The real danger of "AI"

it's not a sci-fi scenario

I’ve recently published three longform pieces on the global Great Game and how it’s been tangled with American politico-economic incoherence during the Trump derangement: “A non-economist on the US, tariffs, China, and Trump,” “The Euuropean conundrum,” and “Trumponomics.”

At this stage of the game, actors like the Chinese and Russian governments are clearly outmaneuvering the erratic and increasingly feeble-minded US Chief of State, and in stating so, it may have appeared that I’ve “taken their side.” In only the most temporary and tactically limited way, there may be a particle of truth in that, but I need to correct any misapprehension that I might anticipate a new and improved future led, for example, by the CCP or any constellation of breakaway states as the US empire stumbles through its now inevitable endgame.

Progress—as most commonly conceived—is a deadly siren.

I recently read Anthony Galluzzo’s “On Paul Kingsnorth, Green Romanticism, and the Technophilic Left,” which I recommend, and that article provoked me to write this particular essay. Anthony and I have exchanged thoughts online occasionally, and while we of differing temperaments owing to background and age, we share a powerful affinity for what remains of enchanted nature—including human nature—and a bitter antipathy to what I somewhat clinically call radical technological optimism. Anthony attacks it as “Jetsonian leftism.”

Likewise we are both refugees from the left who’ve nonetheless remained nominal socialists—even if by that we mean something strikingly different from many others who wear the same badge. Eco-socialists, if you like, or what I’ve called subsistence socialists. Anthony surpasses me by a mile in both erudition and lyrical skill, but he’s been very tolerant of my deficiencies, even when we’ve disagreed. One point at which he and I and Kingsnorth and Nina Power intersect is our indebtedness to and appreciation of the works of Ivan Illich—who was influenced by feminist historian Carolyn Merchant. If you’ve never read her 1980 book, The Death of Nature: Women, Ecology, and the Scientific Revolution, you’re behind. Illich, ditto.

In Anthony’s “Kingsnorth” article, he addresses our cyborg age, which Kingsnorth simply refers to a The Machine—a term that can be read as Moloch, the fearsome ancient deity of Hebrew lore to whom children were sacrificed.

Anthony is a lyricist—a good one, even when he’s being acerbic—and his project, Romanticon is worth a look for the literary-adjacent. I myself am more the foul-mouthed, hillbilly pathologist, at least in my textual guise, and as such I’m appending today’s treatment of so-called Artificial Intelligence.

For the last few years, I’ve tried to avoid the AI buzz in the same way I try to avoid fads and furies. Temptation got the best of me when people started suggesting this shit was comparable to an actual human mind, that is was even on deck to supplant human intellect—an idea so transparently idiotic that it brought out the little Mencken in me to give it the lampooning it deserves.

I remember hearing from a young age, “If you’re so smart, why aren’t you rich?” As things turn out, rich people know how to get rich (by shitting all over others), but great riches can have the same effect on the rich as bovine spongiform encephalopathy—mad cow disease: wealth eats holes in one half of their brains, leaving the other half to imagine colonizing Mars or uploading our personalities into an electronic “cloud.”

Kingsnorth is one who, unlike me, expresses real alarm about AI’s putative potential to become a kind of science fiction horror like Terminator, some demonic supercession of humanity, but I’ve never shared that view. It’s an old NCO’s habit when people panic to bark at them, “Get a fucking grip!” (Panic is dangerous, even when there’s something to fear, because it makes things worse even as it disables us.)

Not that I’d bark at Paul Kingsnorth or anyone else who senses that “technology” has taken a demonic turn. As Kingsnorth, an Orthodox Christian, points out, murder and technology are combined in our Bronze Age origin myth, when Cain kills Abel then goes on to invent technological development and urbanization. (Read it, for real.)

What I need to say out front, as justification for my no-panic position, is that humanity—the part of us that continues to exist apart from the many mutilations of our most infernal inventions—has survived from the Bronze Age until now (as has the origin myth). We’re susceptible to demons, but the divine spark is stronger than the most dreadful dark.

