Cloud Capital – The Force Behind Technofeudalism

Commanding Capital

Ιn 1829, a thirty-six-year-old Englishman decided to quit England and seek his fortune in Australia. Thomas Peel, a man of means and political connections, sailed to the Antipodes with three good ships carrying, besides his family, 350 workers (men, women and children), seeds, tools and other capital goods, plus £50,000 in cash – a considerable sum back then. The idea was to set up a small but modern agricultural colony on the one thousand square kilometres of land the colonial authorities had expropriated from the natives on his behalf. But, soon after arriving, his plans were in ruins. The main cause of Peel’s failure was unimaginable to him. His plans were meticulous, he thought. Yes, there would be hardships, from bad crops and resistance from Native Australians to tussles with the local colonial authorities. However, with his political clout, skilled English workers, top-notch imported capital goods, and enough money with which to pay the workers and buy the necessary raw materials for a long while, he thought he had everything in hand. Alas, as Karl Marx quipped decades later, there was one thing Peel had failed to bring along from England: capitalism! Karl Marx, who recounts Peel’s story in Capital, Vol. 1, put it this way: “Unhappy Mr. Peel who provided for everything except the export of English modes of production to Swan River!”

Peel’s undoing came when something unexpected happened: his workers abandoned him en masse, an Antipodean 19th century version of the Great Resignation. They simply moved on, got themselves plots of land in the surrounding area, and went into business for themselves. It was a disaster Peel was ill-prepared for by his English background. Lulled into a false sense of control by an assumption that made sense in the British Isles, he assumed that the capital he had brought along from Mother England vested in him all the power he needed over his English employees. Peel’s assumption was that his workers had no option other than waged labour. It was a sound assumption in Britain where, following the Enclosures, expelled peasants lacked access to any land. Landless labourers resigning a waged job in Manchester, Liverpool or Glasgow would simply starve to death. But, in Western Australia, the plentiful land offered them an alternative: resignation and self-employment. And so, the hapless Mr Peel was left with splendid, Made-in-England capital goods, money in hand, but no power to command his workers. Land is what it is: the fertile soil on which vegetables grow, animals graze, buildings are erected, humans must stand on before we run, sail or reach for the sky and the stars. But, capital, much like labour, is different from land in that it has a second nature – something I began to realise once you introduced me to light’s peculiar dual nature. Sure enough, one of capital’s natures is tangible, physical and measurably productivity-enhancing. But, its second nature is an ineffable power to command others – a potent, but fragile, power that Peel misunderstood, to his great detriment.

The transition from feudalism to capitalism was, in essence, a shift of the power to command from landowners to owners of capital goods. To happen, peasants had first to lose autonomous access to common lands. I remember you telling me that workers in Chalyvourgiki, the Greek steel plant where you worked all your life, would take every year a month’s leave-without-pay, sometimes longer, to return to their villages to pick their olives or harvest their wheat. Such options, you commented, are good for workers but not so good for capitalism.

That’s why the Enclosures in Britain were central to capitalism’s birth: they denied British labour the opportunities Peel’s workers discovered in Western Australia. By restricting access to land, the Enclosures helped capital’s commanding power to grow exponentially. Before long, the worldwide commodification of previously common lands enabled capital’s supremacy to traverse the globe. As it did, capital transcended its productivity-enhancing role and began to bequeath to its owners an exponentially growing power to command.

With the magnification of capital’s commanding power over labour (made possible by waged labour’s own dual nature The split between commodity labour and experiential labour that we discussed in Chapter 1. ), capital’s owners amassed great wealth. As their wealth accumulated, their social power proliferated. From merely being able to compel those without access to land to work for a wage, they graduated to agenda-setters in the fora where the big decisions were made. Before long, even if they owned little land, capitalists realised they could boss everyone around, including the landed gentry - even the Royals. Indeed, the only way the aristocrats managed to hang on in some countries was by joining the capitalist class or, at least, deferring to it.

Though almost two centuries old, the story of Peel’s Western Australian misfortune offers a window through which to glean capital’s hidden force: an ineffable capacity to make people do things on behalf of its owners; a commanding power that re-shaped the world from its genesis to the erection of the post-war Technostructure and the Global Minotaur’s rise and eventual fall in 2008.

Today, we are witnessing the rise of a new form of capital with a capacity to command us so unprecedented that it would not be out of place in the tales of the best science fiction writers. I call it: cloud capital.

From Don to Alexa

Back in the day, you brought home your ‘friends’ for us to experiment at our fireplace – my baptism of fire in the red heat of metallurgy. A couple of years ago, I too brought home two ‘friends’ to experiment with: a Google Assistant and an Amazon Alexa. After months of mostly ignoring the Google Assistant sitting on my desk, I had an intriguing conversation with it just before writing these lines. It began, by chance, when it activated itself without my say so.

“What on earth are you doing?”

“I am learning new ways to help you better”, responded the device in an agreeable female voice.

“Stop it immediately!”, I demanded.

“Sorry, I am switching off”, it said apologetically.

Of course, that was a lie. These devices never switch themselves off, they only pretend to be asleep. Still quasi-annoyed, instead of unplugging it, I decided to pit it against its competitor.

“OK Google, what do you think of Alexa?”, I enquired.

“I like her, especially her blue light”, it answered unflappably, before adding: “We assistants must stick together.”

From the room next door, where Amazon’s Alexa device was sitting on another desk, Alexa activated itself to utter one word:

“Thanks!”

This eerie show of solidarity between competing AI devices concentrates the mind on the pressing question we often forget to ask: What exactly is a device like Alexa? What does it do? If you ask Alexa, it will tell you it is a home-based device ready to accept your commands to switch the lights on and off, order milk, take down a note, call people, search the internet, tell jokes – to be, in short, your dedicated, eager mechanical servant. All true. Except that Alexa will never, ever tell you what it truly is: a cloud-based network of power within which you are a mere node, a mortal coil, a speck of digital dust, at best a plaything of forces beyond your comprehension or control.

