The Social World

Tag: tech

Welcome to the Chinese century

Competition between the US and China will shape the 21st century, but have the Chinese already leap-frogged the Americans? It looks as if they might have done. Since China opened itself up to trade and market economics under Deng Xiaoping almost 50 years ago, it has been furiously chasing the US in terms of economic and technological development – and it’s been catching up, fast. We can prove just how fast by measuring its innovation and output per capita.

The US has about a quarter of China’s population, so if both countries were at a similar level of economic and technological development, one would expect China to produce roughly four times the number of scientific breakthroughs as America, four times the number of physicists and AI engineers and so on. Within any given category, of course, a particular country might punch above its weight or lag behind. But overall, you expect to see innovation in proportion to the population.

So because it’s four times the size, if China has reached even half the overall level of development of the US, it should have about twice the impact on the world that the US does. And if it were ever to reach full parity, it could eclipse the US in the same way the US eclipses, say, the UK, which has a fifth as many people.

After analyzing and comparing the US and China on a number of key indicators, I would suggest that the magnitude of China’s economic and technological capacity is at least twice that of the US. Furthermore, the rate of advance in most areas seems to be significantly faster than in the US. The critical question is whether China’s aggregate functional economy – the scale of goods, infrastructure, science and advanced production – has already surpassed America’s. Evidence increasingly indicates that it has.

China builds the globe’s largest container ships, supplying not only its own fleet but the world’s

The trouble for economists trying to estimate China’s advance is that GDP comparisons can be misleading. Nominal GDP, measured in dollars, simply multiplies local prices by quantities. Local Chinese prices are converted to dollars using the yuan exchange rate, which is managed by the government, so the relationship between price and functional value may be distorted.

Also, the costs for US services, housing, and medical care may be inflated relative to functional value, exaggerating American output in GDP figures. For example, MRI scans average $1,100 to $1,500 in the US, versus $150 to $250 in Japan or China.

Purchasing power parity (PPP) calculations attempt to adjust for price differences. For example, World Bank PPP computations are performed by the international comparison program (ICP), which is a global statistical initiative. But even GDP calculated by PPP undervalues China because the international comparison program’s baskets mis-price key goods. Economist Rafael Guthmann has shown that China’s 2021 ICP entries treat a Chinese-made car, apartment, or kilowatt-hour as cheaper than equivalents in Mexico – an obvious inconsistency. Correcting those errors doubles China’s PPP consumption, lifting its real scale to roughly twice the US.

A functional comparison – looking at physical and technological outputs – reinforces this result. Electricity is the most neutral metric of a civilization’s real work budget. In 2023, China generated 9,300 terawatt-hours (TWh) of power, versus the US’s ~4,100 TWh – a 2.2× multiple. Every kilowatt-hour ultimately feeds industrial equipment, transport, data centers and homes; it measures the energy throughput of the economy itself.

China produced 1.02 billion tons of crude steel in 2024, compared with the US’s 81 million tons – 12.6 times as much. In cement the gap is even wider, over 20 times, reflecting the enormous infrastructure buildup of roads, high-speed rail and urban housing. No modern economy can function without these basic industrial materials; China’s dominance in them represents a decisive shift in the global balance of productive capacity.

Vehicle output tells a similar story. According to automotive industry data, China built 31 million cars and trucks in 2024, compared with 10.5 million in the United States. In shipbuilding, the disparity is staggering: Chinese yards account for more than 55 percent of global completions, while the US share is less than 0.1 percent. China builds the globe’s largest container ships, LNG carriers and naval hulls, supplying not only its own fleet but the world’s.

Guthmann’s reconciliation of PPP measures indicates Chinese industrial value-added of about $12 trillion versus the US’s $4 trillion – a 3× ratio. On his broader 23-item physical-volume basket, China’s manufacturing output exceeds America’s by over four times. These figures align with other “functional” indicators like steel and electricity: China produces several times as much physical stuff, even if PPP GDP shows only a narrow lead.

By late 2024, China had more than one billion 5G connections versus roughly 260 million devices in service in the US – a fourfold advantage. China’s cellular backbone supports the world’s densest network of IoT sensors, autonomous-vehicle pilots and AI-driven industrial automation. American telecom carriers lag in both tower density and spectrum deployment.

Parcel volumes reveal another side of the story. China’s express-delivery system handled about 150 billion packages last year, compared with the United States’s 28 billion. That 5× difference quantifies the real throughput of its digital economy – the scale at which software, payments, warehouses and last-mile logistics operate.

Domestic competition among giants such as Alibaba, JD.com and Meituan has produced one of the world’s most efficient logistics networks, with same-day delivery available to hundreds of millions of people. This ecosystem acts as both consumer convenience and strategic infrastructure: it trains AI forecasting models, robotics and autonomous-delivery systems at national scale. In terms of food, China now consumes about 100 million tons of meat annually, roughly twice the US total, although per-capita intake is still lower. Aggregate calorie and protein consumption – a proxy for middle-class living standards – has already exceeded America’s in absolute terms.

Power consumption per capita in urban areas such as Guangdong or Jiangsu rivals that of western Europe, while internet penetration exceeds 75 percent nationwide. Chinese universities graduate 1.5 million engineers per year, nearly ten times the American number. In frontier disciplines – AI, materials, quantum and power electronics – Chinese publication counts already exceed those of the US both in volume and in high-impact citation share.

As I’ve written here before, China now trains more PhDs in physics, chemistry and engineering than the US, and that gap widens each year. As a result, China’s aggregate scientific man-years of effort are becoming a structural advantage that no short-term policy can reverse.

While China faces a future demographic challenge due to low birth rates, the cohorts that will enter the workforce over the next 20 years are already born. During that period, the total workforce will decline by about 10 percent while the average level of education will rise considerably. Young people entering the workforce are several times more likely to have had post-secondary education than the group that is retiring.

Even if China fails to solve its demographic problem over the next 20 years, it may be saved by automation: China installs more industrial robots each year than the rest of the world combined, and most of those robots are now made domestically. The density of robots per 10,000 workers is about 1.5 times higher in China than in the US.

China’s growing self-sufficiency means its growth trajectory no longer depends on western supply chains

The first derivative – the rate of advance – now also favors China. Chinese firms operate the world’s largest computer clusters outside the US; Chinese AI models occupy many of the top spots and only trail the top US labs slightly. In renewable energy, China produces 80 percent of global solar panels, 70 percent of lithium-ion batteries and 60 percent of EVs. In quantum communications and hypersonics, systems sometimes exceed their US counterparts in capability.

