The Digital Yoke: Cybernetics and Labor


As we find ourselves nearing “full employment”, we also see a rise in “precarious employment”. This is all in the context of a rising “gig economy” (ex-“on-demand economy”, ex-“sharing-economy”) that has no sign of disappearing or shrinking any time soon. Technorati tell the rest of us to “program or be programmed” while a shrinking middle class grasps for spare dollars from the elite via service jobs pushed to them from their phones. What went wrong here? Weren’t computers supposed to turn us into knowledge workers? Weren’t machines supposed to usher a period of leisure in human history?

How did this transition occur? How is it enforced? I would argue that the key is cybernetics. This present condition can be synthesized from prognostications of cyberneticists working during the past 70 years.


Cybernetics was born in Vannevar Bush’s “Iron Triangle” during World War 2. The setting is fitting – World War 2 was a war driven by information and information computing. It was a war where weather predictions directly affected strategy, leading the Allies to plan the assault on Normandy under specific meteorological conditions1. It was a war where England and Germany competed for the best cryptography and cryptography breakers, and where the leap in computing power brought about by Turing saved lives and took months off the war. Information, its flow, and its control proved to be as powerful as advances in ballistic weaponry. (In many ways, an informational advantage was like having a temporal advantage.)

Norbert Wiener was a leader in the development of early cybernetics. He infused ballistic weaponry with information. Wiener outfitted anti-aircraft guns with technology so that they could aid the human operator by automating the calculations for how much to lead a given airplane based on its distance and velocity.

He saw that integrating the gun-pointer (the human) and the gun could follow two different paths. Using the anti-aircraft gun as an example, ought the machine interpret the human’s intent and calculate the lead, or ought the human interpret the machine’s intent? Conceptually, it is the same: there’s sensors reacting to feedback and ‘successive switching’ – does it matter whether it is a human machine or a metallic machine? Wiener argues that the nervous system is “not only a computing machine but a control machine”.

And this was a transformative idea. Wiener describes this as issuing a “new industrial revolution … replacing human judgment and discrimination at low levels [emphasis mine] by the discrimination of the machine. The machine appears now, not as a source of power, but as a source of control and a source of communication.”2 This new industrial revolution was not about power or energy; it was about information.

Despite Wiener’s belief that cybernetics could benefit humanity, he already had premonitions that it could be problematic. He warns against “[eliminating] the human element in these tasks”, for we might telescope the worker out of the picture, rather than improve the conditions of the worker. Further, he called on leaders of business and government to consider human leisure as their responsibility and their priority.

It goes without saying that his admonishments fell on deaf ears.

Piercing Cyberspace’s Barrier – How Cybernetic Systems Became Embodied via Human Labor

The contemporary labor context is different from what likely was imagined by the Iron Triangle’s cyberneticists. Their context was one of a middle-class composed of manufacturing labor which would increasingly be either aided or supplanted by machines.

Today, the new middle/upperclass is ‘information/knowledge’ workers. The upperclass exclusively deals in information — what is finance other than information now that banking and trading is so computerized? The new middle class either makes tools for the upperclass, or does arbitrary work that can be bet on by the upperclass via financial instruments.

Non-fiat money is little more than information. We’re so far from a ‘ground truth’ via financial instruments and borrowed money that it is impossible to go back to a representative currency, or even to properly audit the value vs actual output of One can view the global economy as a large cybernetic system where financial algorithms act like homeostats, optimizing the allocation of money (for banks).

Meanwhile, the old middle class is under attack and shrinking. While the middle-class in America first hit its peak at the height of manufacturing, the “first world” no longer subsists on manufacturing. The rise of the global economy has sent much unskilled and semi-skilled labor overseas. The result is that many workers are unempowered (in many ways, but especially) financially.

Manufacturing was the industry that drove robotics research. In its absence, robotics has been reserved for highly specialized purposes and for R&D.

Alongside the rise of computerized banking and the financial system’s investment in Silicon Valley, we’ve seen an enormous rise in computing power and capability and availability.

In the shadow of concentrated capital and a national policy and cultural shift towards an “app economy”, we have tens (arguably low-hundreds) of millions of workers falling by the wayside, making ends meet by performing service work on an on-demand schedule of an on-demand type, often switching between delivering food, building furniture, driving people and more within a very short period of time (in some cases, the same day).

With the rise of the internet and computing, there grew a conceptual barrier between cyberspace and “real life”. And immediately, people wanted to pierce that barrier. One of the first internet art projects was Ken Goldberg’s Telegarden (1995). In this piece, he hooked a garden-tending robot up to computers and a webcam. Viewers could visit a website and order the robot to act and view the actions via the webcam. Nichols felt this “tangible embodiment” was inevitable: “Conceptual metaphors take on tangible embodiment through discursive practices and institutional apparatuses. Such practices give a metaphor historical weight and ideological power. Tangible embodiment has always been a conscious goal of the cybernetic imagination.”3 A twist on this Nichols passage is that in contemporary practice, humans are becoming the embodiment of cybernetic systems. Controlling robots via code intuitively makes sense – they share etiologies and, in a certain sense, biologies. Donna Haraway’s Cyborg Manifesto both identifies information as the vehicle across information and humans, and predicts how it will be used: “communications sciences and modern biologies are constructed by a common move - the translation of the world into a problem of coding, a search for a common language in which all resistance to instrumental control disappears and all heterogeneity can be submitted to disassembly, reassembly, investment, and exchange.”4