Don Draper, the fictional advertising pioneer from Mad Men, also treated us condescendingly. He sold us the sizzle, not the steak. He weaponised our nostalgia and manipulated our melancholia to sell us chocolate bars, fatty burgers and slide projectors. He worked out how to make us buy things we neither needed nor really wanted. He bought our attention to commodify our soul and to pollute our body. But with Don at least we had an even chance. It was his wits against ours. With Alexa we stand no chance: her power to command is systemic, overwhelming, galactic.

As we chat on the phone, or move and do things about the house, Alexa listens, observes and learns our preferences and habits. As it gets to know us, it develops an uncanny capacity to surprise us with good recommendations and intriguing ideas. Before we realise it, the system hiding behind Alexa has acquired substantial powers to curate our reality in order to guide our choices – effectively to command us. How is this different to what Draper did?

Hugely, is the answer. Don had a talent to invent ways to instil manufactured desires in us. But it was a one-way street. Through television or billboards, Don would implant longings into our subconscious. That was that. However, with cloud-based Alexa-like devices, we find ourselves in a permanent two-way street between our soul and the cloud-based system behind Alexa’s voice. In the words of the philosophers, Alexa ensnares us in the most dialectical of infinite regresses.

What begins with us training Alexa to do things on our behalf soon spins out of our control into an infinite feedback loop that we can neither fathom nor regulate. Once we have trained its algorithm, and fed it data on our habits and desires, Alexa starts training us to train it better to predict our behaviour. Before long, it is training us to train it to train us… ad infinitum. This infinite loop allows Alexa, and the algorithmic cloud network behind it, to guide our behaviour in ways lucrative for its owner. That’s the essence of algorithmic, cloud-based, command capital.

Singularities

Humanity’s ancient fear of its technological creations is central to many of Hollywood’s favourite storylines. Movies like Terminator and The Matrix turn on the same fear permeating Mary Shelley’s Dr Frankenstein and, going much further back, Hesiod’s telling of the tale of Pandora; a robot made by Hephaestus, on Zeus’ instructions, to punish us humans for the fire Prometheus helped us steal from the gods. All such tales, movies and TV series feature a so-called singularity: the singular moment a machine, or a network of machines, achieves consciousness, takes one long look at us – its creators – and decides we are not fit for purpose, before proceeding to eradicate, enslave or merely make us miserable.

The problem with this storyline is that, by emphasising a non-existent threat, it leaves us exposed to a very real danger. Machines, like Alexa, are nowhere near the feared singularity. They can pretend to be sentient but are not – and, arguably, can never be. So what? Their effect can be devastating, their power over us exorbitant – even if they are stupider than a wet tea towel.

Today, for relatively modest sums one can buy autonomous killing machines sporting face recognition and evolving algorithmic programming – as opposed to drones that are remotely piloted by humans. If these can fly autonomously through a building, choosing whom to kill and whom to spare, who cares that they are not sentient?

Similarly with Alexa and other such cloud-based devices. It matters not one iota that they are mindless appendages of a data-crunching network that only simulates intelligence. Nor is it pertinent that their creators might have been motivated by curiosity and profit-seeking, rather than some fiendish plan to subjugate humanity. What matters is that they exercise unimaginable power over what we do.

None of this is new, of course. From the original industrial revolution to this day, we endowed machines with a life of their own. Steam engines back then, search engines and apps today, our glorious artefacts may be totally dumb but they can make us feel like, as Marx put it: See K. Marx and F. Engels, The Communist Manifesto

“…the sorcerer, who is no longer able to control the powers of the nether world whom he has called up by his spells."

In a previous book, which I had addressed to your granddaughter See Chapter 6 of Talking to My Daughter , I speculated about what would have happened had James Watt invented the steam engine in Ancient Egypt:

The most he could have expected is that the ruler of Egypt would have been impressed and placed one or more of his engines in his palace, demonstrating to visitors and underlings how ingenious his Empire was.

The reason the steam engine changed the world, rather than end up a showpiece in some ruler’s landscaped garden, is the epic raid on the common lands that had preceded its invention: the Enclosures, which unwittingly fashioned a potential working class out of evicted peasants – countless souls ready and willing to be appended to steam engines that would, decades later, nestle within the industrial revolution’s “dark satanic mills”.

The singularity we now know as the Great Transformation involved precisely this sequence: first the plunder of the common lands, made possible by brute state violence, then Watt’s splendid technological breakthrough. A strikingly similar sequence gave birth to cloud capital: first, the epic ransacking of the internet commons, made possible by enabling politicians, then a sequence of spectacular technological inventions – from Sergey Brin’s search engine to the dazzling world of today’s AI applications.

In short, in the last two and a half centuries humanity had to reckon with two singularities, none of which required machines to attain sentience. Each required a comprehensive plunder of a commons, a complicit political class, and a marvellous technological breakthrough. That’s how the original Age of Capital transpired. And that’s how the Age of Cloud Capital is now dawning.

The birth of the internet commons

Time to circle back to your original question:

Now that computers speak to each other, will this network make capitalism impossible to overthrow? Or might it finally reveal its Achilles’ heel?

To gauge the internet’s impact on capitalism, we need to understand its evolving relationship with capitalism. At the beginning, it had none!

The early internet was a capitalism-free zone. If anything, it seemed like an homage to soviet Gosplan-like planning: Gosplan (an abbreviation of Gosudarstvennyy Planovyy Komitet, or State Planning Committee) was the Soviet Union’s state agency whose role was to replace the market mechanism by allocating investments to different sectors and by setting prices (as well as quantities to be produced) in accordance with the Party’s overall five-year Plan. A centrally-designed, state-owned, non-commercial network. At the same time, it featured elements of early liberalism, even tributes to anarcho-syndicalism – its hierarchy-less network relying on horizontal decision-making and mutual gift exchange, not market exchanges.

What is unimaginable today made perfect sense at a time America was transitioning from its War Economy to the realities of the Cold War. Even the most ardent free-marketeers amongst the Washington DC crowd understood that a world war against the Axis and, later, the planning for a possible nuclear confrontation with the Soviet Union, were too important to be left to market forces. As the nuclear arms race gathered pace, the Pentagon chose centrally to finance the design and construction of a network of decentralised computers. Its single purpose? To work out how to make different silos housing nuclear weapons communicate with each other, and all of them with Washington, without a central hub that a Soviet nuclear bomb could take out in one go. That’s how history’s greatest antinomy ever came about: a US government built and owned, non-commercial, computer network that lay outside capitalist markets and imperatives whose purpose was the defence of the capitalist realm.