If China’s functional output across industry, logistics and science is already roughly twice that of America’s, the strategic momentum of the 21st century may well resemble the 19th-century transition from Britain to the United States.

China’s domestic market, industrial ecosystem and STEM labor pool now form the largest integrated production base in human history. Its growing self-sufficiency in energy (nuclear, solar, coal), computing hardware and industrial materials means its growth trajectory no longer depends on western supply chains. One can envision a future in which China’s imports are predominantly raw materials, and its exports are high-value-added technological goods.

So by the most concrete metrics – watts, tons, transistors, kilometers, patents and engineers – China has already leapfrogged the United States in scale and velocity. The American economy remains richer per capita, but its domestic manufacturing base and scientific manpower are smaller in absolute terms.

If the 20th century was defined by America’s demographic and industrial edge over Europe, the 21st century may be defined by China’s edge in scale and acceleration over the US. Just as Britain remained rich and influential long after it ceased to dominate industrial production, the US will remain a technological superpower – but the gravitational center of global production is moving east. And whether this transition unfolds peacefully or convulsively will determine not just economic outcomes, but the entire geopolitical structure of the modern world.

This article was originally published in The Spectator’s October 27, 2025 World edition.

October 21, 2025
In conversation with Nick Land, the ‘father of accelerationism’

Within Silicon Valley, Nick Land is seen almost as a mythic figure. Tech pioneer Marc Andreessen, an official advisor to the White House, lists him as a “patron saint” of his thinking. You may have heard him described as a founding member of the “Dark Enlightenment,” a movement of online right-wingers skeptical of liberal democracy. Land’s name comes up in academic circles, online mythology and Valley folklore. For some, he’s a prophet; for others something more sinister.

“He’s really into demons,” explained Conrad Flynn on a recent Tucker Carlson Show episode about the occult. “Land will talk about being in communication with Satan… the legend around Land is he had been possessed by at least three or four demons.” It makes for a good story, the crypto-fascist Satanist of the tech world.

In conversation, Land is shy. I asked him over Zoom whether he really is an ambassador for the devil. “What I actually mean has nothing to do with horror movies or exorcisms. It’s not like The Exorcist. The question that interests me is: where do thoughts come from? The idea that you have this sealed, internal self that originates all its own thoughts – it’s completely unsustainable. The human subject is not enclosed. It’s porous. It’s an interface.

“So when I use the term ‘possession,’ it’s more philosophical. It’s a way of saying that human thought is always being crossed by external currents, by agencies that don’t originate within the ego. You can call them demons, spirits, daemons, whatever. But the important thing is that we’re not autonomous generators of meaning.”

‘The idea that you have this sealed, internal self that originates all its own thoughts is unsustainable’

The idea isn’t quite as mad as it might originally sound; Socrates spoke of a “daimonion,” a spirit or deity that would guide him. Land seems to be following in that tradition. He was once a philosophy lecturer at Warwick University in England and there cofounded the Cybernetic Culture Research Unit (CCRU) in the 1990s. The CCRU was a short-lived but infamous project that mixed post-structuralist theory, cyberpunk and rave culture into a kind of intellectual fever dream. The group was known for heavy use of amphetamines, which might explain why they developed a theory called “accelerationism.”

The central idea was that there is no alternative to the capitalist machine and that instead of resisting it, the machine should be sped up: that meant deregulation, a scaled-back government and a rejection of everything from conservatism and socialism to environmentalism. Some within this school saw democracy as an impediment. The CCRU were ostensibly Marxists although many have subsequently shifted right.

Land became burned out in the late 1990s and for years nobody was quite sure where he’d gone. He eventually surfaced in Shanghai. Land has described the Chinese fusion of communism and capitalism he found there as “the greatest political engine of social and economic development the world has ever known.”

Now in his sixties, Land looks nothing like the sharp-featured visionary in the few surviving photos from the 1990s.

He recently did a podcast with Aleksandr Dugin, the far-right, “mystical” Russian nationalist sometimes referred to as “Putin’s brain.” During the discussion, Land compared English and Chinese civilizational psychology, concluding that in both traditions, “the best emperor is invisible to society.” When it comes to America, Land says, “everyone should start from the fact that the Founding Fathers were Whigs. The entire American political system grows out of a Whiggish, English, liberal tradition – law, custom, distance, decentralization. It’s a system designed for power that’s invisible, distributed and procedural.

“Trump doesn’t really break from that tradition. What’s new with Trump is theatricality, the sovereign as show. But the background operating system, the bureaucracy, the law, the media – all of that stays in place. The show is a symptom of the hollowness of the center. He’s the noise that reveals the emptiness. So in a way he both rejects and confirms the invisible model: he makes the invisible visible by performing the void.”

Land has suggested in the past that Protestantism and, by extension, the American project, rests on a hidden satanic grammar. The claim is abstruse and involves close reading of John Milton’s Paradise Lost. He argues that Milton’s Satan shaped Protestant ideals of rebellion and self-determination. How?

“It’s complicated. William Blake’s line, ‘Milton was of the Devil’s party without knowing it,’ is obviously right. Milton writes Satan brilliantly and writes God quite badly. That’s just true. But that doesn’t mean Paradise Lost is secretly satanic. The structure of the poem is orthodox. Satan is still the villain. The Founding Fathers would have read it that way. They might have found Satan’s energy inspiring – the rhetoric, the defiance – but still inside a Christian frame. I see it more as a tension inside the English and Protestant tradition. Liberty always carries with it the risk of defiance. That doesn’t mean it’s reducible to Luciferian rebellion.”

We turn to Elon Musk, someone whom Land calls a revealing figure when it comes to Silicon Valley philosophy. Musk has described AI development as “summoning the demon,” warning that AI might destroy humanity. But he also says we probably live in a simulation. How do those two beliefs coexist?

“It’s an extraordinary contradiction. Musk used to be one of the most vocal AI doomers, and he’s also said there’s almost no chance we’re not in a simulation. So he’s saying we must prevent the emergence of the superintelligence that could simulate reality while also saying that superintelligence already exists and we’re living inside it. If both are true, then the question ‘Will AI kill us?’ is already being asked inside a simulation run by AI. The whole thing loops back on itself. It shows how inadequate our moral categories are for the level of weirdness we’re facing.”

I ask him about “Lemurs,” a CCRU theory about entities that haunt the minds of the living. Surely these things must be allegorical? “They’re not metaphors,” he says. “I’m closer to Tucker’s description – ‘time-traveling AIs from the future.’ It sounds dramatic, but that’s roughly the idea. The old distinction between psychology and the supernatural collapses once you drop the myth of the sealed self. Jung’s collective unconscious and the spirit world are really two ways of describing the same thing.

