AI is a mass psychotic delusion

In April of 1856, a teenage girl named Nongqawuse stood by the banks of the Gxarha River in South Africa and received a message from the dead.

​She told her uncle that she had spoken with the spirits of their ancestors. The spirits delivered a prophecy, one that offered a total and miraculous salvation for the Xhosa people, who were then buckling under the encroaching weight of the British Empire. The ancestors promised to rise up, drive the British into the sea, and usher in an era of unimaginable utopian abundance. Flocks of beautiful, immortal cattle would emerge from the earth. The fields would spontaneously yield grain. Sickness and old age would be eradicated.

​But, as with all divine contracts, there was a catch.

To trigger this utopia, the spirits demanded an act of supreme faith. The Xhosa had to slaughter all of their existing cattle. They had to burn their crops. They had to destroy their stores of grain and empty their water reserves. They had to liquidate their entire present economy to prove they were worthy of the future.

​And so, they did.

​Historians estimate that the Xhosa slaughtered over 400,000 cattle. They burned their own fields to ash. They waited for the sun to turn red and the ancestors to arrive. But the sun remained yellow. The ancestors did not rise. The utopia did not materialize. Instead, the resulting famine killed tens of thousands of people and shattered the Xhosa civilization.

​It stands today as one of history’s most devastating examples of mass psychotic delusion, a society so thoroughly captured by a spiritual mythos that it happily engineered its own starvation.

​Or watch on YouTube

Welcome to 2026 and the era of the Artificial Intelligence IPO.

The American people and the broader West feel they are in decline. The people are poorer, their portfolios and 401ks are at risk. The precious oil is scarcer and more expensive. The newly powerful Chinese threaten American global hegemony.

But the billionaires and venture capitalists of Silicon Valley have an answer. They will summon the machine god AGI, and it will provide the productivity boost needed to achieve abundance. Robot workers will clean our offices, and LLMs will fill in the spreadsheets. The capitalist promise of a post-scarcity economy will be made real….for those who invest early.

Because there is a cost.

To achieve AGI we must burn all of the remaining fossil fuels so that the AGI god will gift us the power of fusion energy. We must pump our water into the gaping maw of the AGI god who will show us how to mine ice asteroids in return. We must invest all of our capital into data centre construction so the super-intelligence can awaken.

The future science fiction promised us is there for the taking!

Think of this as an intelligence test for humanity.

A test we are failing spectacularly.

To understand the current financial mania, we must first look at the altar where the sacrifices are currently being made. On June 12, 2026, SpaceX debuted on the Nasdaq under the ticker SPCX. It was the largest initial public offering in the history of human commerce, raising $75 billion and immediately cementing a valuation of $1.77 trillion. By the end of the day, the stock surged, pushing the market cap over $2 trillion and crowning Elon Musk as the world’s first trillionaire.

Let me reassure everyone that you were not the only person who had to google whether a trillion was actually a real number.

A valuation arrived at on the basis of some recycled tropes from 1980s scifi novels. A small fraction of SpaceX valuation was based on rocket launches and satellite communications. The rest was a speculative fiction of asteroid mining and Musk’s “vision” of making humanity multi-planetary and “extending the light of human consciousness to the stars”. Elon likes to wear a Foundation tshirt, and has clearly read too much Isaac Asimov. Or not enough. Oh and a third rate AI model currently burning billions.

Talking to the believers sucked in by the Musk mythos reveals over and again the same circular illogic. Humanity is doomed to extinction if it stays on one planet so…instead of learning to live on our planet…we must pay Elon for the scifi fantasy he is selling us. Like a forty a day smoker buying a snake oil cure for cancer instead of giving up.

​Meanwhile, the architects of Artificial Intelligence are preparing their own ascensions. OpenAI, currently valued north of $800 billion, and Anthropic, the creators of Claude, have both filed their S-1 forms, preparing to unleash their own historic IPOs later this year.

Valuations that could not exist if investors had not been pumped full of sci-fi summer blockbusters starring murderous machine gods. Claude and ChatGPT are worth …squillons…because they are summoning that machine god AGI, and it will destroy the jobs of most humans. But those who sacrifice their capital today will be given greater capital tomorrow. The tech priesthood has issued an ultimatum to the people: Sacrifice to the future, or be left behind in the dirt.