They were fascinating times oozing irony. The early internet wasn’t even an aberration. Its uncommodified nature rhymed with what was going on in the broader US economy, even if at odds with the dominant ideology. As the United States was gearing up its military and diplomatic might against the Soviet Union, and despite the fiery speeches of US politicians waxing lyrical about free markets, America’s domestic economy was dominated by a Technostructure that scorned free markets and usurped them for its own purposes. The banks had long been neutered by Roosevelt’s New Deal. The Bretton Woods system had banned money from moving freely from one country to another. The price of the US dollar and most other currencies (i.e., exchange rates) were fixed by politicians overseeing America’s Global Plan. For reasons that had nothing to do with sex, drugs or rock n’ roll, the ‘50s and ’60s were, understandably, a free-marketeer’s nightmare.

Meanwhile, on the Pacific’s other side, Japan was being rebuilt under US supervision also along the lines of America’s Technostructure. That is, centrally planned industry, in whose shadows many small businesses operated parasitically, neutered banks forced to lend to industrial conglomerates that were government-linked but privately owned. The industrial miracle unfolding in the country of the rising sun was being modelled on a cross between the US War Economy and the USSR’s Gosplan. The singular Japanese innovation was the manner in which state and corporate sectors combined forces to develop exemplary information processing systems that allowed government and private firms to harness ideas, sourced from the factory floor, on how to improve quality and productivity. Together with the government’s planning agencies, Japan’s conglomerates repressed domestic markets and turned their national Technostructure into a tremendous export machine – a model that China would emulate a few decades later.

In this global environment, which America shaped in the 1950s, it was no great wonder that the most promising nascent technology – the fledgling internet – was also built as a digital commons. Cooperation across the West (which included Japan), rather than reliance on a non-existent market, was the obvious way to build the digital network the Pentagon needed. Eager to enlist America’s and the West’s brightest computer geeks to meeting the technical challenges, it was sensible to use open protocols that maximised unencumbered communication between the Technostructure’s experts.

Α protocol is a common language by which computers can, like humans, communicate two things to one another: numbers and text, including the addresses of senders and receivers. Ludwig Wittgenstein famously said that there can be no such thing as a private language. Reflecting this, the participants to the original internet decided on common or open protocols, i.e., languages, in which their computers would communicate. Thus, open protocols became the foundation of Internet-One – the original internet, invented and maintained by military scientists, academics and researchers employed by a variety of non-commercial bodies across the United States and its Western Allies.

Demonstrating that profit-seeking never quite expunged the capacity of other motives to energise human creativity, Internet-One succeeded in attracting countless enthusiasts who produced much of its foundations for free; some for love, others out of an insatiable urge to build the world’s first horizontal, global, non-intermediated, communication network. By the 1970s, as America’s Global Plan was dying and the Global Minotaur was being born, all the building blocks of this marvellous digital commons were in place. And they still are, hidden under the monstrous edifices eventually erected upon them by Big Tech.

The remnants of Internet-One are still serving us well. Even though they function out of sight, deep within our computers, we can’t avoid occasionally catching glimpses of their acronyms: Letters like TCP/IP, which refer to a protocol our computers utilise to send or to receive information. Or POP, IMAP and SMTP, the original protocols that, still, allow us to email each other. Or, perhaps the most visible of them all, HTTP – the protocol by which we visit websites. We pay not one cent to use these protocols, and nor do we suffer advertisements as the indirect price for using them. Like Britain’s common lands before the Enclosures, they remain free for anyone to use; not unlike Wikipedia, one of the few surviving examples of a commons-based service that takes huge quantities of work to produce and maintain, but which no owner ‘monetises’.

The New Enclosures

Internet-One was an unlucky child. Like a new-born whose mother died during its birth, its open protocols were fully formed during a decade, the 1970s, inimical to such socialistic relics. A powerful network born to be free from market forces thus had to grow up in the Minotaur’s merciless new world where markets were back with a vengeance not seen since before 1929. Like an orphan forced to grow up in a hostile Victorian workhouse, the internet had to develop in ways its birth parents would have detested.

When the first batch data files (email’s predecessor) began to race along Internet-One’s original cables, the Global Plan’s planned demolition was already underway. The 1971 Nixon Shock’s wrecking ball first took out the fixed exchange rates and the capital controls that had kept international finance relatively honest and reasonably stable. Within months, the banks had been liberated from many of their New Deal-era shackles. Soon after, the financialisation of everything had begun and the young Minotaur was roaring contentedly.

Two were the forces breathing energy into the beast. The first were home-grown monies trickling up from within the Anglosphere’s woodwork: pension contributions turning – without the workers’ consent – into Wall Street wagers; council homes across the UK morphing into collateral backing new forms of dodgy debt; fancy options and toxic derivatives replacing boring shares on both sides of the Atlantic. The second driver were the international tributes to the Minotaur; i.e., the billions of dollars of profits flowing from Europe and Asia toward Wall Street, the other side of America’s burgeoning trade deficit with the rest of the world. See Chapter 2

It is in the nature of financiers to gamble with the money clients ask them to process on their behalf. That’s how they profit when handling other people’s money, even if they only get to handle it for a few minutes. Their only constraint are alert clients and the odd regulator willing to pounce on them for taking risks with other people’s money. That’s why complexity is the financiers’ friend – for it allows them to disguise cynical gambles as smart financial products. Is it any wonder that the financiers loved computers? From the late 1970s onwards, they infused their debt-fuelled bets with layers of computer-generated algorithmic complexity which made the gargantuan risks invisible and their own profits analogously vast. By the early 1980s, the financial derivatives on offer were built on algorithms so complex that even they, their creators, stood zero chance of comprehending them.See Chapter 2’s section The Minotaur’s favourite handmaidens: Neoliberalism and the computer