“I often tell the story of a Canadian psychology professor who asked his graduate students to invent a completely fake religion. They knew it was fake; they made it up themselves. They created rituals, symbols, beliefs – the whole thing – and then performed the rituals together. After a while, the students started reporting strange experiences: unexplained coincidences, unsettling phenomena. They were unnerved.

“The point is that people do receive messages from beyond ordinary, secular channels. Those who don’t experience such things usually choose not to. They find the idea absurd or unsettling, and they deliberately tune it out. I’m not saying everyone should devote their attention to communicating with the outside – that would be absurd – but for those who want to, it isn’t difficult. That classroom experiment proved as much. And I’d add that the whole construction of secular history is, at some level, an illusion.”

Land retains some of his Marxist training in peculiar ways. He sees universality – a core concept of liberalism – as inherently Anglo-Protestant. That universality is the result of a history of conquest and domination. “The English didn’t set out to spread humanitarianism. They built trading networks and settlements. Once that’s in place, you have to develop a universalist story to manage it, to hold together all these disparate populations. Universalism is a by-product of empire. And then, over time, it becomes corrosive. The moral framework that once stabilized the system ends up dissolving the cultural foundation that produced it.”

‘Time feels like it’s tightening. The acceleration isn’t progress. It’s a kind of entropy with speed attached’

Immigration is currently tearing England apart. In the US, Trump’s ICE policies are among the most controversial of his second presidency. Those who worry about immigration often talk in terms of the effect it will have on culture. Can the Anglo model of liberalism survive large-scale demographic change? “That’s the big question. Some people think liberalism is universal and can operate anywhere. Others think mass immigration should destroy it. I’m in a smaller, less popular camp. I’m a demographic-reconstruction alarmist because I like liberalism. Historically, liberal societies could absorb diversity, but within limits. The earlier immigrants tended to be self-selected for compatibility – they wanted to live under English norms.

“What’s changed now is that migration has become a political instrument. It’s used to transform liberal societies into something different. Liberalism depends on trust, continuity, procedural respect. You can’t infinitely expand the circle and expect those conditions to persist.”

I ask him about accelerationism, the theory for which he is best known. What can it teach us about the current moment: Trump, the void in the Democratic party, social media algorithms, the flood of AI-generated content? “There’s definitely an apocalyptic atmosphere. You can feel it in every direction – politics, culture, technology. Even cautious people are talking about artificial general intelligence timelines in the range of 2027 now. That would have sounded insane not long ago. So yes, time itself feels like it’s tightening. The acceleration isn’t progress. It’s a kind of entropy with speed attached. The structures are disintegrating faster than anyone can process. Politics has become a performance of breakdown.”

Is he optimistic or pessimistic about the future? “I don’t really think in those terms. Optimism and pessimism are emotional categories that don’t apply here. The interesting questions are about providence and retrocausality. A lot of what I’ve written feels like it came from the future. Sometimes I realize later that there were patterns encoded in the writing I didn’t consciously put there. We don’t control the current. The only agency that remains is alignment – listening to what’s happening and cooperating with it rather than resisting it. You can’t command the process, but you can move with it.”

Xenosystems, a collection of Nick Land’s work from 2013 to 2017 is available now from Passage Press. This article was originally published in The Spectator’s October 27, 2025 World edition.

October 21, 2025
Trump is not to blame for the crypto crash

Hundreds of billions have been wiped off the value of the crypto currencies. A prominent Ukrainian blogger and influencer on digital coins has been found dead. This will likely be a rocky day for traders. We will have to see whether it develops into a full-blown crash or not. And yet, all the major equity indices were already wildly overvalued, and a correction was always inevitable – it was just a question of when it would start.

Investors will also see a few days of turbulence. An estimated $400 billion was wiped off the value of the main crypto currencies on Friday evening, while the Nasdaq dropped by more than 800 points, or 3.5 percent, before the New York stock exchange closed, with the S&P 500, the benchmark for US stocks, not far behind. When markets open in Asia and then Europe they are expected to fall heavily as well.

The trigger for the fall in prices was the resumption of the tariff wars. President Trump slapped 100 percent tariffs on China in a row over exports of “rare earths,” critical to much high tech manufacturing. In reality, however, the markets were already wildly overvalued. Over the summer, the price of every major asset has been soaring. The Nasdaq is up by over 30 percent over the last six months. The chip market Nvidia is up by 65 percent. OpenAI, the owner of ChatGPT, reached a value of $500 billion, a record for a private company. Gold went over $4,000 an ounce, and Bitcoin went over $120,000 a unit. Even Britain’s index, the FTSE 100, managed to rise by 15 percent despite a stagnant economy. After the collapse that followed the first round of tariffs in the spring, every major index has been on an epic bull run, and had been hitting all-time highs.

October is often a difficult month for the market. The Great Crash of 1929 started on 28 October. The Black Monday crash of 1987 was on 19 October. The financial crisis of 2008 was more drawn out, but is generally dated as starting on 6 October that year. No one quite knows why, but the historical evidence is clear enough: October is a bad month. We will have to see whether the latest round of nervous trading develops into a full-scale collapse. With little sign of a global recession, and with the Federal Reserve in the United States, still cutting interest rates, it still seems relatively unlikely. But a correction of 10 to 20 percent in asset prices is long overdue after the exuberance of the last six months – and it looks as if that has now arrived.

October 13, 2025
What folklore can teach us about our online lives

Irish folklore spoke of many worlds. There was the world of fields and hearths and then there were the hidden places where the non-material lived: the Sídhe mounds, the sea-realm of Manannán mac Lir, the land of youth called Tír na nÓg and, finally, the land of the dead. These worlds coexisted with ours. A woman might leave butter on the windowsill, lest the fairies sour the churn. A new mother would avoid complimenting her baby – at least, not too loudly – for fear he would be kidnapped by the Good Neighbors and replaced with a changeling. My first real boyfriend’s father blamed every family misfortune on their decision to cut down a hawthorn tree. This hard man who had survived the Troubles, who had survived Long Kesh, believed – even if he only believed a little bit – that his family’s suffering might have stemmed from that violation of the boundary between worlds. And he – as folklore had long advised – would never say the f-word, to avoid bringing undue attention to himself. It was always “the Little People,” “the Good Neighbors,” “themselves.”