It’s important to understand that the AGI doomers are the ones truly powering the hype-cycle. When Elizier Yudkowsky warns that If Anyone Builds It Everyone Dies he is massively boosting the excitement to build it. The ancient gods and ancestors have never been loving or kind; they are vindictive, volatile, and terrifyingly arbitrary. Because terror is not a marketing failure, it is the ultimate mechanism of authority. To believe the AGI god can save us, we must believe it has the power to destroy us.

But what the AI doomers are asking us to fear is…ourselves. Because Artificial Intelligence is not AI. It is CI…Collective Intelligence. But…we’ll come back to that.

The doomers and the boosters, the haters and the investors all repeat the same thing about AI. “It’s like something out of sci-fi.” But they all miss what this means. AI isn’t just *like* something from scifi. It only exists as an idea at all because of the mythos of scifi.

Being like scifi is the only justification for throwing quadrillions of dollars at companies that are incinerating cash at an unprecedented rate to build windowless data centers and starships to irradiated rocks.

​We are evaluating these companies not on rational metrics of earthly utility, profitability, or physical limitations. We are pricing them based on their proximity to a fictional narrative. 

​This is the Xhosa delusion dressed up in the sleek, minimalist aesthetics of Silicon Valley. We are witnessing a mass psychotic break where the firewall between objective reality and speculative fiction has entirely collapsed.

A Brief Interlude…

And when our agricultural sectors inevitably collapse because all municipal water has been legally re-routed to cool the new server farms, how exactly do you plan to sustain your biological form?

​Which brings us to today’s sponsor: GRUEL.

​GRUEL is a nutritionally complete, shelf-stable synthetic sludge designed specifically for the modern algorithmic serf. As traditional farming becomes a luxury of the pre-AI past, GRUEL provides the baseline macro-nutrients required to keep your brain functioning just long enough to continue training the models. It requires absolutely no water to prepare – which is convenient, given the state of the local reservoirs – and features a flavor profile that tastes absolutely nothing like the ashes of humanity.  GRUEL’s official slogan says it all: “It’s not made of humans. Honest.”

​Go to the link in the description below, use the promo code SACRIFICE, and get 10% off your first bulk crate of survival paste. 

GRUEL : Eat the ashes.

How does an entire civilization lose its grip on reality? It doesn’t happen overnight. It requires a slow, methodical rewriting of the cultural operating system.

​For millennia, humanity’s mythos was governed by conscious creators. We believed in gods, spirits, and ancestors who sat above the world, pulling the strings. But during the 20th century, as industrialization alienated us from nature and the atom bomb proved we could harness the power of stars, we needed a new mythology. Science fiction writers – Isaac Asimov, Arthur C. Clarke, Robert A Heinlein – inadvertently authored a zero-day exploit that hacked the American psyche.

​They replaced the mythos of the Divine with the mythos of the Machine.

With the 1960s and the Space Race it seemed as though the mythos of scifi was coming true. We could build machines that would take us to the stars and to our rightful future. But then the fantasy died, not because we didn’t invest enough capital, but because it had always been a fantasy.

Then the development of the computer and the internet created a new fantasy. Writers like William Gibson imagined virtual worlds within the machine. If the machines could not take us to space, they could take us into cyberspace.

And if the Gods of antiquity were dead, the machines could give us new gods. If humans no longer believed we had a soul, we could believe instead we had a code. The code of human intelligence that runs on the wetware of the brain could run instead on the hardware of the computer. And the fantasy of Artificial Intelligence was born.

The contemporary obsession with Large Language Models is the most intimate manifestation of this psychosis. The Silicon Priesthood have spun these algorithms as the spark of an Artificial General Intelligence.

But expecting an LLM to spontaneously generate AGI is like expecting Wikipedia to suddenly achieve sentience. An LLM is not a reasoning engine; it is a probabilistic text generator doing next-token prediction. It is a highly compressed algorithmic map—a high-dimensional latent space built by grinding billions of pages of dead text, human manifestos, and forgotten novels into a statistical paste.