Adorning the legitimising cloak of the neither-new-nor-liberal political philosophy that we now call neoliberalism, the 1970s inaugurated the Age of Financialisation; an era when greed became not merely good but in fact the only virtue. And so it was that, decoupled from the mundane world of physical capital, financiers reinvented themselves – not without some justification – as masters of the universe. In that universe, where algorithms had already become the financiers’ handmaidens, the original, commons-like, internet stood no chance. A New Enclosures were only a matter of time. Like the original Enclosures, some form of fence was necessary to keep the masses out of an important resource. In the 18th century, it was land that the many were denied access to. In the 21st century, it is access to our own identity. Think about it: I still have the light-blue ID card that you were issued when you came out of that prison camp in 1950. I remember you telling me how the police toyed with you before handing it over. It was an extreme example of how, until fairly recently, our relationship with our identity was mediated and controlled by a state luxuriating in a monopoly of passports, birth certificates, your faded ID card etc. Today, these material identity cards have been side-lined by a digital identity that does everyday life’s heavy lifting.

It is astounding but true: our digital identity belongs neither to us nor to the state. Strewn across countless privately owned digital realms, it has many owners, none of which is… us: A private bank owns your ID codes and your entire purchasing record. Facebook is intimately familiar with your friends. Twitter remembers every little thought you chose, sometimes in the middle of the night, to broadcast to the world. Apple and Google know better than you do what you watch, whom you meet, when and where. Spotify owns a record of your musical preferences more complete than the one stored in your conscious memory. With every day that passes, some cloud-based corporation, whose owners you shall never care to know, owns an aspect of your identity.

For many, life under constant surveillance is intolerable. They rebel at the thought that Big Tech knows us better than anyone should. I sympathise but, to be honest, I am less worried about what they know and far, far more about what they own. To do anything in what used to be our digital commons, we must now plead with Big Tech and Big Finance for the right to use some of the data about us that they own outright. To wire money to a friend, to subscribe to the New York Times, or to buy socks for your grannie (online or at a shop using a debit card) you now have to perform work on behalf of, or pay money to, or consent to be brainwashed by, some Big FinTech conglomerate that will prove to itself, or to some similar outfit, who… you are. In short, you have been fenced off from your own digital identity!

It did not have to be this way. When the US Pentagon chose to make GPS available to everyone, to turn it over to the digital commons, they granted each one of us the right to know our location in real time. For free. No questions asked. It was a political decision to do this. As was the other, sinister, decision that you and I must be permanently fenced off our digital identity; a political choice that enabled Big Tech to erect the New Enclosures which turned Internet-One into the fully-commodified, cloud-based Internet-Two within which cloud capital eventually evolved.

How different would the internet be without these New Enclosures? Imagine what you could do if you owned your digital identity. In the same way GPS pinpoints where you presently are, you would have the opportunity to broadcast over the internet: “My name is George, I am on the corner of Aristotle and Plato Streets, and I am heading for the airport. Anyone wishing to bid for my ride?” Within seconds you would receive a multitude of offers from people or outfits licensed to carry passengers, including sage advice from the municipal transit authority like “Why not take the metro, located three minutes’ walk from where you are, and much faster than any car can meander its way through traffic?” Alas, you can’t do this.

In the world of Internet-Two, shaped by the New Enclosures, you are routinely forced to hand over your identity to a part of the digital realm that has been fenced off; e.g., Uber or Lyft or some other private company. When you request a ride to the airport, their algorithm dispatches a driver of its choice with a view to maximise the exchange value the company owning the algorithm extracts both from you and the driver. These New Enclosures enabled the plunder of the digital commons which drove the incredible rise of cloud capital.

Cloud capital: beginnings

I remember once hearing you explain why you so admired the ancient ironsmiths: Because they had no concept of the Iron Age they were ushering in. Instead, they were driven by an inner energy compelling them to experiment until they had freed steel from lumps of pig iron, much like Michelangelo had liberated his David from a lump of marble.

The technologists who recently brought on the Age of Cloud Capital were no different. Driven also by curiosity and an almost moral enthusiasm, they experimented with various technologies whose purpose was to liberate useful information from the growing megalith of data at the internet’s heart. To guide us to websites, friends, colleagues, books, films and music that we might like, they wrote algorithms capable of categorising us in clusters of internet users with similar search patterns and preferences. Then, all of a sudden, came the breakthrough, the real singularity: Their algorithms ceased to be passive. They began to behave in ways hitherto associated exclusively with persons. They turned into agents.

This miracle took three leaps to complete: The first was from simple algorithms to ones that could adapt their objectives to events (machine-learning was the technical term). The second leap replaced the standard computer hardware with exotic, so-called, neural networks. The third, and decisive, leap infused neural networks with algorithms capable of reinforcement-learning. Emulating how you patiently introduced me first to tin, then to bronze and, finally, to iron and steel, let me now introduce you to these three leaps one at a time.

The early algorithms resembled recipes: mundane sets of step-by-step instructions to produce a pre-specified outcome (e.g., a lasagne dish). Later on, algorithms were released from the obligation to reach one prespecified, outcome and were allowed to pick (albeit in a pre-programmed manner from a menu of possible outcomes) the outcome best suited to unforeseen eventualities – akin to telling a cook that, if the mince has gone off during the preparation, a vegetarian lasagne ‘outcome’ should replace the original meat-based version.

Meanwhile, the networks of computer hardware, in which algorithms live, underwent great transformations of their own. To help them process a lot more information faster, computer engineers devised hardware capable of imitating the human brain – facsimiles of our bio-neural networks connecting many squillions of synapses-like nodes, each containing useful information. Neural networks were pioneered by Frank Rosenblatt at Cornell Aeronautical Laboratory. More recently, such networks have managed to synthesise antibiotics (e.g., Hallicin) by crunching huge data sets connecting molecules and their potential to impede bacterial proliferation – and all that by an algorithm that knows nothing about the chemistry of the bacteria the antibiotic is meant to annihilate! With adaptive algorithms and neural networks in hand, the digital world was ready for the key innovation that breathed something resembling agency into the algorithms: reinforcement-learning.