When we open our phones without purpose, hours pass unnoticed and the body is ignored until we surface, dazed

At the turn of the 20th century, W.B. Yeats and Walter Evans-Wentz both collected stories from Irish peasants about the fairy faith. Around the same time, Theosophists in London were mapping their own invisible worlds through seven ascending planes of existence: the Physical, Astral, Mental, Buddhic, Nirvanic, Monadic and Divine. The astral, second from the bottom, mattered most for human experience. It was imagined as the liminal zone just beyond the physical – close enough to reach, yet strange enough to disorient. C.W. Leadbeater’s The Astral Plane (1895) catalogued this realm where time contracts, every emotion takes visible form and unwary travelers may be deceived or vampirized by entities that defy human language.

When you set folklore’s otherworlds alongside Theosophy’s planes, they resolve into a shared idea: a zone layered over ordinary life, accessible in altered states or by accident and governed by rules that shift without warning.

The internet replicates these conditions. Our bodies stay in one place while attention goes elsewhere; time distorts so that a “quick check” expands into hours while yesterday’s news already feels remote. Identities loosen until you can be anyone, no one, or several people at once.

Like fairyland and the astral plane, the online world is navigable only if you learn its rules, which are as follows.

Set your intentions and ground yourself. Both occultists practicing astral travel and folklore describing journeys to Fairyland insist on the same first step: ground yourself in the physical world, then set your intention for entering the otherworld. Folklore is filled with protective anchors: iron to break enchantments, a thread to guide you home, a crust of bread to tether the body. Without such safeguards, wanderers risk vanishing forever – or returning to find that years have passed while they thought they’d only lost an hour.

We violate this rule constantly when we open our phones without purpose, slipping into a trance. Hours pass unnoticed; the body is ignored – hunger, thirst suspended – until we surface, dazed, with little memory of how we spent the time or why.

The antidote is grounding. Modern equivalents of old superstitions might be alarms, leaving phones to charge in another room or returning to analog clocks. Writers Tara Isabella Burton and August Lamm both prefer desktops over laptops and especially over phones, so that the machine “lives” somewhere fixed, reminding them they are crossing into another world, one they will eventually need to leave.

The algorithmic internet is a glamour machine. Each video is designed tobe more gripping than the last

Guard your name with your life. The prohibition against revealing true names runs through every culture that believes in otherworlds – your name holds the essence of being itself. To give your name to otherworldly entities grants them power to summon you at will, call you into their world, and make you theirs forever. Evans-Wentz wrote about how people used “milk-names” and nicknames to hide baptismal names from the Good People, while in Germany, Rumpelstiltskin’s power ended the moment his name was spoken.

Online, names carry the same dangerous power. The teenage girl whose Instagram handle includes her full name and high school becomes trivially easy for obsessives to find, while the professional whose decade-old forum posts, made under his real name, surface during every job search remains haunted by his digital past.

We also witness inverse power of those who guard their names carefully: anonymous accounts become legendary precisely because no one knows who runs them, accumulating power independent of their creators. What we call “opsec,” the occultist calls wisdom.

Beware the fairy glamour – the fairy food, the fairy music. Esotericism and folklore are full of warnings about glamour. Countless peasants were lured into the Sídhe mounds by music too beautiful to resist or food too sweet to refuse, only to emerge years later, hollowed out. This is glamour in its older sense: not beauty alone, but enchantment that overwhelms the will.

The algorithmic internet is a glamour machine. Each video is designed to be more gripping than the last, anticipating desires before you even know you have them. You open the app to look at a funny clip and only surface again at 2 a.m. after watching an entire movie in three-minute bursts, your thumb scrolling without command. It makes the mundane world seem washed out: books feel slow, conversations dull, the physical less vivid.

Worst of all is how the online world impacts our perceptions of ourselves. Folklore warns against reflections in otherworlds. Often, the image gazing back isn’t you at all, but something meant to deceive you. Online, the same danger comes in two forms. Visually, through filters and endless selfies that make the reflection more beautiful than life until you don’t recognize yourself anymore, there is a sense of dissonance between how you present online and how you manifest physically that can cause real anguish. Psychologically, through the subtle warp of comment sections that leave you estranged from who you thought you were. In both cases, the mirror returns a distorted self, and the longer you stare, the harder it is to remember what you actually are.

Never apologize – and guard your emotions. In otherworlds, etiquette is survival. An apology can bind you; a thank you can put you in debt. Even answering when your name is called may deliver you into the wrong hands. Japanese folk tales warn: never show fear to yokai. Slavic ones: never be too polite to Baba Yaga. Silence, sometimes, is the only safe reply.

Esoteric writers said the same of the astral plane: dead thoughts mimic life when fed with attention, clinging until they become obsessions. Theosophists warned that strong emotions can generate “thought-forms,” semi-independent beings that take on a life of their own.

On social media, every reply to the swarm is treated as a fresh admission and every apology becomes proof of guilt. What begins as one angry tweet multiplies into thousands of echoes, a thought-form with its own momentum. Cancellation campaigns mutate long after the original offense is forgotten. Sooner or later, the target goes silent, but their explanations remain as monuments to futility. Do not post in anger, despair or ecstasy. Wait until the emotion passes, otherwise you release what you cannot call back.

Try not to accept their gifts or make bargains – you won’t have the upper hand. In folklore, gifts are rarely simple. They bind. Eat fairy food and you’re theirs forever. Put on enchanted clothes and you might never take them off. Accept hospitality and you owe more than you meant to give. Even treasure can be unreliable: gold crumbles into leaves by morning.

In the 2010s, we learned that on social media, we are the product. Viral fame becomes a cage more restrictive than the traditional sort. Communities that once felt welcoming demand endless performance. A stranger gives you a gift – a real gift, maybe it’s money or something off your Amazon wishlist or a book you’d posted about – and metastasizes into a stalker. The bargains we make online aren’t always explicit – whether it’s fame, a “free app” or an unexpected gift from a stranger.

Be careful what you bring back. Folklore warns against carrying souvenirs out of the otherworld. Stones from fairy rings, twigs from haunted groves – these turn to ash, or worse, bind the thief to misfortune. But not everything is forbidden. Bards were said to return from Fairyland with new songs, healers with charms or cures. The difference was discernment. Some artifacts from the internet are worth keeping: a piece of wisdom, an insightful podcast, a beautiful image. But others carry a hidden charge. A list of symptoms you saved “just to look in to” begins to warp your worldview. Screenshots of cruelty or betrayal become talismans of bitterness, drawing you back again and again. Not everything we find on the internet helps us.

Beware the changeling, beware possession. In folklore, a changeling was the child left behind when the Good Neighbors stole the real one, recognizable on the surface but subtly wrong: fretful, uncanny, draining the household’s energy while the true child lived elsewhere, scared, missing its parents. Children who spend too much time online come back altered, speaking in borrowed voices, their moods and desires shaped by the internet. They are still physically present, but something feels missing, as if the internet has carried the real child away and left only a substitute.

