I am writing this on what could be the last day of an entire era of human history.
The AI bubble is stretched to its breaking point, and it does resemble an actual, birthday party soap bubble: massive, misshapen, hovering a foot or so above the spiky grass of the back yard. Even a child knows there’s no saving it. All anyone can do is hold their breath in horrified anticipation of the moment when it splatters. It may not be today–NVIDIA may be able to string out the shell game for another quarter or so–but it’s coming and it’s coming soon.
I’ve read something on the order of 10,000 words about what will happen when the AI bubble pops, from writers who know more about economics and technology than I ever hope to learn. They predict some general ripple effects:
- The AI industry will consolidate around a couple of major players, who might have some prayer of turning a profit once they’re no longer forced to compete.
- The stock market, which has been propped up on a single, load-bearing toothpick for the past few years, will crash and the economy will plunge into a recession.
- The tech industry writ large will undergo a period of contraction and mass layoffs.
I find all these predictions basically plausible, but they have left me dissatisfied. I think it’s because they are all too small. The cyclical churning of markets doesn’t reflect how all-consuming the AI narrative has been for these past couple of years. This was the greatest invention since the internet! No, since fire! This would make work obsolete! Also art! Also, for good measure, women! It would kill us all, or it would save us all.
The implosion of a story this large demands an equal, opposite reaction. AI has been done to us these last three years, and when it collapses, its architects can’t be allowed to cheerfully walk away from the rubble. There must be a reckoning.
It turns out I don’t want predictions; I want prophecy. So I’m making one myself, in the hopes that, like the prophets of myth, the act of speaking it will make it real.
The collapse of the AI bubble is not a tragedy–though it will doubtlessly do tragic damage to people’s lives–but a priceless opportunity. It is a chance for a referendum on the whole, rotten monstrosity of late-stage capitalism and the filthy tentacles it has shoved into our social, political, and economic lives.
AI is the perfect crystallization of the deranged logic that has brought us to this point in history. It has nothing to contribute except: more, faster, cheaper. It offers nothing except to consolidate power and wealth for the people who control it, while stripping what remains of life’s joys for its victims. (And there are many, many AI boosters, especially in the corporate world, who are mistaken about which side of the leash they’re on.) It is the monstrous reflection of the ideals of people who have profoundly missed the point of being alive. And their joyless myopia threatens us all.
Here’s what I mean by that. There’s a famous thought experiment in the world of AI safety called The Paperclip Problem. In this exercise, an AI exists whose only directive is to maximize the number of paperclips in the world. It excels at this. But in the absence of other values coded into its programming, the AI creates paperclips at the expense of every other goal. Every forest, every city, every person can either be reduced to the raw materials with which to create more paperclips, or an obstacle standing in the way. At the end of the story, all life on Earth has been extinguished and replaced by endless rows of gleaming paperclips, and one lonely robot feeling vaguely as though it missed a step somewhere.
This story is supposed to illustrate the supposedly existential threat of a rogue AI, but it has a rather gaping logical hole. Because the thing is: we already have the paperclip-producing machine. It’s just called capitalism, and instead of paperclips, the thing it produces at the expense of all else is profit.
Capitalism is burning the world alive to support its own delusions. We have lived under its logic for so long that it takes an effort of will to question it, and questioning it puts you outside the mainstream of either major American political party and the majority of mass media. The ubiquity of the capitalist delusion renders us helpless in the face of obvious absurdities. Why are we burning fossil fuels even as renewables become cheaper? (Not coincidentally, the fates of AI and dirty energy are intimately linked.) Why are we failing to rein in single-use plastics even as they pile up inside our bloodstream? Why must we let insurance companies weigh our lives against their revenue? Well it can’t be helped, you see. Not as long as long as somebody, somewhere is turning a profit.
But nothing goes on forever, not even this merciless onslaught of absurdities. And with AI, the wave has finally crested. The disconnect between what we are told to believe about AI and what it actually is is so jarring that our senses rebel and we balk like horses being led into a charnel house.
One of the most nauseating parts of the AI narrative has always been its aura of inevitability. I was recently listening to an executive who kept saying that AI is here, “whether we like it or not.” The more he said it, the hotter my blood ran. Because, hey, what if we don’t fucking like it? Does that not matter? Are we truly so helpless in the face of this two year-old technology that we must lie in front of it like a live action trolley problem? That might be the case if AI had something to offer that truly made it inevitable, but it patently does not. Environmental, ethical, and aesthetic issues aside: the damn thing doesn’t work. And that disconnect creates an opening for dissent.
We can opt out of a future we didn’t choose and don’t want. We can see LLMs for what they are: the last, desperate gambit of an industry that has stopped giving things to the world and started taking things from it.
AI has nothing to give us1, and when the bubble bursts – whether it’s with today’s NVIDIA earnings report or some other event – this fact will at last be laid bare. That moment may be our best opportunity to end our cultural deference to the boy kings of Silicon Valley and the worldview they represent.
It helps that the bubble is bursting just as tech’s allies in the Trump administration are also feeling their tower beginning to teeter. Skepticism toward one bleeds easily into the other because AI and fascism are two strands of the same proudly ignorant nihilism. In many ways, they are both reactions to humanity’s terror in the face of climate change. In the United States, individualist denialism has been more appealing than collectivist reality.
But the victory of nihilism is no more inevitable than AI, and there’s no reason why this crash should only serve to further consolidate wealth and power. It is an opportunity for whoever seizes it. As Naomi Klein wrote in The Shock Doctrine (herself quoting Milton Friedman), “He observed that ‘only a crisis – actual or perceived – produces real change. When that crisis occurs, the actions taken depend on the ideas that are lying around.”
Increasingly, there are some good ideas lying around for people to pick up. Zohran Mamdani has some about how we imagine a more socially-oriented society. The new Luddites have some about how technology might be made to serve us, instead of the other way around. Cory Doctorow, Brian Merchant, Meredith Whitaker, Emily Bender, Nilay Patel, Timnit Gebru, and Ed Zitron all have incisive criticisms as well as constructive ideas for the future of tech. Only 54% of Americans (Americans!) have a positive view of capitalism. This is an opening.
So here is my prophecy about what happens when the bubble bursts:
The stock market crumbles, for a time. Many retail investors lose their savings. Many blameless tech workers lose their jobs. The rich and powerful snap up as much as they please, as they always do. So far, it feels depressingly familiar.
But amidst the garment-rending on the pages of the New York Times, there’s a new note of anger and derision. It starts in the comment section and works its way up. The people called to explain this moment aren’t the same people who created it. The story going around is that this was all a con job, and people are sick of being conned.
Politicians unfurl their sails to catch the breeze of this popular anger. “Regulate big tech” becomes a winning proposition in districts both red and blue.
Data centers are left unfinished, just cinderblock ruins of a civilization that never was.
The next time Mark Zuckerberg, or Peter Thiel, or Sam Altman comes oozing back with some shiny piece of trash to sell, they are booed off the stage.
No one wants what they are selling, because they have picked up some ideas they found lying around, and they are busy building something new.
- I will slightly moderate my position here to say that I’m aware that AI does have a small number of unobjectionable use cases, but their relationship to the industry’s hype is roughly the same relationship of a Segway to a space shuttle. ↩︎