I write today about the dangers of AI, but not about killer robots—shit, we already have them—or mind-replacement—we already have that, too, but never in some irreversible and totalizing way. Nothing, I repeat, nothing technological can ever even approach the powerful and mysterious thing we call mind . . . or soul . . . because—and this may be controversial—machines are not people, and people are not machines. You may as well try to replace trees with rat traps.

The dangers I’m going to unpack are hardly different, except in form, from the same dangers that have attended human existence since Cain bumped off his little brother: power games, abuse, precarity, displacement, dependency, war, hunger, madness, disordered love, and malignant hatred.

So, we’re going to look at “AI” technology as an arms race, an economic time bomb, an environmental catastrophe, and finally a spiritual pathology. Generically speaking, nothing new here. We’ve seen all these before, the difference being this is now a far more universal form—a geopolitical constant, if you will, with biospheric collapse and nuclear weapons hovering in the background. In that sense, AI is very small-c catholic.

In Anthony’s re-posted article on Kingsnorth, his introduction states that he’s moved further away from materialism—including the Marxian materialism that was formative for him. I identify. Nonetheless, there was a kernel of truth in the Marxian base-superstructure framework. In Marx, it was all about the dialectical interplay between basic “material conditions,” meaning environment, methods of productions, and so forth, and superstructural culture and ideas, with the base—as stubborn material reality—playing the more determinative role.

My first challenge to the Marxist dogma—not shared by all Marxists btw—that methods and means of production being the most basic of bases, was related to gender (in the second wave sense), in particular how constructions of masculinity—influenced by the transhistorical phenomenon of war—were perhaps even more powerful transgenerational basics than productive methods and relations.

There’s a chicken-egg dilemma there, but my own experience led me to highlight the underappreciated role of masculinity read as conquest and domination. This idea was met, on the one hand, by waves of biological determinism—with frequent references to testosterone and lions and such—and on the other hand, by gender ideologues of the third-wave type, who were frantically trying to dissolve the “oppressive” categories of “man” and “woman.” Así es la vida.

This thesis led to my own inability to decide which was the base of the base—conquest masculinity or war—because I came to realize these reproduced each other. Where these ruminations led me further afield from dogmatic Marxism is the subject of an essay I posted here in September (2025), called “The epistêmê of war & related matters,” which suggests—as the evidence suggests to me—that war, even more than the desire for accumulation, might be the primary driver in capitalist technological development. More chickens, more eggs, of course, but the idea deserves enough attention, I think, to either rebut or validate it. (I could begin the rebuttal myself with, say, the Hanseatic League or something, but the chickens and eggs keep alternating.)

I bring this up because the whole China-US rivalry over “AI” is literally an arms race. Politicians, “experts,” investors, and tech bros have all been sowing panic about AI’s war applications; and I’m quite sure the CCP’s emphasis on AI weapons development is a response to this war-mongering by the war-addicted US. For now, I’ll set aside the macho vibe that enervates the war epistêmê, having already given it a richer treatment elsewhere.

Mike Gallagher, “Defense Head” at Palantir Technologies, called AI the next watershed in warfare after the invention of nuclear weapons; that is, the nuclear arms race of the Cold War with the Soviets is now the AI arms race between the US and China.

“We cannot allow,” D-Head Gallagher said, “the Chinese Communist Party to control the commanding heights of critical technology.” ← (Conquest masculinity thesis—validated)

Well, too late, Cowboy, because they already control 90 percent of the rare earths exports that form the (ahem) material base of that critical technology. They already have an advanced and strategically constructed energy grid to support its further development. They crank out 77,000 STEM PhDs every year—that’s three for your every two, Hoss—and they all get jobs with heavy government support. Tsinghua University graduates publish five times more patents in machine learning and AI than all the grads of CalTech, MIT, Princeton, Harvard, Stanford, and the University of California combined.