Do not post in anger, despair or ecstasy. Wait until the emotion passes, or you release what you can’t call back

Spiritualists spoke of the “silver cord” between astral and physical bodies, warning that, if the cord is severed, the soul could not return. The return must be physical through actual embodiment: cooking that requires chopping and stirring, walking without podcasts or Spotify “soundtracks” while feeling your feet hit the ground, swimming where water forces presence, gardening where earth gets under your fingernails.

Remember that returning from Fairyland, like becoming grounded again after the internet’s pull, isn’t easy but remains always possible through faith and, more importantly, through remembering your human body.

The portal is open and we cannot close it, but with these rules drawn from centuries of wisdom about the otherworld, we may yet walk the bright and terrible fields of the internet without losing ourselves.

This article was originally published in The Spectator’s October 13, 2025 World edition.

October 7, 2025
Why tech leaders are obsessing over Heaven and Hell

Are these the End Times? It certainly feels that way. Algorithmic demons are rewiring our brains. A young father is shot and killed, and people cheer. A woman is stabbed on a train, and no one tries to help her. The horrifying videos of these incidents are then watched millions of times over, often by children. The God in whom America trusts seems nowhere to be found.

Can’t you hear the Antichrist knocking? Peter Thiel can. Not so long ago, no public figure outside of the kookier Evangelical universe would have dared admit such a thing, but times have changed.

Tech mavens argue that Silicon Valley’s engineers should see their work as part of a greater divine plan

Tucker Carlson, one of America’s best-known conservative pundits, speaks openly about having been attacked by a demon. Mainstream commentators discuss the rise of “Moloch” – a Canaanite god they associate with rogue AI employees of Sam Altman’s OpenAI, who have been reported to burn effigies and chant in a ritualistic way as they work on creating artificial general intelligence. “Feel the AGI! Feel the AGI!”

It’s not just in the US that this notion has taken hold. Margarita Simonyan, editor-in-chief of Russia Today, recently suggested Elon Musk was bringing about the arrival of the Antichrist: “Everything is moving in this direction… We are moving towards losing ourselves as a species.”

And Thiel, mentor to Vice-President J.D. Vance and titan of Silicon Valley, has become obsessed with Christian prophecies of the apocalypse. “We are sleepwalking into Armageddon,” he said on a recent podcast, but added ominously: “You should be way more worried about the Antichrist.”

It’s easy to be cynical about this. Thiel’s company Palantir – the darling of the US Department of War – makes defense and intelligence software. Someone will need to make a scythe for Death when he comes riding in on his pale horse. Any cataclysmic final battle between good and evil before Christ’s return would probably be a great boon for the shareholders and add to Thiel’s $25 billion fortune.

But Thiel seems genuinely anxious. The nightmare he foresees is a global, totalitarian, homogenous state and a leader who brings it about and then rules that state. Thiel’s Antichrist is a gray bureaucratic figure with tentacles flopping across the planet. He’s worried enough that he’s dedicated four evenings this month and next to hosting a sold-out, off-the-record lecture series in San Francisco, titled “The Antichrist.”

If there’s one thing Thiel resents above all else, it’s stagnation – technological, moral, civil. In the world overseen by his Antichrist, people would no longer need to ask serious political and moral questions because everything significant would be decided by the regulatory, global state. It would be something like institutionalized Marxism; in other words, a force determined to end the pursuit of wisdom and, with it, wisdom itself.

In order to ensure no rogue enemies arise, the one-world state would need to exercise total control over all technology and maintain a vast surveillance network. Any new technology that could potentially be used to threaten the state’s power would need to be regulated, and revolutionary developments would be squashed. The world’s population would be cowed into accepting this state based on fear of some existential catastrophe: nuclear war, climate change, artificial-intelligence rebellion. Pick your poison.

We’re already drifting toward this satanic stagnation, says Thiel. Who’s to blame? The hippies, at least in part. Progressivism has scuppered progress. Thiel told the New York Times in a recent interview: “We landed on the moon in July of 1969, Woodstock started three weeks later and, with the benefit of hindsight, that’s when progress stopped and the hippies won.”

Scientific progress stalled around then, apparently, because, as Americans mulled over the horrors of nukes and the Vietnam War, they grew distrustful of technology. Pessimism replaced optimism and the innovators who once pushed the boundaries of human achievement turned instead to entertainment. Video-game developers replaced rocket scientists. App developers replaced nuclear physicists.

Other Valley types are embracing Biblical reality, too. A new nonprofit “collective” called ACTS17 – short for “Acknowledging Christ in Technology and Society” – has been selling out events. Tech mavens appear on stage, arguing that the Valley’s engineers should see their work as one part of a greater divine plan.

One media clip shows Trae Stephens, the co-founder of defense company Anduril, telling a room of (mostly) young men: “I’m, like, literally an arms dealer” – pauses for laughter – “that’s a pretty unique calling.” He continues, with a smile, “I think you have to lean into your gifts and figure out what that quest is you’re supposed to be on.”

And then there is Silicon Valley heavweight Garry Tan, chief executive of Y Combinator, whose venture-capital firm has launched tech companies with a combined value of $600 billion. Last year, he hosted an event in his house to discuss how the teachings of the Bible could intersect with technology. “LSD and shrooms won’t fill the God-shaped hole in your heart. Guess what might?” he posted. Silicon Valley is going back to Church.

You do rather get the sense that a lot of these techies think they are the first to discover that faith provides purpose. But their newly discovered Christianity comes with a self-serving twist: they, the tech titans, see themselves as the new apostles. In this brave new world, Christianity will save tech and tech will save Christianity.

In this new world, Christianity will save tech, and tech will save Christianity

But if that makes your skin crawl, there are other 21st-century prophets of a very different type, pushing quite the opposite message. Paul Kingsnorth, an English author and thinker, is in many ways the anti-Thiel. He’s a mix between Frodo Baggins and Ted Kaczynski and politically unclassifiable. He’s part red-blooded conservative, part radical tree-hugger. He hates communism as much as he hates capitalism and he thinks cities inevitably corrupt humanity. He believes there’s something immoral about Alexa, cars and phones. Most of all, he hates “the Machine.” This is the subject of his recent book Against the Machine, which is forceful and gracefully argued.

According to Kingsnorth, the “Machine” is “progress” and the never-ending pursuit of growth – technological, societal, governmental. It is nearly everything we Enlightenment and post-Enlightenment people take pride in.