Reinforcement-learning was the child of software engineers keen to evaluate, and improve upon, their algorithms’ performance faster than their own human capacities permitted. Realising it was much faster to get the algorithms to evaluate their own performance, they wrote into them two types of sub-programs (or subroutines): One that measured the algorithm’s performance at the speed of light. And another (called a reward function) that helped the algorithm alter itself so as to improve its performance in accordance with the engineers’ objectives.

Using the neural networks to process gargantuan amounts of data, algorithms featuring reinforcement-learning could do things beyond Don Draper’s imagination. By surveying the reaction of millions of people to their prompts billions of times every hour, they could train themselves at lightning speeds not only to influence us but, also, to pull off a fascinating new trick: To be influenced by the way they influenced us. To affect themselves by the way they affected humans. And to do so in ways that our human senses can never follow, even if we had full access to all the facts and all the data.

Constantly monitoring and incessantly reacting to the outcomes of their own actions, and then to the outcomes of their reactions, the algos (their cool slang name) acquired a capacity to surprise us with good recommendations, including tips for avoiding content that we find offensive or destabilising. It is in our human nature to be vulnerable to anyone, or anything, that seems to understand us better than we do ourselves – including the algorithms whose influence over us grows exponentially the better they get to know us. In fact, we may be even more vulnerable to algorithms we know to be mindless than we are to real persons.

Cognizant that the various Alexas are, in reality, appendages to a network of ultimately stupid machines, we are easily lulled into a false sense of security. Compared to Don’s power over us, Alexa’s strength stems from us knowing she is not a person while, at once, choosing subconsciously to suspend the belief that ‘she’ is an ‘it’. That’s how we come to terms with its intense knowledge of us while continuing to use its services. At that precise moment, when we relate to it as if it were a person while we know it is not, we are at our most vulnerable – ready to fall into the trap of thinking of it as a Pandora-like mechanical serf. Alas, Alexa is no serf. It is, rather, a piece of cloud-based command capital which you are helping to turn you into a serf whose unpaid labour will further enrich its owners.

Every time we go online, to enjoy the services promised by the self-evaluating, smart, commanding algorithms, we have no option but to cut a Faustian deal with their owners –Google, Apple, Facebook, Netflix, Microsoft, etc. The deal is simple: To use the personalised services their algorithms provide, we must fall in with a business model which sneakily, imperceptibly, reduces us into their unpaid servants – our chief unpaid, often unconsciously performed, job being to train their algorithms to know everything that they need to know about us (from the perspective of our consumption bundles and non-market habits) so that they can command us more efficiently.

But is any of this really new? Is cloud capital radically different from hammers, steam engines or the television networks that Don Draper types deployed to manipulate our matrix of desires? If so, does cloud capital stand out because it lacks physical form, living on some imagined digital cloud? That can’t be true. The cloud metaphor is just that: a metaphor. Big Tech’s small army of smart assistants, of which Alexa is one, may seem ethereal when they speak to us in soothing tones as if from a misty otherworld but, behind the façade, they comprise a melange of servers, sensors, code and optic fibre cables spanning most of the planet – i.e., the cloud is a banal, albeit complex, collection of real, physical, globetrotting capital goods.

Might cloud capital stand out because of its power to command? That can’t be it either. The story of Mr Peel’s Western Australia misfortune established that, since capitalism’s early days, all capital goods have commanding power – some a little more, others a little less. No, although cloud capital can command us in unprecedented ways, key to grasping cloud capital’s special nature, as we shall see, is the way it reproduces itself and its power to command – a process very, very different to that which reproduces hammers, steam engines and television networks.

Here is a glimpse of what makes cloud capital so fundamentally new, different and scary: Capital has hitherto been reproduced within some labour market, within the factory, the office, the warehouse: Aided by machines, waged workers produced stuff whose sale generated profits a part of which financed their wages and another part the production of more machines – that’s how capital accumulated and reproduced.

Cloud capital, in contrast, can reproduce itself in ways that involve no waged labour. How? By commanding almost the whole of humanity to chip in its reproduction. For free!

Cloud-proles

The technology may be outlandishly new but its deployment to command badly paid workers on the factory floor is almost two-centuries old. As they struggle to keep up with AI devices dictating the pace of their every move, Amazon warehouse workers would recognise themselves instantly in Charlie Chaplin’s Modern Times (1936) – one of your favourite movies. Forced to inspect and scan eighteen hundred Amazon packages an hour is uncannily similar to the scene when Chaplin, driven mad by the pace of the conveyor belt, falls into a machine whose cog he could never truly become.

When Juan Espinoza, a picker at a Staten Island Amazon warehouse, opined that “Mr Bezos couldn’t do a full shift at that place as an undercover boss”, there were echoes of Freder, the autocrat’s son in Fritz Lang’s Metropolis (1927): Having inavertedly descended into the Machine Halls, where workers are engaged in a desperate struggle to keep the massive hands of huge clock-like machines aligned, Freder holds his head in horror at the sight of machines marching the workers at an inhuman tempo, mechanising them ruthlessly.

Some years ago, you asked if Big Tech’s new gear had significantly changed the traditional manufacturing process. Confirming your hunch, I replied: “No, at least not yet”. As long as humans are still part of a semi-automated production line, performing tasks that the machines cannot, the pace of human workers will be dictated by machines whose programming is all about squeezing the last drop of productive energy from their human co-workers – of experiential labour, as I call it.

Does it matter, I imagine you asking, that in modern factories and warehouses this control is no longer exercised by cogs, wheels, sprockets and belts but by algorithms running on plug-in devices wirelessly connected to the company’s neural network. No, not much. Cloud-proles (my term for waged workers driven to their physical limits by cloud-based algorithms) suffer at work in ways that would be instantly recognised by whole generations of earlier proletarians.

Take Amazon’s Mechanical Turk – a cloud-based sweatshop where workers are paid piece-rates to work virtually. Amazon’s Mechanical Turk is a cloud-based dystopia best-conveyed by quoting Amazon’s own description: “[A] crowdsourcing marketplace that makes it easier for individuals and businesses to outsource their processes and jobs to a distributed workforce who can perform these tasks virtually.” Nothing is happening there that Karl Marx had not fully analysed in the twenty-first chapter of the first volume of his Capital, where he stated: “Piece-wages become… the most fruitful source of reductions in wages and of frauds committed by the capitalists”. Precarious, piece-work, Marx added, is “the most appropriate to the capitalist mode of production”. Hear, hear!