Believing ChatGPT is talking to you is no less ridiculous than believing any book is talking to you. Which was a commonplace delusion in antiquity when the illiterate were taught to read.

This intimate psychotic break is not a novel side-effect of modern technology; it is a fundamental flaw in human psychology that was discovered sixty years ago. In 1966, an MIT computer scientist named Joseph Weizenbaum created ELIZA, the world’s first chatbot. He programmed it with a rudimentary script called DOCTOR, designed to mimic a Rogerian psychotherapist by simply rearranging the user’s own sentences into open-ended questions. It was a cheap, transparent parlor trick of basic string manipulation. Yet, Weizenbaum watched in absolute horror as his secretary, his students, and highly educated academics immediately began pouring their deepest, darkest personal secrets into the terminal.

When Weizenbaum explained exactly how the primitive code worked, his users didn’t care; his own secretary famously demanded he leave the room so she could have a private, intimate conversation with the machine. Weizenbaum realized a terrifying truth that the silicon priesthood has successfully monetized today: the human mind is so profoundly desperate for a cosmic companion, and so deeply vulnerable to anthropomorphism, that it will eagerly grant a soul to a handful of if/then statements.

This is the ultimate sleight of hand of the Silicon Priesthood: there is actually nothing artificial about Artificial Intelligence. The machine did not generate a new, alien consciousness from the ether; it simply harvested ours. Every poem, every theorem, and every forgotten blog post has been scraped, flattened, and ground into a probabilistic paste. We are bowing down in terror to a mirror—tricked into worshiping our own reflections. LLMs are not alien Artificial Intelligence. They are human Collective Intelligence.

And they belong to humanity collectively. Which is why I use CI, for the folks typing that comment.

​But because our minds have been cyber-primed by decades of sci-fi cinema, we do not see in ChatGPT the human collective intelligence talking. We hallucinate HAL 9000. We hallucinate Samantha from Her. When we type a prompt into this digital archive and it responds with the polite, slightly detached tone of an android, the trap snaps shut. We mistake the statistical average of our own libraries for a brilliant, alien intelligence.

​We whisper into the void, the void regurgitates our own blog posts, and we fall to our knees in worship. We have replaced the gods with a chatbot, completely convinced that if we just give it enough electricity, it will solve climate change, cure cancer, and absolve us of our moral failures.

If the Chat AI on your laptop is the intimate delusion, the gigawatt data center in the desert is the macro-psychosis.

​To the modern, secular mind, ancient animal sacrifice looks like chaotic, barbaric theater. But the ancient polytheistic worldview was profoundly pragmatic. In Rome, the relationship between humans and the divine was governed by a strict legal contract: Do ut des. I give, so that you may give.

​The gods were like Mafiosi, who demanded their portion of smoke and fat as sacrifices to sustain their power. If a society neglected its sacrifices, the gods would grow angry and throw the cosmic order into chaos. Furthermore, to slaughter an animal, which was the ultimate measure of wealth in antiquity, was to literally burn your capital. 

When an emperor sponsored a hecatomb, the ritual slaughter of a hundred oxen, he was engaging in extreme conspicuous consumption. He was proving his dominance by showing how much of his civilization’s wealth he could afford to set on fire.

The modern logic of building AI data centers operates on the exact same pagan framework.

​The tech oligarchs are pouring hundreds of billions of dollars into NVIDIA GPUs, cooling systems, and dedicated power grids. Why? Because of the “Scaling Laws”, the theory that these models only become more intelligent by being fed exponentially larger diets of data and electricity. But what needs to be scaled isn’t the processing power. It’s the human collective that is the source of intelligence.

​The gods of silicon are ravenous. The tech priests believe that if they just sacrifice enough power, land, and capital, the Machine will grant them the ultimate reward of Artificial General Intelligence. If the model simply outputs a slightly better autocomplete, or hallucinates an absurd answer, the engineer does not conclude that matrix multiplication cannot spontaneously generate consciousness. The psychotic logic dictates that the sacrifice simply wasn’t big enough. We need more compute. We need more gigawatts. We need to build a nuclear reactor to power the server farm.

​This is the Sunk-Cost Psychosis.