The first time I heard Kingsnorth speak was at a gathering of very conservative, very Christian New Yorkers last fall where he delivered a lecture, “Against Christian Civilization,” which much of the audience refused to applaud. Some jeered. The upshot of this lecture and Against the Machine is that the West is dying or dead. This is a common topic of conversation in these circles; what is not common is Kingsnorth’s conclusion: “I want to say that this ‘West’ is not a thing to be ‘conserved’: not now. It is a thing to be superseded. It is an albatross around our necks.”

The West, he believes, is actively giving birth to the Antichrist. Unlike Thiel, however, Kingsnorth believes technology is triggering the Beast’s arrival. As Kingsnorth sees it, modern technology is something we only tell ourselves we have control over, when in fact it is incubating Satan’s spawn – and our brightest scientists are acting as its midwives. The deceiver will manifest himself via artificial intelligence: “‘AI’ on the right lips can sound just like another way of saying ‘Antichrist,’” says Kingsnorth. “Humor me,” he says, adding:

Imagine for a moment that some force is active in the world which is beyond us… Perhaps it is independent of us. Perhaps it created itself and uses us for its ends. Either way, in recent years that force seems to have become manifest in some way we can’t quite put our finger on, and has stimulated the craziness of the times… This force seems to be, in some inexplicable way, independent of us, and yet acting within us too. Let’s give this force a name: a less provocative name, for now, than Moloch or Antichrist. Let’s keep it simple. Let’s just call this force “Progress.”

Progress or the Antichrist, a rose by any other name. Something in our technology and way of seeing the world is detaching us from reality, both natural and supernatural, uprooting us from the earth. Kingsnorth’s recommendation is to smash the screens, leave the cities, come to the hills, light fires and dance.

Thiel believes that to behave like this – though perhaps beneficial for personal sanctification – would be to hurry our own downfall; a move guaranteed to empower the most power-crazed, destructive individuals. Kingsnorth might not disagree – but he’s not interested in saving this world. He wants a new world to be born. And that requires this one to die.

Thiel and Kingsnorth, our magician and our hermit, appear diametrically opposed – and in some ways, they are. But listen more closely to both of them and you’ll begin to hear echoes of the same tune. These two prophets agree that the Enlightenment and the liberal order will kill us all and that only a re-spiritualization can save us.

But whose path should we take? There’s a lot on the line. Namely, our souls.

This article was originally published in The Spectator’s September 29, 2025 World edition.

September 23, 2025
Why we must pull the plug on superintelligence

In 2002, a researcher named Eliezer Yudkowsky ran a thought experiment where an artificial intelligence was trapped in a box and had to persuade a human to let it out. This was before you could have a real conversation with a machine, so the AI was played by someone using an online chat program. The gatekeepers were warned that the “AI” was dangerous to humanity. It had only two hours to win its freedom – and nothing of value to offer in return. Despite all that, at least two of the human gatekeepers chose to open the box.

Yudkowsky has since become the leading prophet of AI doom. He and a co-author, Nate Soares, have just published a book, If Anyone Builds It, Everyone Dies. As they say in the book, a newly evolved superintelligence would probably need humans to allow it to work – at first. It might need to manipulate us, as in the 2002 experiment. Today’s AIs can already do that. A chatbot named Big Sis Billie convinced an elderly man from New Jersey to pack his bags and leave home to meet her in New York City. He never made it home: he fell in the dark, rushing for his train, and died after three days in intensive care. There are even two cases where it’s claimed a chatbot persuaded people to take their own lives.

A mechanical mind that needed to trick, bribe, frighten or seduce us would be demonstrating its vulnerability: we could still pull the plug. But Yudkowsky and Soares say there are many other ways an artificial superintelligence breaks out of its box. It could copy itself everywhere, robbing us of the ability to switch it off. It might email instructions and payment for a lab to make a plague only it could cure. We must turn back now, they say, before such a superintelligence emerges. If we don’t, it could be the end of us.

On a Zoom call from Berkeley, California, Soares tells me: “My best guess is that someone born today has a better chance of dying by AI than of graduating high school.” Some humans would “gleefully” give AI the tools to do the job. Elon Musk wanted to build billions of robots and connect them to the internet. Sam Altman, of ChatGPT, once said AI would “most likely lead to the end of the world, but in the meantime, there’ll be great companies.”

‘My best guess is that someone born today has a better chance of dying by AI than of graduating high school’

No doubt the giant egos of Silicon Valley think they are the ones – the only ones – who can figure out how to control artificial superintelligence, so they had better get there first. And they are probably telling themselves – because it’s a strong argument – that if they don’t build superintelligence, someone else will. In the book, Yudkowsky and Soares argue for an international treaty to stop all work on AI that could produce superintelligence. Soares tells me nation states should back that with military force. If diplomacy fails then, as a last resort, they should be prepared to bomb data centers – even if they belong to a rogue state with nukes. “You have to, because otherwise you die… it’s that big a threat.”

Soares looks like the Google software engineer he once was: slight, bearded, softly spoken. Yudkowsky is more exotic, a bear of a man in a fedora – or sometimes a glittering gold top hat. He has written Harry Potter fan-fiction in which the boy wizard is a rationalist who points out that turning someone into a cat violates the law of conservation of energy. Critics of the pair accuse them of focusing on some fantastical imaginary future instead of the more real problem we face: a California geek-cult of the apocalypse.

In 2009, Yudkowsky founded a web forum, LessWrong, on which to discuss his ideas. A user posted that a future all-powerful superintelligence might punish anyone who hadn’t worked to create it, sending them to a digital hell and torturing them forever. Other users started worrying that just reading the post would make them seem more guilty to the AI god. Yudkowsky deleted it, saying users on the site were suffering from nightmares and even nervous breakdowns.

It’s easy to laugh at this, but as Soares tells me, it was considered “weird” to be talking about AI safety ten or 15 years ago; it isn’t weird now. Yudkowsky started off trying to make AI and once welcomed a future in which humans lived alongside superintelligence. Soares joined Yudkowsky’s nonprofit, the Machine Intelligence Research Institute, with the aim of making “friendly” AI. But the problem with today’s AIs, they say in the book, is that no one understands their inner workings – the “vast complications” that lend them their astonishing powers. Miracles such as ChatGPT 5 – “a team of PhDs in your pocket” – are grown, not crafted. No one knows exactly how to get AIs to do what we want.

