There are certain experiments in the course of human civilization in which the future itself seems to echo. They are the historical prologue of a new social reality. For the present technological age, where computers, digital networks, and algorithms are inescapable, we might count Project Cybersyn as one such artifact.
It was July 1971, the thick of the Cold War, when Stafford Beer, a management consultant based out of London, received a letter from Chile. The South American country had recently elected one Dr. Salvador Allende to power through a coalition of left-wing parties, and he had begun a program of sweeping nationalization. The authorities in charge of the newly nationalized sectors had made a firm decision against copying the Soviet system, whose worrying inefficiencies were even then getting critical analysis from top thinkers and leaders in socialist movements. Instead, they wanted to create a new, modern system of economic planning. Allende and his ministers had turned to Beer to do just this.
Beer was an expert in the field of cybernetics, the science of automatic feedback and control mechanisms. He viewed this framework as having the potential to develop the most advanced methods in scientific management. The letter he received was from an academic, Fernando Flores of the Technological Institute of Chile, who had been teaching cybernetics in Santiago. Flores had been appointed by the new government to oversee the total reorganization of Chile’s public sector, and he saw it as the perfect chance to apply cybernetic principles on the level of a whole national economy. Beer jumped at the opportunity.
Beer and Flores’ team tirelessly drafted up a series of ambitious plans. Besides creating the physical infrastructure for digitally connecting the whole economy, new software had to be created from scratch, as well as the user interface.
It’s important to keep in mind that Chile was considered a Third World country at the time. While ideally they would have procured a computer for each factory and economic institution, much like we would today, the fact is that it was not financially feasible. Instead, they had to make do with a handful of Telex machines that a previous government had acquired but did not know how to use, and two central computers in Santiago.
The idea was that managers in factories would send statistics on inputs and outputs using an index that would record all the connections between industries and enterprises, such that a flowchart for the whole economy could be produced. Sales information and orders for goods, coming both horizontally and vertically, would use the network for transmission. Much of the effort of the Cybersyn designers was put towards creating instantaneous feedback mechanisms to control production from the bottom-up.
The user interface was carefully crafted to ensure that ordinary workers could utilize it, and there were even sociological experiments conducted to document what people responded to best. From Beer’s perspective, the human brain itself was a key participant in the cybernetic recursion, and all the information feeding into it had to be regulated according to what the brain could biologically accept.
All these digital and practical designs were intended to be integrated into a series of ‘Opsrooms’ where controllers—committees of ordinary workers at the lowest level, and state managers at the higher levels—would monitor production. If any identified issue couldn’t be solved by one Opsroom, it would automatically be passed up the chain to the next one. The photograph of the Opsroom commonly associated with Project Cybersyn was the experimental prototype which was designed to exact specifications before it was built. No existing government office could meet the dimensions of the Opsroom, so they were forced to lease the building formerly used as the headquarters for Reader’s Digest in Chile.
The fledgeling network, running on incomplete software, would prove invaluable to the socialist government as the country came under economic siege. International trade was sharply cut off, and sabotage from the CIA and small business owners increased in intensity. Various retailers and distribution centers were already organized into associations called Gremios, many of which owned small trucking fleets. Unlike the oligarchs and foreign managers who had long-since emigrated and formed the basis for Allende’s international isolation, these more local business owners formed the basis for internal political opposition. In particular, they feared the nationalization of their logistical networks so they could be integrated into a planned economy. A trucking strike by these Gremios nearly paralyzed the country, putting the government at risk, but Project Cybersyn allowed Allende’s administration to mobilize all available transportation to move resources where they were needed most. Although the Opsroom was not yet ready, the infrastructure was ready to facilitate a network of communication centers relaying information and transmitting orders across the country in real time. This instantaneous flow of information for economic coordination was a monumental feat, and in some ways surpassed even the most advanced war-time logistical systems of developed countries at the time.