But what qualities does the human laborer possess when acting as a cybernetic embodiment? Nichols describes the consequences: “In cybernetic systems … In place of human intersubjectivity we discover a systems interface, a boundary between cyborgs that selectively passes information but without introducing questions of consciousness or the unconscious, desire or will, empathy or conscience, saved in simulated forms.”5

Computing advanced much faster than robotics. Cybernetic systems want physical manifestations (/human engineers want their code to have physical actions on the world outside of the silicon). Now humans are cheap and highly able, quickly repurpose-able robots. Wiener’s fears came true.

This all plays out in a perversion of where mechanization was intended to take us. The industrial revolution replaced energy of man and animals with machines. The early cybernetic revolution replaces human judgment and discrimination with that of machine — intended to be the repetitive/laborious aspects of judgment and discrimination. But in the cybernetics of late capitalism, we have machines employing their best judgment and discrimination to extract the most value from human laborers — an inversion of the industrial revolution.


So where are we now, in our late-capitalist cybernetic economy?

We’ve transitioned from dominant media companies to dominant media platforms. The media platforms do not monetarily compensate content creators or curators for their labor. Instead, the platforms offer analytics (information) and attention (eyes, clicks, action).

However, the gig economy of the 2010s complicates contemporary work. These articles assumed that we would be mechanizing more labor, turning more humans towards knowledge work, where what we've instead seen is a bifurcation: we have high-salaried employee knowledge workers devising the control systems for the on-demand contracted manual laborer. For the contracted laborer, suddenly, software is used for the subjective judgment and the person is used as little more than a very complicated actuator. At what point does a combination of policy and artificial scarcity turn the contracted laborer into a biological machine? And what does it mean that, after we replaced the energy of steam with the energy of electricity, we return to the energy of humans?

As for the advertisers, growth hackers, UX designers, and programmers, they are the new white-collar workers. These are the people designing and implementing the systems themselves. They have become gatekeepers, or instruments of the gatekeepers, working in the realm of information. Gatekeeping used to be primarily the realm of those controlling money; now those with control over the flow of information have similar amounts of power. Although controlling flow of information often is capitalized and thus associated with control of money, there are cases where purely controlling information can be hugely influential, cf Edward Snowden, the Panama Papers, and so on.

Katherine Hayles remarked that “The Marxian monster of metal and flesh would just be updated to that of a world-spanning network where computers use human beings as a way to allow the system of machinery” to function and imagined “immobile bodies of the hackers, electrodes like umbilical cords connecting them to the matrix, appendixes to a living, all-powerful cyberspace”6. Cyberspace was not realized as then envisioned. We retain the image of the immobile white-collar worker tranced by a screen. But now networked computing technology is mobile, and many markets of ’cyberspace’ are embodied via humans carrying phones. These humans are directed by cyberspace instead of directing cyberspace themselves.

Case Studies

In the following case studies, I will highlight systems designed to subvert, deaden, or obviate human intelligence and judgment.

Social Media — the montage of the cybernetic age?

Facebook, Twitter, Tumblr, have embedded the concept of streams of media curated by humans (informed by data from the platform, and secondarily curated by machines reacting to human information) in our cultural practice, likely for a long time (I wouldn’t be surprised if it lasted a generation or more). On these platforms, content from all over the web, content created by users, and content captured from other mediums by users, can be posted, reappropriated, recontextualized, remixed, etc. Per Henry Jenkins, we now can expect to see all media on all channels. And, based on the ‘feed’ model and the algorithms powering it, we now can see and expect to see anything next to anything else, algorithms, filter bubbles, and user self-selection notwithstanding. Content creators and curators now get feedback they did not have access to prior to these platforms. Now we see impressions, reactions, and in some cases, the types of reactions. Now that all the services have at least partially non-linear feeds, users may even guess as to what content is prioritized or deprioritized by the platform, which itself is making decisions heavily based on user actions and reactions. It’s a cycle that shapes headlines into clickbait which then shapes off-platform action that is suited to clickbait style headlines. We need look no farther than Trump to see an example of someone who has understood how to tailor his brand of outrageousness to varying mediums, platforms, and cultural contexts. And now he has set new records for free press exposure – in the mid 10-figure range – by hammering on catch-phrases that are intended to trend. (Most recently “playing the ‘woman-card’”.)

Lazzaratto again nails the double-productive function of the user of these platforms. “The general public tends to become the model for the consumer (audience/client). The public (in the sense of the user - the reader, the music listener, the television audience) whom the author addresses has as such a double productive function. In the first place, as the addressee of the ideological product, the public is a constitutive element of the production process. In the second place, the public is productive by means of the reception that gives the product "a place in life" (in other words, integrates it into social communication) and allows it to live and evolve. Reception is thus, from this point of view, a creative act and an integrative part of the product. The transformation of the product into a commodity cannot abolish this double process of "creativity"; it must rather assume it as it is, and attempt to control it and subordinate it to its own values.”7 But now the user has a triple-productive function: their actions generate data that hones the system and can be sold to third-parties (mostly advertisers, but also more nefarious companies like Palantir).