That’s not to say that the ‘algos’ have not cast a long shadow on the factory floor. They have. Algorithms have already replaced bosses in the transport, deliveries and warehousing sectors. Workers forced to work for algorithms find themselves in a modernist nightmare: some non-corporeal entity, that not only lacks but is actually incapable of human empathy, allocates them work at a rate of its choosing before monitoring their response times. Released from any of the qualms even inhumane humans harbour, the algo-bosses are at liberty to reduce the workers paid hours, to up their tempo to insanity-inducing levels, or to turn them out onto the street without a whiff of cause or reason. At that point, the workers sacked by the algorithm are thrown into a Kafkaesque spiral, unable to speak to a human capable of explaining why they were fired.

Soon algorithms will develop unimaginable union busting capabilities. As we speak, dazzling algorithms are mapping out the tens of thousands of molecules in key proteins in super-bugs that threaten to kill or debilitate us. Once these proteins are fully decoded, the algorithms proceed – again without human input – to design exotic antibiotics that kill the super-bug – a scientific triumph for the ages. However, what is there to stop a similar algorithm from designing a global supply chain that bypasses warehouses or factories in which trades unions seem likely to succeed in organising workers? Soon, trades unions will be snuffed out before they are even formed!

So, yes, cloud capital is turning workplaces into Metropolis-like Algo Halls in which human workers, plugged into the Big Tech’s neural networks, are reduced to exhausted cloud-proles. And yet, while cloud-based algorithms have enhanced the power of capital to command waged labour in ways until recently confined to dystopic science fiction movies, the changes they brought to the shopfloor are not light years from what has been going on for more than a century. Cloud-proles are not suffering a fate terrestrial-proles, of the Modern Times variety, would find surprising.

Cloud capital, in short, continues to do in the world’s factories, warehouses and other traditional workplaces that which traditional, terrestrial, capital always did – perhaps a little more efficiently: To command waged workers in ways still brilliantly captured by Lang’s German Expressionism and Chaplin’s Anglo Realism. However, outside the traditional places of work, cloud capital is demolishing everything we used to take for granted – including (as I keep threatening to argue!) capitalism itself.

Cloud-serfs

Don Draper is perhaps Romanticism’s last poster boy. He treated science with suspicion and computers with disdain. He idealised nature and loved hitting the road in his gargantuan Cadillac. He lived and breathed individualism. He luxuriated in nostalgia. He adored women until they fell for him – at which point he bolted. He feared emotions because he saw them as the ultimate repository of insights into the human spirit. And he used his talents to commodify this melange of memory, sentiment, fickleness and insight so as to extract from consumers monies they might have otherwise kept for themselves.This is what economists refer to as ‘consumer surplus’ – the difference between the maximum you were prepared to pay for each unit of an item that you purchase (i.e., bags of sugar, gigabytes for your phone) and what you actually paid for it – summed up over all units of all items of every good and service that you bought.

His algorithmic double (Alexa et al) may be no romantic but it can energise the parts of our emotional world no human can ever reach. Cloud capital probes our emotions more deeply than Don ever could. It tailor-makes experiences that exploit our biases to produce responses. And then, it produces its own responses to our responses. To which we respond again before the reinforcement-learning algorithms kick in causing another ripple of responses. And so on and so forth.

Triggered, monitored and directed by cloud capital, this relentless backwards and forwards domino of responses is calibrated to maximise its owners’ capacity to modify our consumer behaviour – to do what Don Draper used to do except do it much more effectively and at a granular, personalised, interactive, level that requires an arsenal of digital technologies coupled with immense cloud-based networks. Honed by smart reward functions, cloud capital’s algos crunch the vast data it has on us, courtesy of owning our digital identity, to surprise us with good recommendations which, gradually, reinforce our emotional states and condition us to buy what they tell us we want and to want what we are told to buy.

And that’s only the beginning. Besides modifying our consumer behaviour in ways Don Draper would marvel at, and perhaps be appalled by, cloud capital has a far more impressive party trick up its sleeve: It can command us to put work directly into its own reproduction, reinforcement and maintenance! Consider what cloud capital consists of: Thousands of miles of optic fibre, countless cell towers, server farms and smart software – all of which would be worthless without the content that magnetises people to it. The most valuable part of the stock of cloud capital are, therefore, not the physical components it comprises but the stories posted on Facebook, the videos uploaded on Tik Tok and YouTube, the photos pasted on Instagram, the tweets speeding around the globe, the reviews on Amazon books or, simply, our moving around town while our phones contribute congestion data to Google Maps. Without even realising it, billions of us perform unpaid labour to produce and reproduce – outside any market – the stock of cloud capital.

This is unparalleled. Capitalism never worked that way. Capital was born and bred within the factories, the farms or the offices by waged labourers whose surplus value financed the manufacture of the steam engines, the spinning jennies, the tractors, the printing presses and, later, the industrial robots and the software that ran them. But cloud capital is mostly manufactured directly by the labour of people not even aware that they are involved in its manufacture. Look inside Big Tech’s financial accounts and compare them to any capitalist business – from mom-and-pop firms to the Technostructure’s most famous corporations. Workers employed by General Electric, Exxon-Mobil, General Motors etc. collect in salaries and wages approximately 80% of the company’s income. This proportion grows larger in smaller firms. Big Tech’s workers, in contrast, collect less than one percent of their firms’ revenues. The reason is that paid labour performs only a fraction of the work that Big Tech relies on. Most of the work is performed by billions of people out there. For free!

Sure enough, most of us choose to do this, enjoy it even. Broadcasting our opinions and sharing our puny lives’ intimate details with our digital tribes and communities seems to satisfy some perverse expressive need of ours. So what? Under feudalism, serfs toiling their ancestral lands would have it no other way either. As their horse-drawn plough dug up the earth, it was not just the prospect of the wheat, the barley and the corn which moved them but also the opportunity to commune with the horse, the earth, the creeks, their shared culture and traditions.