After a military coup led to the death of Allende and put General Augusto Pinochet in power, Cybersyn was destroyed. Chile’s constitution was suspended, and the country became the testing ground for the infamous free market Chicago School boys. The technologically innovative and unprecedentedly democratic system sketched on paper was never fully implemented.
But its logic did not die alongside Allende and the dreams of the Chilean socialists. Within a couple of decades, similar technologies, logistical structures, and forms of economic planning were being constructed by the giants of American capitalism themselves.
Walmart was among the biggest success stories. The company began to pioneer a system of planned coordination among its suppliers through databases of information on costs, demand, and location for its inventory at every level. These ensured suppliers could coordinate production to meet demand as soon as it was registered.
This was no easy task. A single commodity on the shelves of Walmart could, for example, have a dozen different components, each with a different manufacturer. Say you want to buy a lawn-chair with a seat cushion. There could very well be a different supplier for the tubes of metal and plastic, the screws, and the rubber bands that make the back of the chair, not to mention the cushion itself. Each of these suppliers in turn have their own suppliers for various raw materials, further multiplying the number of inputs. Walmart has made a wildly successful model out of cybernetic planning: harnessing huge amounts of data to update their supply chain with the goal of being able to provide any customer requirement at any time.
Much like how the computer proposed by Charles Babbage in 1837 laid out the future path of computation over a century later, despite never itself being fully materialized, Project Cybersyn has its counterpart in the most advanced logistical systems on earth today. It wasn’t that the successors of Beer and his collaborators sought out to emulate these visions, but that those who came afterwards were forced to rediscover their insights independently. The course of scientific and economic development demanded it.
Up to the appearance of Cybersyn, the economic planning of socialist and capitalist countries alike was done largely by quotas. The USSR would begin a planning period by drafting budgets for its enterprises, both in terms of material resources and money, and demand a certain amount of output. The information fed to planners by enterprises had to be clawed back through a network of party operatives overseeing production at every level, and even then consistently adjusted for misinformation.
Compared to planning, the market made for a much more effective automatic feedback mechanism—if a commodity did not have sufficient demand, the price could fall below the cost of production, forcing less to be made and vice versa. Plans, on the other hand, were generally only adjusted either on far longer timescales, or when necessity demanded otherwise. But primitive market pricing was also imperfect as a feedback and control mechanism.
Markets similarly create miniature cycles of over and under production if they rely solely on price signals for information; the more complicated the final good, with the greater number of inputs, the greater these cycles can be. If a retailer sees a sudden increase in sales of lawn chairs and decides to put in a larger order to the final manufacturer, this information will not reach the parts makers until the final manufacturer makes their own purchases. If the supply of parts is not increased to meet the retail demand in time, then the retailer could be faced with the embarrassment of empty shelves.
Retailers respond by planning and coordinating with suppliers. This solves the problem and also makes them far more competitive against rival retailers who rely on the open market. Walmart built its empire by using data collection and logistical strategy to keep its suppliers responsive and low-cost. Other companies will often end up simply acquiring or creating supply chains outright, a process called vertical integration.
The giants of capitalist economic planning now dominate the contemporary landscape, whether Walmart, Amazon, or Alibaba. But there is a key problem with applying cybernetic principles to social organization: humans ultimately control all the inputs and outputs. When a single team of people organizes production and information, the relationships are direct and face-to-face. A high trust team will more easily work toward the same goal. Even if the team is low trust, it is easier to find out when someone is acting in bad faith. But as these systems scale up, managers and bureaucratic middlemen become necessary. Manipulating even these human systems is easier than manipulating a close team. In a system of metrics, automation, and digital feedback, the incentives to game the system are even higher.
To overcome these misaligned incentives, many of these firms have opted to use the same technologies to introduce new metrics to control the old ones. This increasingly takes the form of new and invasive methods of surveillance. For example, Walmart’s database system is in part based on the data pulled directly from cash registers. To ensure this information is being reported correctly, Walmart has an AI monitoring agent that watches the registers. However, this AI is notoriously error prone—it’s more likely to disrupt innocent interactions and cause headaches than actually catch shoplifters.