The US has the most advanced “chips,” which rely on Chinese supply chains, and we are about 20 years—if at all—from constructing an energy grid sufficient to support whatever fantasies underpin AI development—which even then would require a second Manhattan Project (explicated below). Our universities are being turned into academic shit-holes, and the AI apps we have are turning students into cheats and dunces.

The language of these calls to arms is all about “dominance,” “winning,” “existential threat” (as opposed to pre-Husserlian threat, I spose), and psuedo-military jargon like ^^^ “controlling the commanding heights.”

If you’re of a suspicious bent, you may have noticed that the most vocal of these laptop generalissimos are dudes (mostly dudes) who are heavily invested in so-called AI, or politicians who receive vast sums of campaign cash from them. It ain’t rocket science. What does require a pattern of thought transcending quid pro quo and qui bono is the menu of effects generated by AI investment.

I’ve already explained in an earlier article why the US is not going to actual kinetic war with China, and why the Trump government is getting its ass kicked in the more metaphorical trade war. For its size and cost, the US military is perhaps the most inefficient and ineffective military in the world—costly, cautious, and overstretched . . . and now under-resourced as Ukraine and Israel have eaten into stocks, even as China has a choke hold on those weapons-essential refined rare earths. All the war blather is a cover for the economics—which are not good.

Let’s start with intent. What do the AI mavens intend—apart from making tanker loads of cash and attempting to become immortal?

As a pro-union fella, who for a brief time bagged groceries as a member of the UFCW, I remember the first time I saw one of those scanner aisles in a store. It took approximately one nanosecond to realize what this thing was intended to do: replace a human being. Electronic scabs.

AI is not intelligence; it’s super-computing—different things—but as we all know, the purpose of taylorization was to turn thinking human beings into appendages for machines. To divide a process into its simplest, most mindless tasks, so that it required no complex skills. This not only sped up production lines, it made each worker easier to replace. Harry Braveman’s canonical book Labor and Monopoly Capital: The Degradation of Work in the Twentieth Century described this in great sociological detail.

The crux of this was the “deskilling” of the worker, which not only made him or her easier to replace, but more dependent on the employer and more subservient to the employer class. The ideal “reserve army of labor,” in a mechanized and taylorized society is one with no transferable and independent skills—think artisans and craft workers. Unskilled and easily replaceable labor works for less, and every employer knows that the margin of profit is most easily increased by minimizing the pay of workers. Marx 101 (and the bearded one was pretty much on the money with this one—employers are the most adept Marxists in this regard).

The intent of AI in production is to replace humans. Trump’s administration has hitched its wagon to AI as the future of America’s industrial renaissance, which in true Trumpian fashion insults our intelligence and betrays their own stupidity.

Premise 1: Re-industrializtion will benefit the American working class.

Premise 2: AI is the future of industrial production.

And yet . . . AI is specifically designed to produce more with fewer actual workers. AI—once broadly applied—is a job-killer.

Of course, what the employer class ignores, it having business-cycle-attention-deficit-disorder, is the rest of Marx regarding living and dead labor, competition, and realization—that is, the inevitable profit crisis when everyone has the same (dead-labor) robotic work force.

China, on the other hand, is already leading the pack in robotic production.

But there are some key differences between China and the US, beyond the “AI” advantages listed above, i.e., China’s economy under the soft-supervision of the state, China’s government is not—like the US—run by rentiers (finance capitalists), and China is still in the exponential phase of development, whereas the US has entered the death phase.

Explaining the super-simplified curve above, “development,” in an economic “growth” paradigm, is driven by the combination of potential, thermodynamics, limits to growth, and saturation. Potential characterizes the exponential up-curve, and saturation characterizes the decadent down-curve. Think “room for development,” in this regard, then think of development as mechanization and urbanization.

^^^ Saturation

943.5 million people live in urban China, as opposed to 464.8 million in rural China—or around 67 percent urban. In the US, we are around 80 percent urbanized, with another big difference being urban planning and infrastructure.