Uber & other on-demand labor platforms - Mediated Judgment and Control

Wiener warned that machines could telescope the worker out of the picture. The on-demand labor platform is a case where machines control the physical labor and high-level decision-making of humans, who exist only to drive the car where the computer tells them to. At least until self-driving cars are mass-deployed in 5-10 years. The human is only used because computers cannot perceive driving conditions, which can be highly variable and unpredictable, as well as humans. Perhaps the most damning quality is that many of these platforms do not allow the workers to turn down jobs. The workers cannot exercise judgment for preference, for efficiency with their lives (does a driver want their last drive of the night to take them three hours out of the city?), or otherwise appropriately react to ground truths that the system may not know about (say, protests in downtown). This is the sort of judgment I would argue that an human most ought to be able to exercise. On top of that, some of these platforms have ‘friendliness’ as part of their brand, requiring the employee to smile, bump fists, and offer hospitality to their customers. Lazzaratto summarizes it best: “The management mandate to "become subjects of communication" threatens to be even more totalitarian than the earlier rigid division between mental and manual labor (ideas and execution), because capitalism seeks to involve even the worker's personality and subjectivity within the production of value. Capital wants a situation where command resides within the subject him- or herself, and within the communicative process. The worker is to be responsible for his or her own control and motivation within the work group without a foreman needing to intervene, and the foreman's role is redefined into that of a facilitator. In fact, employers are extremely worried by the double problem this creates: on one hand, they are forced to recognize the autonomy and freedom of labor as the only possible form of cooperation in production, but on the other hand, at the same time, they are obliged (a life-and-death necessity for the capitalist) not to "redistribute" the power that the new quality of labor and its organization imply. Today's management thinking takes workers' subjectivity into consideration only in order to codify it in line with the requirements of production. And once again this phase of transformation succeeds in concealing the fact that the individual and collective interests of workers and those of the company are not identical.”8

Drones - Mediated Ethics

How do you kill people with a computer?

The advent of drones welcomed an age of mediated cyberwarfare. In some ways, this is Wiener’s cybernetic ideal – the human only has to make high-level decisions, then machines carry out the minute details (eg how to fly, how to fire a missile [and then in the case of the missile, how to home into its target]) and the physical labor (of flying, of firing the missile). Yet the human operator also is being controlled. The human operator sees a plain screen with a slowly updating photo-feed of high-altitude low-resolution images from the drone’s point of view. It’s cold, clinical, impersonal. Recall the tactics that are employed. The “double-tap” – the practice of striking twice: once for the targets, again for the responders – would it be possible to convince troops on the ground to fire a mortar in such a fashion? The system “decides” (or the system architect has pre-decided) what information the drone operator has access to, and I would argue that it is designed to occlude emotional response.

Online Dating - Cybernetic Affect

How can one use a computer to make people fall in love (or “love”)?

Online dating is an odd ouroborosian practice. Assuming a monogamous context, companies need to tantalize users with the potential of sex or relationships, yet cannot deliver long-term relationships because those users then become inactive. Tinder functions by enticing users act in a mechanized fashion in which they are sorting and selecting each other and performing affection for each other while they are also performing for the company and its advertisers. At the same time, we must be aware that the company, in its black boxes, is sorting and selecting users as well. It also has control over rates of matching, and how many people each user sees per day. It also uses Big Data to (attempt to) assess a user’s desirability and present them with similarly assessed matches. Writing about how this reinforces unhealthy social attitudes (eg, racism, sizeism, cis-sexism, etc) could occupy its own paper.


We dreamt of the internet bringing about a collective intelligence and instead it realized a system of machine-ordered human labor. What is the place of humans? What is it that humans are now good for to capital and to machines? And how do we subvert these systems controlling our information and, in turn, us?

  1. Wark, McKenzie. Climate Science as Planetary Infrastructure. ↩︎

  2. Wiener, Norbert. “Men, Machines, and the World About” in Medicine and Science, 13- 28, New York Academy of Medicine and Science, ed. I. Galderston,New York: International Universities Press, 1954. ↩︎

  3. Nichols, Bill. “The Work of Culture in the Age of Cybernetic Systems” in Screen 21 (1): 22-46. Winter 1988. ↩︎

  4. Haraway, Donna. Simians, Cyborgs, and Women: The Reinvention of Nature (Kindle Locations 3350-3352). Taylor and Francis. Kindle Edition. ↩︎

  5. Nichols ↩︎

  6. Terranova, Tiziana. “Free Labor: Producing Culture for the Digital Economy” in Social Text, 63 (Volume 18, Number 2), Summer 2000, pp. 33-58 (Article). ↩︎

  7. Lazzarato, Maurizio. Immaterial Labor. Radical Thought in Italy: A Potential Politics, 1996. ↩︎

  8. Lazzaratto ↩︎