Amazon maintains even more stringent surveillance of its own employees in order to maximize speed and hit its targets. By tracking the rate of scanned items per minute for each employee in its distribution centers, Amazon can automatically issue warnings when workers fall behind the breakneck pace. This surveillance is more successful. But instead of breaking the feedback loops, the feedback loops break the human beings with exceptionally large numbers of serious injuries in the workplace. When you cannot expect people to cooperate naturally with the cybernetic system, the only choice is to reduce them to almost literal machine-cogs with minimal agency.
Walmart, Amazon, and their cohort managed to solve half the problem. Their technologies are more effective at economic planning than anything Cybersyn had access to. What they have not managed to solve is the tendency of these systems to escape human judgment. The result is gaming, cheating, and the creation of alienated workers and managers alike—be it on a Muscovite assembly line or at a Walmart supercenter.
If the surveillance systems and sophisticated methods of social control employed by these firms are not effective, the whole basis of the system is in jeopardy. Workers, consumers, and managers will find glitches that they can exploit. However, if the systems are effective, then we will begin to see a return to the Fordist era of production, a mechanistic world with new forms of the assembly line. Observing the development of the factory system in its infancy, Marx noted that the organizing of workers and the rise of trade unions was a natural consequence of this “socialization of production.” Bringing so many workers together into a central location, linking them together in direct cooperation, and forcing upon them intolerable conditions naturally spurred an organized response. It was no coincidence that the end of Fordism oversaw a vast reduction in unionization.
This uneasiness of cybernetic planning and dictatorship also applies to the state, and for the same reasons. China is one of the biggest pioneers in cyber-systems of social control, especially as wildcat strikes have become far more common since 2014 compared to the previous decades of its industrialization. The implementation of these cybernetic systems has also, on occasion, broken in spectacular fashion. The most notable of these breaks occurred when local officials in Wuhan undermined the system which health authorities used to monitor novel pneumonia cases, hiding the initial COVID-19 outbreak. The system was state of the art, and was supposed to be automatic, feeding information to central authorities from hospitals much in the same way Chile’s Opsroom would receive production statistics directly from factories. But instead of inputting information into the system, hospitals turned towards local officials—probably aware that their necks were on the line as well if they caused an embarrassment for Beijing.
The driving force behind this kind of digital securitization is not the technology itself, but the class dynamics in which its development plays out. Owners, users, regulators, and developers all participate in these lines of conflict and respond to them. Just as Cybersyn was intended to reproduce the values of Chilean socialism, the technologies of Amazon and Walmart ultimately reproduce those of capital.
It is no wonder, then, that the original theoretical work which elevated cybernetic planning to a vision for the future integrated it into a political system of direct worker control over the use and management of these technologies. Beyond Cybersyn, this is reflected in the work of computer scientists and Marxian economists Paul Cockshott and Allin Cottrell in their 1993 book Towards a New Socialism. It describes a society with all the same science as ours, with all the same desires for consumption. One might go to work, have a family, and shop at the local supermarket just as they would today. It is a world of direct worker control over the cybernetic apparatus and where class divides between owners, users, and creators no longer define their development. The book was originally written in the late 1980s as a program of reform for the communist bloc and would have been translated into Russian soon after, had it not been for the fall of the Soviet Union.
The cybernetic socialist commonwealth is, on paper, a difficult system to cheat. The algorithms that govern it are seemingly free of the systemic errors of Soviet Communism. But we can assume that if there are incentives to cheat it, people will find a way. This is the same principal-agent problem we have observed in capitalist cybernetic planning systems.
Instead of relying on Orwellian surveillance and social control methods to solve this problem, Cockshott and Cottrell look to find ways to put ordinary workers in the driver’s seat. They admit that previous theoretical and practical attempts—whether based around parliaments or parties—have failed. They propose a new political form based around classical, direct democracy on both the macro and micro levels. While these mechanisms likely require vigorous development in their implementation, it is a first step towards reconciling human control with the possibilities of cybernetic systems-building.