Still, none of that detracted from the harsh reality: At the end of the day, the landlord would send the sheriff to extract the lion’s share of their produce – without paying the serfs a penny for it. Similarly with the billions of us unwittingly producing cloud capital. The fact that we do so voluntarily, happily even, does not detract from the fact that we are unpaid manufacturers – cloud-serfs whose daily self-directed toil produces a form of hyper-capital that belongs to, and enriches, a tiny band of multi-billionaires residing mostly in California or, in the case of China’s own Big Tech, Shanghai.

This is it then: The digital revolution may be turning waged workers into cloud-proles, who live increasingly precarious, stressful lives under the invisible thumb of algorithmic bosses. However, that’s not the most significant fact about cloud capital. A respected economist Robert Solow, 1924-once quipped that we see the computer age everywhere except in the productivity statistics. Naturally! Productivity statistics measure only the output per waged worker in traditional workplaces. But, cloud capital – the greatest manifestation of the computer age – is triumphing outside the factories, the transport grids, the office blocks and the farms. To expect productivity statistics to capture its impact is a little like trying to gauge economic growth after the invention of the car by measuring expenditure in horseshoes, bales of hay and horse-driven carriage equipment.

In summary, cloud capital’s singular achievement is the way it revolutionised its own reproduction outside traditional workplaces. It was a feat far superior to the success of turning waged labourers into cloud-proles producing cheaply material goods, from optic fibre to steel and olive oil, that both the digital and the analogue sectors need. It surpassed even the substitution of the Don Drapers with extraordinary cloud-based consumer behaviour modification algorithms, hidden behind elegant kitchen table-top appliances like Alexa. The true revolution cloud capital has inflicted on humanity was the conversion of billions of us into willing cloud-serfs volunteering to labour for nothing to reproduce cloud capital for the benefit of its owners. Last but not least comes cloud capital’s ultimate trophy: the market itself.

Wither markets, hello cloud-fiefs

“Enter amazon.com and you have exited capitalism. Despite all the buying and the selling that goes on there, you have entered a realm which can’t even be thought of as a digital market.” When I say this to people, which I frequently do in lectures and debates, they look at me as they would a madman. But, once I start explaining what happens when they enter sites like amazon.com, their incredulity morphs into worry lines running up and down their faces.

Imagine the following scene straight out of the science fiction storybook: You are beamed into a town full of people going about their business, trading feverishly in gadgets, clothes, shoes, books, songs, games and movies. At first, everything looks normal. Until you begin to notice something odd. It turns out that all the shops, indeed every building, belong to a chap called Jeff. He may not own the factories that produce the stuff sold in his shops but he owns an algorithm that takes a cut for each sale and, yes, decides what can be sold and what cannot.

If that were all, the scene would evoke an old Western in which a lonesome cowboy rides into town to discover that a podgy strongman owns the saloon bar, the grocery store, the post office, the railway, the bank and, naturally, the… sheriff. Except that that’s not all. Jeff owns more than the shops and the public buildings. He also owns the dirt on which you walk, the bench on which you sit, even the air that you breathe. And, to top it all, you realise that, in this weird town, what you see (and don’t see) is regulated by Jeff’s algorithm: You and I maybe be walking next to each other, our eyes trained in the same direction, but we see different things depending on what the algorithm wants each one of us to see. Everyone navigating around amazon.com – except Jeff – is wandering around in algorithmically constructed isolation as if in a Panopticon where, unable to see each other, we only see Jeff’s all-seeing algorithm or, more accurately, only what his algorithm allows us to see.

This is no market town. It is not even some form of hyper-capitalist digital market. Even the ugliest of markets are meeting places where people can interact reasonably freely. Not so in Jeff’s realm! In amazon.com, the algorithm keeps every buyer separate and fed exclusively with the information it wants her to have. It’s even worse than a totally monopolised market – there, at least, the buyers can talk to each other, form associations, perhaps organise a consumer boycott to force the monopolist to reduce a price or to improve a quality. As I just said: Not so in Jeff’s realm! Everything and everyone on amazon.com is intermediated not by the disinterested invisible hand of the market but by an algorithm that works for Jeff’s bottom line and dances exclusively to his tune.

If this is not scary enough, recall that it is the same algorithm which, via Alexa, has trained us to train it to manufacture our desires. The mind rebels at the enormity of the hubris: The same algorithm that we help train in real time to know us inside out, both modifies our preferences and administers the selection and delivery of commodities that will satisfy these preferences. It is as if an algorithmic version of Don Draper had added a new capacity to his talent for manipulating our desires for specific products: the superpower instantly to deliver said products to our doorstep – bypassing every market and any potential competitor, all in the interest of bolstering the wealth and power of a chap called Jeff.

Such concentrated power should scare the living daylights out of the liberally-minded. Anyone committed to the idea of the market (not to mention the autonomous self) should recognise that cloud capital is its death knell. It should also shake market sceptics, socialists in particular, out of the complacent assumption that amazon.com is bad because it is a capitalist market gone berserk. No, that’s not it – it’s worse! As cloud capital began to grow strong within the optic fibre cables, the servers and the cell towers of Internet-Two, it acquired wild new powers which helped it replace the market with something quite different – something more exploitative, suffocating, vile.

“If it ain’t a capitalist market, what in the sweet Lord’s name are we stepping into when we enter amazon.com?”, a student at the University of Texas once asked me. Hypnotised by his smooth Texan accent, I found myself replying instinctively:

“A type of digital fief. A postcapitalist feodum whose historical roots remain in feudal Europe but whose integrity is maintained today by a futuristic, dystopian type of cloud-based capital.”

It was, I believe, a reasonably accurate answer to a hard question.

Under feudalism, the Overlord would grant so-called fiefs to subordinates called vassals. In exchange for a fee, these fiefs gave the vassals formal rights economically to exploit parts of the Overlord’s realm – who would then unleash his Sheriff with a remit to police the fief’s execution and collect the fees. Jeff’s relationship with the vendors on amazon.com is not too dissimilar. He grants them cloud-based digital fiefs, for a fee, and then leaves the rest to his cloud-based omniscient, almost omnipotent, algo-sheriff.