There is latent potential for growth in continued urbanization, but also in the continued development of infrastructure, which is supported by finance capital, but not—as in the US—driven by it.

At some point in the future, China will level off on the stationary phase; but if it doesn’t de-center the growth paradigm, it, too, will hit saturation and begin the death phase. There are energetic and material limits to growth—any growth, all growth. Full stop. Right now, though, driven by its “socialist” developmental agenda and under competitive international pressure and the perpetual threat of war, China will continue to “grow.”

As I suggested, there are not mere quantitative differences afoot here, but qualitative ones as well. Political continuity, political soft-control of finance and production, less stringent environmental regulation, subsidized education, etc. Emphasize control of finance capital, or what some have called rentier repression.

Returning now to the US, rentiers (finance capitalists ← speculative capital) have amassed such fortunes in fungible fictional value (valuations on “pieces of paper”) in a pay-to-play political system that they now select our political “representatives.” This is not news. What may be news is the current composition and disposition of finance capital. This is where AI comes onto the stage.

AI is the hype that’s pumping tech stocks, the AI component of which is massively overvalued and to which the overall market is now massively exposed. As this is written, NVIDIA’s value is jumping up and down like a kangaroo.

AI will dump its value as a gambling asset, because more and more people realize that AI cannot live up to the hype. The AI bubble is not alone on the field, however. All the parts of the financial system are put together like a Rube Goldberg device.

Any disruption in one component has devastating potential for all the other functions.

The AI bubble is happening at the same time as a crypto-bubble, a precious metals bubble, stagflation (here now—remove the bullshit trades among five tech companies, and we’re already in a recession), a private credit and private equity meltdown, dollar devaluation, a repo market crisis, a failed trade war, and an insane clown executive branch.

What do we mean by overvaluation? Simple. It can’t live up to the hype.

For a while, when stock prices are going up, everyone wants in on the action. Invest $10 today, get $15, tomorrow, invest that $15 the next day, get $25 the next. Yeehaw! The values is being bid upward like a rave vibe—do another bump . . . dance harder. Recent reports on NVIDIA stock dropping were met in the following days by a sigh of relief when the stock was bid up again (bump), and rentier commentators celebrated, “There’s no bubble, see? The music has started again.” (Elon should rename X as MDMA.)

At the end of the party, though, reality returns.

AI is a financial pig in a poke. The reason there’s not a real pig in the bag, so to speak, is that material conditions thing again.

Alf Hornborg was the first person to describe money as an ecosemiotic phenomenon. It’s semiotic inasmuch as it’s a sign, a signal, a representation of value without being the use-value itself. No one eats money. It’s only used for exchange (which is why it’s so dangerous ^^^ when people just exchange value-representations with other value-representations on bets, creating fictional value—value with no use in itself). Money is ecological inasmuch as it facilitates material exchanges very quickly, so quickly in fact that (1) it outruns natural adaptive and reparative processes and (2) it facilitates material projects—using credit and debt—that (1) may or may not work as planned or (2) will create all kinds of unanticipated mischief or (3) will create anticipated mischief the consequences of which will be palmed off those who aren’t invested.

If I were to add a (4), which is maybe not as directly eco-material, it would be that money is a magnet for crime.

(We’ll just mention, in passing, that fictional value, having no comparable value in actually circulating material goods, has a double potential—first as inflationary value and second as the oh-fuck disappearance of value . . . the latter of which is now puréed together with your pension plan.)

AI boosters are telling everyone this is the thing of the future, while they show all kinds of magical-fantasy shit to the public, but fail to mention what is materially required to achieve this kind of uber-computing, which—as we said—may or may not work as anticipated (hint, it won’t).

First of all, facilities are required to house and operate the data systems. These are called data centers. Sounds about as threatening as a Yorkshire puppy. In fact, a new hyperscale data center—the kind that AI boosters want—is an enormous agglomeration of very expensive and environmentally-taxing materials that eat more electrical power than a whole city.

