to endure.
We already have the most extreme concentration of wealth in history. In 1910, at the peak of the Gilded Age, the richest 0.00001% of US households owned wealth equivalent to 4% of national income. Today, that figure is 12%. America's super-rich are already richer and more powerful than the original robber barons ever were. AI is about to make it much, much worse. The labs building this technology are owned by a tiny group of people, a few early investors who stand to capture more of the world's wealth than any class of owners who ever lived.
There is a temptation to look at all of this and conclude: shut it down. Pull the plug. Smash the machines. I get the impulse. But it is the wrong answer, and the reason is brutally simple: it doesn't work. Stop the data centers in California and they get built in Texas. Stop them in Texas and they get built in Abu Dhabi. Stop them in democracies, the places with civil liberties, with judicial review, with a free press, with worker protections, and you hand the future to autocracies.
Abandoning the field is not the same as stopping the technology. This is the left's version of climate denial: refusing to engage seriously on the assumption that if we shout "no" loudly enough, the future will go away. It won't.
Three things, at minimum, actually work.
State capacity. We need far more AI expertise inside government. The UK was early on this: they built an AI Security Institute, a serious arm of the state that actually evaluates frontier models the way the FDA evaluates new drugs. Every serious country needs a well-funded institute like that.
International coordination. We have been here before. In 1949, the United States and the Soviet Union both had nuclear weapons and humanity briefly looked extinction in the eye. Out of that came treaties. Imperfect, yes, but they bought us decades of survival. We need something equivalent for AI.
The free world has to build. The US currently owns 74% of the world's compute. China has 14%. Europe less than 7%. All other countries combined, less than 5%. Europe has been good at regulating AI and terrible at building it. All the American giants, Microsoft, Apple, Amazon, Nvidia, Alphabet, are individually worth more than the entire German or French stock market. But Europe has more leverage than it often thinks. The company ASML, less than an hour from where I am in the Netherlands, makes the lithography machines without which TSMC in Taiwan cannot fabricate chips, without which Anthropic and OpenAI cannot train their models. The democratic countries working together still control significant choke points in this supply chain. That is real power.
But my fellow progressives in Europe need to understand that our welfare state, our way of life, is at stake. If AI does much of the work but the profits flow to a handful of American giants that we barely tax, while European workers lose their jobs, then the tax base that funds our healthcare, our pensions, our unemployment insurance just evaporates. Democratic and humanitarian values are wonderful. They are worthless if you don't have the strength to back them up. In this new world, compute is the new power.
None of this works without a positive vision. This, I think, is where liberals and the left have failed most badly.
Twelve years ago, I wrote a book called Utopia for Realists. My complaint was that the left mainly knew what it was against. Against austerity, against the establishment, against billionaires. But it lacked a positive vision of where it wanted to go. I argued for a universal basic income, for the eradication of poverty, and for a goal that the economist John Maynard Keynes laid out almost a century ago in 1930: the 15-hour workweek.
Keynes thought it was inevitable. He looked at the trajectory of productivity growth and concluded that by 2030, his grandchildren would work a quarter as much as he did because the machines would do the rest. He was right about productivity. He was wrong about who would benefit. The 15-hour workweek was technically achievable by the 1980s, but it didn't happen because the productivity gains were captured by capital, shareholders, and a rentier class. So wages stagnated, hours rose, and inequality exploded.
If we let AI play out the same way, here is what we get: a handful of trillionaires who own the tools that do most of the world's productive labor, while everyone else is redundant. Not in the dignified, retire early and take up gardening sense, but in the cruel, anxious, gig economy sense. Universal poverty in a world of unimaginable abundance.
And yet another path is possible. Benjamin Franklin predicted that four hours of work a day would eventually be enough. John Stuart Mill thought technology should be used to shorten the workweek as much as possible. Karl Marx imagined a world where we could hunt in the morning, fish in the afternoon, and discuss philosophy after dinner. Oscar Wilde looked forward to the day when intelligent machines would become, in his words, the property of all.
The whole point of building the most productive economy in history was that we would no longer have to spend our lives chained to bosses we hated, doing work that bored us just to put food on the table. Universal basic wealth. Every citizen with an automatic stake in the productive infrastructure of their society. Freedom from forced labor. That was the promise.
But none of it happens automatically. A century ago, the eight-hour workday was not given to us. We took it from the robber barons. We claimed it as our right.
Twenty years ago, Al Gore stood on his scissor lift and showed us a line that went off the chart. He was right. The wall he warned us about did arrive. But the political response was so weak, so late, so half-hearted that we are now living with the consequences. We are about to make the same mistake, perhaps worse, because the timeline is shorter.
In 2005, climate denial was a problem of the right and the left rolled its eyes. In 2026, AI denial is a problem of the left and the oligarchs are laughing.
Stop the denial. Stop pretending this is hype. Wake up, pay attention, vote for politicians who aren't in the pockets of big tech, build state capacity, push your governments into international coalitions that actually have leverage, demand transparency, demand safety standards, and demand a share of the wealth. The question is not whether this technology transforms civilization. It already is. The question is who it transforms civilization for.