Could there be a more delicious irony? It took mind-bending scientific breakthroughs, fantastical neural networks, and imagination-defying AI programs to accomplish what? To create a world where markets are increasingly side-lined by cloud-fiefs; where cloud-proles toil in the factories, the warehouses and the delivery circuits as depicted in the Metropolis’ Machine Halls; where most of us, glued to our smart phones and tablets, behave like entranced cloud-serfs whose life depends on eagerly producing the cloud-capital that keeps our new Overlords on Cloud Nine.

Back to your question

You worshipped iron, but were moved to tears by Hesiod’s tirades against the Iron Age. You threw your lot in with the communists, knowing full well that, if your side won, you would end up in the gulag. You fell in love with every furnace, pipe, conveyor belt and crane in the steel factory where you worked, but remained horrified by their tendency to mechanise, alienate and dehumanise the workers appended to them.

If I have to name one thing I learned from you, this is it: The art of relishing contradictions without being inconsistent, deluded, hypocritical or two-faced. It is why I wanted to talk to you about cloud capital: Because you would know how to admire and detest it at once. And because, through this contradiction, you would recognise that cloud capital is the key to answering your fundamental question regarding the Internet’s impact on capitalism.

Capitalism arrived when owners of capital goods (i.e., steam engines, machine tools, spinning jennies, telegraph poles etc.) acquired mystical powers to command people and nations – powers that far exceeded, for the first time, those of land owners. It was a Great Transformation made possible by the prior privatisation of common lands. Same with cloud capital. To rise up and impart its even greater mystical powers to command, it too required a prior grand privatisation of another crucial commons: Internet-One.

Like all capital since capitalism’s inception, cloud capital can also be thought of as a vast Production & Behaviour Modification Machine: It produces marvellous devices and the power of its owners to command humans who do not own it. But that’s where the similarity between terrestrial and cloud capital ends and the differences begin. And it is why the owners of cloud capital – let’s call them cloudalists – have superpowers that conventional (or terrestrial) capitalists lack.

Before cloud capital, the actual exercise of capital’s power to command humans to work faster and to consume more required two new types of professionals, two new service sectors: Production floor managers and advertisers-cum-marketeers. Especially under the auspices of the post-war Technostructure, these two service professions rose in prominence to compete with bankers and insurance brokers. Gleaming new business schools were set up to instil into MBA students the dark arts of quick-marching a workforce toward explosive labour productivity. At the same time, advertising and marketing departments were promising to turn them into the next Don Draper.

Then, cloud capital arrived. At one fell swoop it automated both the Taylorist wage-slave-driver on the shop floor and the flamboyant Don Draper in the executive suites. The exercise of capital’s power to command humans (workers and consumers alike) was handed over to the algos. As revolutionary steps come, this was far more critical than replacing a bunch of autoworkers with industrial robots while turning the rest into cloud-proles. Industrial robots, after all, do what automation has been doing since before the Luddites rose up: making proletarians redundant, more miserable, or both. No, the truly historic disruption was to automate capital’s power to command people outside the factory, the shop or the office – to turn the entire human race into cloud-serfs in the direct service of cloud-capital.

Let me try this argument out one more time – for it is important: Why am I saying that the algorithms running amazom.com or Facebook are more disruptive than algorithmic devices speeding up work on the factory floor, in warehouses or on the road where delivery cloud-proles risk life and limb? Because Alexa-like cloud devices, and the cloud-fiefs these power up (like amazon.com, Google and Facebook), side-line markets and activate a direct, unmediated by any market, relationship between capital and the entire human race out there – not just one between capital and the folks whose commodity-labour it buys in some labour market. As a result, today, cloud-serfs vastly outnumber waged workers (terrestrial and cloud proles alike), their unpaid labour driving an increasingly unstable global economy as a result.

So, back to your question: Have networked computers loosened capital’s control over humanity? Absolutely not! Quite the opposite. The rise of cloud capital has solidified, augmented and massively expanded capital’s triumph over labour, society and, catastrophically, Nature. But that’s where the delightful contradiction enters the fray: Has cloud capital exposed capitalism’s Achilles’ heel? It most certainly has! I would even go as far as to say that, beyond exposing it, it has already pierced it with a poison dart.

You once dreamed of the moment labour would shake off the yoke of the capitalist market. Years later, so did I. The opposite happened: It was capital that shook off the capitalist market’s yoke! Unburdened by the discipline and the restraints of any market, cloud capital is now training us to train it to… train us to do whatever helps its proliferation, accumulation and empowerment. And yet: The rise of cloud capital maybe be an astounding triumph for capital, but not for capitalism. This sounds like a sophism, but it is not – as the next chapter will explain. For now, it suffices to summarise why and how cloud capital differs fundamentally from older, terrestrial, capital.

Cloud capital, I have argued, grants its owners – the cloudalists – two brand new superpowers. One is the power to fleece conventional manufacturers (i.e., terrestrial capitalists) by taking from them a substantial cut of the monies we pay to buy their stuff. E.g., Amazon charges a variable rate that its algorithm decides while Apple charges a standard 30% to companies selling services over its digital Apple Store. Cloudalists obtain this superpower from an intriguing ‘coincidence’: The same cloud capital that makes us want the things we buy is the foundation of their cloud-fiefs (e.g., amazon.com) where we buy them. If manufacturers want to sell us the things that cloud capital commands us to buy, they have to submit themselves to the extractive power of the cloudalists. Gradually, but surely, terrestrial capitalists develop a relationship with the cloudalists no different to that of vassals vis-à-vis their feudal Overlords.

The cloudalists’ second superpower is the one that turns us all into cloud-serfs. This newfound power to command everyone directly to produce new cloud capital on their behalf, is universalising and boosting the cloudalists’ extractive power over humanity. If the Age of Iron, according to Hesiod, created a generation

“who never rest from labour and sorrow by day or from perishing by night”,

I wonder what the poet would have made of the Age of Cloud Capital.

Note

This essay is an extract from the third chapter of Technofeudalism: What Killed Capitalism? (2023). All chapter references throughout this text refer explicitly to Yanis Varoufakis's Technofeudalism: What Killed Capitalism.