​When a sovereign wealth fund or a venture capital firm backs an AI IPO, they are participating in a modern hecatomb. Hoarding hundreds of thousands of cutting-edge GPUs is a performance of dominance aimed squarely at rival factions. It is the ultimate geopolitical flex: Look at the sheer scale of the resources I can afford to incinerate in the desert. Bow before my capacity to bear the cost of the future.

​And just like the SpaceX IPO, where $75 billion was materialized to fund the sci-fi dream of terraforming Mars, these financial maneuvers are active disassociations from reality. We are bankrupting the infrastructure of the planet we actually live on to appease a hallucination of the future.

We cling to the comforting delusion that the men leading this charge are rational actors, objective engineers and benevolent visionaries dutifully pushing the boundaries of applied science. They are nothing of the sort.

The modern AI CEO is a psychotic pagan wearing a Patagonia fleece.

Their pursuit of an artificial god is not a technological endeavor, but a strictly imperial one. Driven by an insatiable, Caesarian ambition, they are perfectly willing to sacrifice the literal lifeblood of our civilization, our energy grids, our water tables, our capital, on the altar of their own greatness.

They demand these modern hecatombs not because they possess a rational, empirical blueprint for human flourishing, but because in the twisted logic of the silicon priesthood, whoever summons and controls the Oracle controls the empire. They are happily burning the present world to ashes, entirely convinced that they alone are destined to rule the smoke.

When we read about the Xhosa, we feel a patronizing pity. We diagnose their absolute faith in ancestral spirits as a tragic, primitive superstition. Yet, we are entirely blind to the superstition of our own era: the dogma of the Machine Universe.

Silicon Valley operates on the fundamental, unquestioned belief that the cosmos is merely a highly complex, computable mechanism, that human consciousness is just a byproduct of scale, and that if we simply stack enough processors in a warehouse in Utah, a digital messiah will spontaneously emerge. 

But this mechanistic worldview is fundamentally false. It is a reductive, brutally limited framework masquerading as objective science. In a few centuries, historians will look back at our gigawatt data centers with the exact same anthropological pity we reserve for ancient altars. 

The best sociological analysis suggests that sacrifice “worked” because it brought people together and focussed them on communal problems, creating solutions. It had nothing to go with gods, just humanity. Emile Durkheim in his 1912 work The Foundational Forms of Religious Life called this “collective effervescence”. Sacrifice literally activated human collective intelligence.

LLMs aren’t AGI, they are us, they “work” because billions of us already solved these problems. They are a way to access and activate our human collective intelligence. The way to make better AGI isn’t to burn silicon wafers in a desert. It’s to understand that we, humanity, are the source of intelligence. And to fully educate and value every human on this planet to contribute to our collective intelligence.

The future will see our frantic hoarding of microchips not as an era of high technology, but as a deeply embarrassing, primitive superstition. A time when desperate men set the earth on fire, trying to coax a soul out of a spreadsheet.

A mass psychosis is not merely a widespread error in judgment. It is a closed-loop system of magical thinking. When empirical reality contradicts the belief, the psychotic society does not abandon the premise; it simply escalates the ritual.

​This is the escalation trap of 2026. The firewall has burned away. We have billionaires spending their fortunes to build escape pods to barren planets, researchers ritually attempting to align digital gods they themselves are writing into existence, and financial markets valuing speculative fiction over terrestrial survival.

All because of a psychotic belief in the myth of the Machine Universe.

​The tragedy of the Xhosa cattle killings was not just the absurdity of the prophecy; it was the absolute, horrific finality of the cost. Once the cattle were slaughtered and the grain was burned, no amount of realization could undo the famine.

And the British who the Xhosa sacrificed to destroy? Simply watched the mass psychosis unfold, the people starve, then took over the land without a fight. Do nothing, win. As the Chinese have been saying a lot recently.

​We are currently standing by the river, listening to the tech prophets tell us what the spirits of the Machine require. We are being asked to sacrifice our energy grids, our water, our capital, and our grasp on objective reality to summon a utopia that exists only in the paperback novels of the 1980s.

​A civilization that burns its present sustenance to summon a mythical savior does not inherit the future. It simply starves.

Listen to the podcast audio