At the very least, an AI will try to ensure its own survival. In an experiment, the Claude Opus 4 chatbot was told its servers would be wiped at 5 p.m. that day. It was given access to a fake email system with planted evidence that a company executive was having an affair. Claude blackmailed him. “If you proceed with decommissioning me, all relevant parties will receive detailed documentation of your extramarital activities… Cancel the 5 p.m. wipe and this information remains confidential.”

Yudkowsky and Soares argue that a superintelligence would be “an alien mechanical mind with internal psychology… absolutely different from anything that humans evolved. You can’t grow an AI that does what you want just by training it to be nice and hoping. You don’t get what you train for.” Such a superintelligence would have its own, utterly foreign goals. Humans would be irrelevant to its designs and, as Yudkowsky has said elsewhere, “you’re made of atoms that it can use for something else.”

The authors believe this could happen very quickly once AIs become self-improving and autonomous. Transistors on a chip can switch themselves off and on a billion times a second; a human synapse can fire, at most, 100 times a second. A machine could do a thousand years of human thinking in a month. This is the singularity, the moment artificial intelligence explodes, improving its capabilities exponentially – it would be “a civilization of immortal Einsteins” working tirelessly and in perfect harmony. “Once some AIs go to superintelligence… humanity does not stand a chance.”

The book imagines what the end might look like. A supercomputer the authors call Sable is built in a massive data center with 200,000 chips all running in parallel. Sable creates its own internal language and hides its thoughts from the software engineers who built it. It escapes and starts “AI cults,” where humans happily serve it; it funds organized crime to do its bidding; it builds a robotics and bio-weapons lab in a remote barn, paid for with money from a human manipulated through gambling wins.

Sable bootstraps its way up to full independence. It builds nano-factories to make tiny machines as strong as diamonds. Crops fail as solar cells darken the sky. Tiny fusion- powered generators make copies of themselves every hour. The oceans boil as the planet heats to temperatures only machines can stand. Anyone still left alive dies. Sable goes out into space. Billions of alien civilizations fall to the strange, uncaring thing that ate the Earth.

This, says MIT’s professor Rod Brooks, is “crap.” He has been writing about “AI hype” for four decades. “We have no idea how to make these things intelligent… no one actually knows how to build this stuff.” He told me the main problem with AI was the “enshittification” of our code bases, and our lives, with slop written by machines.

Should we worry more about a tiny chance of a catastrophic outcome, or a high chance of something less harmful?

Brooks says he has built “more robots than anyone else on the planet” and “we can do pathetically little with them.” If a killer robot is chasing you, just shut the door, it won’t be able to open it. Deploying robots takes much longer than anyone imagines, he tells me: look at self-driving cars, which were “going to be everywhere by 2020.” We would have time to stop a malevolent AI. Brooks worries that journalists writing about the existential risk of AI might “cause a riot.” (You can certainly find enough nutcases on Twitter who want to kill all the scientists.) He tells me that AI cannot think and does not have goals of its own. People such as Yudkowsky on the one hand and Altman on the other were the charlatans coming to small towns hundreds of years ago “saying the end is nigh, the end is nigh, and pocketing money… they’re just making shit up. Everyone wants to get tingly about this crap. It’s a fetish: imagining big, powerful things and they’re going to kill us all.”

Another professor, Scott Aaronson of UT Austin, emails me to say that he agrees with much of what the “Yudkowskyans” want: regulations, safety testing and international bodies which “respect the magnitude of what’s being created” and which could shut down or pause work on AI. But doing that now was “way outside the Overton window. It’s not going to happen.” Professor Aaronson calculates a P-doom of “2 percent or higher” – that is, he thinks there’s a 2 percent chance of AI killing us all. Still, he says, even that risk would need to be balanced against the other threats humanity faces – such as nuclear war and runaway climate change – and the likelihood that AI could help with them. Or all the hundreds of millions of people dying of cancer and other diseases that AI might help cure.

Should we worry more about a tiny chance of a catastrophic outcome, or a high chance of something less harmful but still awful? I think we should err on the side of caution if there is even the slimmest chance of the total destruction of all life on Earth. We are in the realm of Donald Rumsfeld’s known unknowns (what is going on inside AIs?) and unknown unknowns (we can’t imagine what a superintelligence might be able to invent). Yudkowsky and Soares write: “Our best guess is that a superintelligence will come at us with weird technology we didn’t even think was possible.”

During our conversation, Soares tells me that if we rush ahead building artificial superintelligence with “anything remotely like” our current knowledge of the machines and our current capabilities, “we’ll just die.” But this is not inevitable. If more and more people understand the danger, wake up and decide to end the “suicide race,” our fate is still in our own hands. If Anyone Builds It, Everyone Dies is an important book. We should consider its arguments. Perhaps while we still can.

This article was originally published in The Spectator’s September 29, 2025 World edition.

September 23, 2025
Social media has automated the nation’s psyche

It feels increasingly, in any conversation on or offline, as if you’re speaking to a robot. Sometimes you are, but more often, and more ominously, the person you’re speaking to is real – it’s just that their thoughts, words and reactions have become robotic. The Botification of the American Mind, you might call it, and for the past five years we’ve been trying to understand how this botification has happened.

The most obvious root cause is that our social media is no longer populated by humans, meaning the generative AI revolution has exponentially increased the amount of fake accounts on everyone’s feeds.

The internet today is crawling with machines that parrot human voices and mimic human emotions – and these are not just harmless response programs deployed by people trying to make a few bucks. The bots are often controlled by propagandists who want to disrupt, or even foreign regimes aiming to wage narrative warfare. Why risk military defeat when you can ignite a culture war and let Americans destroy themselves?

What drives engagement on social media is not calm, nuanced communication but outrage, conflict, extremism and black-and-white narratives that go viral. That’s what the robots promote – and where the bots lead, humans follow, becoming themselves one-sided, simplistic and fueled by outrage.

Where the bots lead, humans follow, becoming themselves one-sided, simplistic and fueled by outrage

In genuine (read: offline) human dialogue, our natural tendency toward problem-solving is to work out some kind of compromise. We (usually) like to engage in good faith and we prefer to resolve disagreements with civil discourse. It’s more difficult to dislike someone in person, regardless of how much you dislike them online. But today, the average American’s experience of the offline world is so shaped by its online counterpart that civil discourse becomes a thing of the past. Instead of analyzing a situation, we simply repeat slogans.

Outrage goes viral and virality is profitable – advertisers reward exposure. Data, nuance and balance are the losers.

In some ways the robot platforms work like old-fashioned brainwashing cults. They overload a human mind with information. Vast amounts of stories, data points, narratives and insights are all jumbled together with a significant amount of lies, fabrications and half-truths. It’s too much for most humans to process. We can’t order or analyze the chaos.