Cockshott and Cottrell sketched out a basic set of structural changes in how such a system might operate. On the micro-level, the committee members and administrators who would run institutions and enterprises would be chosen by sortition—that is, random selection—to prevent the emergence of an economic elite. Your average worker would therefore be expected to hold several important positions in their workplace, community or government in their lifetime, much like how every Athenian citizen would be expected to serve as a member of the assembly at some point. On the macro-level, economists would draft multiple plans and craft budgets for sectors not subject to the consumer market—such as healthcare, education, R&D and investment—which would be subject to approval by larger representative bodies.
The control mechanisms of such a society would be designed with the same goals as Project Cybersyn: to streamline and simplify such that anyone could wield them. Institutions based on sortition have a similar educational outlook as a successful bureaucracy. Instead of relying on particular geniuses to come along to lead in some spectacular fashion, everyone must be trained and prepared to take on leadership roles at a standard level. This is not a new notion, but similar to changes affected by the historical Prussian bureaucracy to improve it against its French rivals. Although more specialized roles could only be filled with a more limited number of people, the overall ethos nonetheless changes as a result. Within the organized body working toward a particular mission, everyone has the ability to replace or be replaced by someone else.
Compare this ethos of administration by those doing the actual work with administration in a standard company. A simple concrete example: anyone who is familiar with popular business software—whether a database, accounting, or correspondence—can attest to the clumsy and confusing user interface so common to all of them. The interests expressed in these tools are not ones of accessibility or ease of use for those working with them, but the incentive for ongoing profits from a service relationship.The expectation is that professionals and managers be specially trained, as well as possessing superior intelligence, and therefore perfectly capable of wading through any obtuseness.
While the cybernetics approach of both capitalist firms and authoritarian states have hardly mastered the worker as a node of feedback, we can be certain that a cybernetic society remains the future. China’s high-tech systems were unable to prevent the COVID-19 pandemic, but they were robust enough to contain it domestically. U.S. policymakers had access to fairly advanced systems in theory, but acted in accord with factional political incentives in practice, as well as outsourcing decisions to international bodies. Our epoch is commonly referred to as the Information Age; but what makes the information of our present period unique is its integration into cybernetic systems. Our history is the history of the computer, of Cybersyn, of Walmart’s database, Amazon’s warehouses, China’s health reporting, and even Cockshott and Cottrel’s utopian vision.
It is a truly odd situation where the political economy of cybernetics is repeating itself. For small- and medium-sized businesses, the threat of massive computerized enterprises swallowing them up is no longer embodied in socialist projects like Allende’s Cybersyn, but rather in their multinational corporate rivals.
Business owners, including those with small businesses, were essential to the rise of neoliberalism both in Chile and the U.S. In both cases, their chief program was the atomization of production and the disruption of unions. However, the return of consolidation and the Fordist production framework is already a reality. The pressure on consumption caused by a falling rate of profit was arrested by the neoliberal era, but only partially. A return to the era of working class agitation and industrial titans will doom the class status of small and medium business owners, weakened as they already are by the fallout of COVID-19.
Should these business owners gather enough political strength to shut down this development, they will be remembered much like those classes of bygone eras who shut down technological progress in order to preserve their lifestyles and their power, whether it be the Roman emperors, the merchants of Florence, or the medieval guilds. The future where local and national-level business interests succeed in using anti-trust to break up what they perceive as technological threats, manage to collect further subsidies from the government to fund their existence, or secure special exemptions for the hyper-exploitation of labor, is the future that will be handed to Chinese planners on a silver platter.
What if they should fail in this endeavor? In that case, an opportunity might arise to let the needs of human labor shape the future of these technologies. Rather than serving as a band-aid solution to surveil, discipline, or replace labor, it can be harmonized with it. Technological development can be conditioned by something other than the dynamics of class self-preservation. The question will then be who dares first to integrate the cybernetic economy with the worker, rather than the other way around.