^^^ to hell with habitat

As of this writing, 5,427 data centers are in the US; that’s up from 2,667 in 2021. So there is an immense and rapidly progressing material investment in these energy-sucking behemoths, which is included in the overall investment costs which are now debt which has been been bet on a very uncertain future. These projects are heavily subsidized (thanks, citizens). The tech bros are already talking about data centers that will consume 5 gigawatts annually. That’s equivalent to dropping the entire island of Cyprus into an existing regional energy grid—the equivalent generation of five nuclear power plants . . . for one fucking data center. If this sounds insane, it’s not. Merely. It’s absolutely ketamine-fueled psycho fruit-loop bat-shit bonkers.

This year, $475 billion was spent building these centers—think the GDP of Norway.

This year, data centers ate 4 percent of America’s electricity. By 2028, it’s expected to be 12 percent. This energy wasn’t added to the existing grids, but added as demand, even as the grids themselves are already antiquated (the existing network is 100 years old), stressed, and in a state of general disrepair. This added demand represents a rapid progression toward multiple breaking points.

Grids were already stressed by climate destabilization as we’ve seen with serial brownouts and blackouts.

Meanwhile, this demand has driven up your electricity rates, a lot. Those rates you’ve seen go up? Inflation accounts for 37 percent, while 63 percent of that price increase you can blame on data centers, which have driven producer costs up—last year’s capacity auction, e.g., saw prices rise eight-fold. Data centers in action.

What would it take to “modernize” the US energy grid? Well, for starters, we’d need a convocation of the best minds in the world to study the existing grid and make bold recommendations. Then, we’d need to educate enough scientists and engineers to plan the replacement grid and transition, including diversifying away from fossil hydrocarbon dependency, and developing a network of resilient interactive and non-ramifying small-grids. We’d need reliable long-term supply chains abroad. After that, we’d need to implement a nationwide federal jobs program, which would entail training programs on par with organizing a national military, with hiring priority given to those who’d lose their jobs during transition. Then we’d need a subsidiary bureaucracy to localize projects, with progressive taxation and bond markets to finance them.

Easy. Except this is America, where that shit ain’t gonna happen.

Right now, the tech bros are ripping off this entire country to the tune of trillions with the enthusiastic assistance of a criminal enterprise know as the Trump administration—which is walking the nation into the aforementioned financial crises. Yay.

Trump has decided to double-down on fossil energy and attack the safer alternatives; but he has bigger issues now, as Epstein’s ghost stalks him through the darkened halls of Mar-a-Lago, the courts begin sweeping up his messes, his grand plans fail grandly, his base fractures, his nervous loyalists look for the exits, and his body continues to betray him with senescence.

A recent poll measured Trump’s approval on ten key issues. He was up only two—Israel/Hamas ceasefire and border security. On immigration, he was -10, tariffs -26, economy -28, Russia-Ukraine -34, inflation/cost of living -44, Epstein -48, government shutdown -56, and the Argentina bailout -58. Tough times for the Boyz n Barbies.

So here we find ourselves—waiting for that black swan—even as we are ourselves all caught in the gears of growth, while ice caps melt and ocean currents collapse.

All that.

Perhaps the greatest danger of AI is that it’s nothing new. It’s the fever dream of the digital age, and here is our worst deskilling.

It’s a crisis of imagination, a simulacrum of re-enchantment to address late modernity’s spiritual malnutrition with empty calories.

It’s the belief in AI itself and all that entails.

In this, we are closer to a reckoning than ascendant China, and therefore closer to finding our way past it—painfully and dangerously, perhaps, but this is the best I can do at doomer optimism.

I’m just a resident alien—a witness passing through.

The world still abounds with glory if we’ll just stop, look, and listen.

Those romantics might be onto something after all.

Paul Kingsnorth has said more than once—and I’ll not dispute him—the basis of all these tribulations is a crisis of the spirit.

The politics is a superstructure.