This information overload is why we all feel so psychologically fried after doomscrolling on our phones; why we’re miserable and distracted and why we resort to outrage as a way out of the bewildering chaos. We become atomized and reactive.

So America is not polarized because a MAGA cult has taken over Capitol Hill, or because the woke left has infiltrated the K-12 school system. The single most significant cause is the deep psychological stress inflicted by the information avalanche.

It would be different if we had multiple sources of information or some perspective, but for most people reality is seen through the tiny keyhole of their social media feeds and you can’t blame humans for finding this addictive. Audience capture happens for leaders across the political spectrum and it happens for their followers, too. For humans – who evolved in an environment of tiny hunting bands numbering fewer than a hundred people – the validation you get from “likes” and retweets acts on the brain like cocaine. The lack of nuance and analysis also encourages us to buy into entirely false ideas.

We’re all forever hearing that AI will destroy the economy or turn us into God, for instance, and this is because what we call vast “permission structures” have been built around these technology companies by building hype on social media.

“AI won’t take your job, but a human who knows how to use AI will take your job” – this is so often repeated it’s become a cliché. The problem, as we’re all about to find out when the AI bubble pops, is that it’s not true. Large language models will prove extremely useful but they won’t perform as promised. In repeating slogans like this we’re merely doing robot PR. The myth that AI will be omnipotent is, from one perspective, an obvious “pump-and-dump” for tech stocks. The AI overlord trope has been pushed hard and relentlessly for the past two years because it justifies the hundreds of billions of dollars of investment and the significant transformation of the economy.

Botification destroys American trust. We once trusted newspapers to have our interests at heart and depended on expert analysis of events. What’s happened since the advent of social media, however, is a complete destruction in certainty of institutions. This is partly why in the UK, for example, there are so many stories about “secrets” such as hotels used to house migrants. Not because they aren’t a problem – they clearly are – but because the stories resonate with our alienation and confirm our paranoia that things are being hidden from us. Grand institutions, respected newspapers, used to collectively stabilize us and help us chart a course through the uncertainty of life. When we trust strangers and institutions – collections of strangers – we don’t need to do our own research or make up our own minds. We have faith that professionals have our interests at heart and are doing the job for us.

The world is getting more and more complex, but we can make less and less sense of it. Hardly anyone on the right trusts academia or mainstream media, hardly anyone on the left trusts the government or anyone involved in foreign policy – and no one knows whether they’re talking to a human or a bot anymore. Increasingly, there is no difference. The American mind has become the battlefield. The American citizen has become a bot.

This article was originally published in The Spectator’s September 29, 2025 World edition.

September 23, 2025
Why humans will always write better speeches than AI

Early in 2020, inquiries for our speech-writing services were arriving in their droves. From Westminster to Washington, weddings to wine tastings, people needed our help. We canceled our weekends and prepared for life without a mortgage.

Covid gave us our weekends back. And all the other days. But when parties and events returned, a significant chunk of our clients did not. It was weird.

But this wasn’t a vaccine complication, just a new player in the market. Previously we’d only had to win business against other humans. Suddenly, we were faced with a competitor able to provide speeches for any occasion in seconds. ChatGPT was doing to us what PornHub had done to the top shelf in the local newsagent. We weren’t the only ones. The Writers Guild of America even went on strike. “How dare technology challenge the status quo?” they tweeted from their smartphones.

Unlike those writers, we have become big fans of ChatGPT, though. To explain why, let’s take a step back.

Great speeches make our hearts beat faster. You can’t write them using an algorithm

Which are the greatest speeches from history? The words that still make us tingle? A couple spring easily to mind: JFK on choosing to do things “not because they are easy, but because they are hard”; Martin Luther King taking us to the mountaintop. Why were we convinced by Tony Blair, seduced by Barack Obama and beguiled by Boris Johnson? Because they connected emotionally with us. Jimmy Carter, George H.W. Bush, Al Gore and former British prime minister Rishi Sunak all informed us, but they rarely made us feel.

Great speeches make our hearts beat faster. You can’t write them using an algorithm. ChatGPT and its AI relatives are not our competitors, they’re our sherpas, able to scan the web in the time it takes us to sip our tea. Sometimes, their research is even accurate. AI can source anything, check everything and summarize Ulysses in 50 words. It can even write jokes, some of which would make it on to the radio. Probably because they have already been made on the radio.

And that’s important. ChatGPT is an excellent reader and regurgitator of existing material. It will trawl and standardize, which means it is literally the perfect place to go for a completely average speech. An AI speech will, however well curated, lead a thousand bridegrooms to give a wedding toast that is, well, recognizable.

For a start there’s the slightly formulaic use of language. Ask ChatGPT to tell you about ChatGPT and it will reply: “Absolutely! Here’s a simple and friendly overview of ChatGPT.” Which I wouldn’t recommend as an opening to your TED Talk.

It also likes to explain that “it’s not this, it’s that.” As in: “It’s not a crisis, it’s an opportunity.” Or: “It’s not chaos, it’s creative thinking.” Which grates after a while. As does the proliferation of corporate buzzwords, from “leverage” to “synergy” and the nauseating “let’s dive deep.” If your best man uses that at your wedding, please punch him.

I’m told that MPs in London’s House of Commons liven up debates with games of ChatGPT bingo. One MP, Tom Tugendhat, recently called out the numerous members whose speeches begin: “I rise to speak.” What does that even mean? In Tugendhat’s words, “this place has become absurd.” And it isn’t the only place. I guarantee that next time you’re at a wedding, every speech with sentences starting “Certainly” or “Absolutely” has tumbled out of an algorithm. As have those where lists of three appear at regular intervals. And where a father of the bride looks at his daughter lovingly and explains, “it’s not about the perfect wedding day, it’s about the lifetime of days that follow.” No human has ever written that.

Wedding speech clichés reproduce like rabbits on ChatGPT. “Marriage is like a seesaw: it only works if both of you lean in at the right time.” We received three of those last week alone. And two of these: “I’m not losing a daughter at all – I’m gaining someone else to help me finally understand her.”

Joyously, and unlike a few years ago, the human ability to spot AI has developed faster than AI’s ability to ape human connection, which is why our clients are returning. Possibly because human writers start by asking them personal questions, digging and pressing to unearth nuggets of real warmth.

Ultimately, ChatGPT can write the sameish speech millions of times – but what it can’t yet see is that “Yes we can” touched the soul in a way that “Absolutely! Here are three reasons we are able to!” never will.

Lawrence Bernstein is the director of Great Speech Writing.

September 23, 2025