Janus, the Roman god of beginnings and endings, did not have two heads, but two faces fused into one divine being. One face stared into the past, the other turned to the future. So too does artificial intelligence have two faces fused into one extraordinarily powerful network of machines. In this duality lies the answer to the questions we most often ask today: is AI just a bubble that will soon burst? Or will its unstoppable progress prove the skeptics wrong? The answer to both of these questions is yes.
The face looking back, weary and furrowed by memory, truly represents a bubble about to burst. This side of artificial intelligence participates in classic capitalist commodity production. Subscribe to ChatGPT is sold like any other commodity, like a subscription to The New York Times or car leasing. And then in January last year, a Chinese startup DeepSeek disrupted the AI product market by offering a free service, far cheaper to produce and just as good quality as ChatGPTThis has shattered the dreams of big tech companies of having a monopoly on AI goods (or subscription services).
Today, we have solid evidence that the value of companies like OpenAI, which make money from selling AI subscriptions, is inflated. While the costs of large language models (Large Language Models, LLM) are growing exponentially, subscription revenue is growing mostly linearly. This is unsustainable and the moment of truth is approaching when the wrinkled, backward-looking face of AI will wither and perhaps die.
But what about the other face of artificial intelligence? When DeepSeek started the race, AI companies began to move their technologies from the capitalist market sector into the existing technofeudal economy. Long before LLMs were able to mimic human interaction, cloud capital interfaces like Alexa, Siri, Google Assistant, or Amazon’s recommendation algorithm were already allowing big tech companies to collect hefty cloud rents from all of us. As soon as you enter amazon.com, you have effectively exited capitalism. This is where the other face of artificial intelligence amplifies the extractive power of big tech companies.
Imagine what an advanced AI bot like Alexa will soon be able to do. It will know you perfectly, and you won't be able to tell whether it's a human or a machine. It will remember everything you've ever said, done, or wanted. It will perfectly remember your every whim, musical preference, shopping habit, and disappointment. It will be there for you through thick and thin. It will be able to predict when one of its recommendations might upset you. It will push you to the very edge of tolerability, without ever going overboard.
It will be completely different from ordinary surveillance capitalism, the current system in which machines spy on you so that advertisers can more precisely target their ads. With AI interfaces, Amazon, Meta and Google will soon speak to you in ways that are unimaginably sophisticated today, just as the program did eight years ago. AlphaZero astonished computer scientists when he taught himself to play chess, condensing 3.000 years of human chess experience into a single day.
Thus, powerful and almost human-like AI bots like Alexa will cause neurodependence in their users. They will become an indispensable part of our mental everyday life. Switching to another bot will not be a simple change of store where we buy e-services, but a painful divorce. This is where all hope for the future of market competition dies.
Massive economies of scale create natural monopolies. It’s one of the few useful things we teach economics students. That’s why a city can’t have multiple water companies, each with its own water pipes running down its streets and buildings. For similar reasons, a handful of tech giants own and control the means to produce AI. For any company or startup that wants to build serious AI capabilities, access to infrastructure is not just a convenience, it’s a fundamental necessity. That gives tech giants an incalculable influence not just on the pace and cost, but also on the direction in which the other face of AI is headed.
Forward-looking, clear-eyed, hopeful, and immune to the bursting of the AI bubble, the other face of artificial intelligence is looking down paths no one has yet taken, toward a future we can’t yet imagine. Interestingly, these are areas far removed from anything we would describe as a capitalist market. There is ample evidence that this is already happening. One example is Instacart, a corporation that has brought traditional grocery stores into the technofeudal sector by spreading AI cloud capital across thousands of retail businesses of all sizes and locations. The result is extreme price discrimination that nullifies the elementary features of the capitalist market.
This practice of “perfect price optimization” has already been adopted by all serious retailers in the United States and beyond. Instacart not only does it extract what economists call “consumer surplus” (the difference between the price a consumer pays and the maximum price they are willing to pay) but, more importantly, it virtually prevents price competition. Capable of remembering every pattern of consumer behavior, this platform applies a pricing model that makes it impossible to compare prices across customers and across time periods, as well as across stores and platforms. Not only that, Instacart It allows both physical retailers and digital platforms to collude - sometimes even without their managers' knowledge - in ways that are not obvious. The system is designed to be unnoticeable while dismantling any meaningful market mechanism.
The other face of artificial intelligence, the technofeudal one, prevails, and will continue to do so after every AI bubble burst. Between the uncertain profits that every startup like DeepSeek can be canceled overnight, and the cloud rents that AI machines can capture in the long term, the tech giants have opted for the latter. Any company that wants to make money from delivering AI goods today will have to move to extracting cloud rents in the technofeudal sector or it will fail.
What does this mean for us, for the future of humanity? Techno-optimists are convinced that artificial intelligence will open up new horizons of pleasure, productivity, and wealth. I do not share their enthusiasm.
The traditional capitalist and AI-technofeudal sectors of the economy are in dynamic conflict with each other: value is created in the technofeudal sector, and is usurped by the traditional sector. When profitability in the capitalist sector rises above a certain threshold, AI-powered cloud capital accumulates excessively in the technofeudal sector, and the rate of rent extraction from the cloud eventually declines. And when rent extraction rates from the AI cloud fall below a further threshold, productive capital (including AI machines) moves into the capitalist sector, which drives down the rate of profit. And so on in a circle.
Similar to the predator-prey relationship: if predators overeat, the prey will become scarcer, and they themselves will begin to starve. The prey population then recovers, which gives predators a new opportunity to hunt. Ultimately, this circular dynamic leads to systemic collapse.
My grim forecast is based on a bad hunch. The greatest achievement of the tech giants is that their machines and AI algorithms have managed to recruit billions of us to work for free - by posting our videos and texts, writing reviews - which increases their capacity to extract profits from cloud rent. But what happens if AI cloud rents rise relative to the cost of wages at big tech companies? The overall spending power of society falls, so do the profit rates of capitalist enterprises, and eventually the cloud rents of big tech companies fall. The instability and structural stagnation that have always been endemic in capitalism thus escalate into a cycle of alternating collapses in cloud rents and capitalist profits.
It's the story of the other side of artificial intelligence and how an amazing technology with Promethean potential will end up like the story of Daedalus. In ancient myth, Prometheus was the personification of the hope that technology would help humanity. Daedalus was himself a Promethean hero, a brilliant engineer who built a labyrinth for King Minos to trap the Minotaur, but was also a victim of his own constructions. When he invented flying machines for himself and his son Icarus to escape Crete, the venture met a tragic end. So it is with artificial intelligence today. Promethean engineers conceived and developed it, but the other, uglier side of artificial intelligence trapped them - like all of us - in an ultra-complex, all-encompassing labyrinth of a technological system.
And so, when friends ask me why I engage in electoral politics, which I dislike, I reply that there is only one path from Daedalus’s trap to the Promethean ideal: radical democratic politics. That means starting with small, regulatory steps—for example, by passing interoperability laws or repealing regulations that strengthen the already enormous power of tech giants—and then moving on to bigger tasks: building a digital monetary commons and rethinking ownership rights over data and capital in the cloud. Only then will we have a chance to turn artificial intelligence into a benevolent human assistant.
(Translated by Milica Jovanović)
See more:
Download the app and follow the news
FOLLOW US ON
