Liberty Machines and Dark Tech

Liberty Machines and Dark Tech

An interview with Evgeny Morozov on The Santiago Boys, a podcast about a project inside the Allende government that tried to use cybernetics to manage an economy under assault.

A computer rendering of the Project Cybersyn Operations Room (Wikimedia Commons)

Fifty years ago, the democratically elected socialist government of Chile led by Salvador Allende was overthrown in a U.S.-backed military coup. In the ensuing decades, the Allende government, in partial power for just three years, has become a touchstone for leftists around the world. One significant part of that legacy is Project Cybersyn, an ambitious effort to manage newly nationalized state firms amid economic warfare waged by both the U.S. government and right-wing forces within Chile.

On his new podcast, Evgeny Morozov uses extensive archival research, secondary sources, and hundreds of interviews to reconstruct this story. The Santiago Boys—a title that draws a sharp contrast with the “Chicago Boys,” the group of neoliberal figures who helped reshape the Chilean economy during the dictatorship of Augusto Pinochet—is about the people who built Project Cybersyn. Morozov places this history against the bigger backdrop of the Allende government: its aspirations, its internal tensions, its travails, and its enemies. This context gives us a richer account of what has sometimes become a simple story about a utopian, futuristic, and computerized system of economic planning and management. It sheds new light on what Cybersyn teaches us about aspirations to build socialism in the twenty-first century. 


Nick Serpe: What was Project Cybersyn?

Evgeny Morozov: Project Cybersyn—short for “cybernetic synergy”—aimed to aid the Chilean state in managing the enterprises being nationalized by the Unidad Popular government. A significant hurdle was the lack of sufficient managerial staff to oversee them. Allende’s opponents, including the U.S. ambassador, were making things even harder by encouraging managers and other professionals to flee the country.

As with most science and technology projects, the path toward Cybersyn was not linear. It didn’t emerge as a culmination of some strategic plan to use computers in management; the whole process was more chaotic—and even its name came at a later stage. It all started with an effort to bring some external expertise to Chile. Fernando Flores, a high-ranking member of the Allende administration, felt that he needed help in dealing with all these nationalized companies. So he sought the guidance of the British management consultant Stafford Beer, even going to London to meet him. That encounter resulted in Beer agreeing to go to Chile. This collaboration eventually blossomed into what we now recognize as Project Cybersyn.

Serpe: What were some of the more technical aspects of the project?

Morozov: Project Cybersyn drew on software developed in Britain by Beer’s associates; a couple of U.S.-made computers to run that software; and a network of Telex machines [devices that exchanged text-based messages over telephone networks], which linked the factories participating in the project to the central office in Santiago.

The hope was that, since managerial expertise was scarce, Cybersyn’s computing infrastructure could make up for that. For example, the project’s engineers would model factories by identifying key indicators that best described the productive activity in each. Thus, it would be possible to predict, based on monitoring fluctuations in those indicators, whether a given factory was about to run into major problems with, say, supplies or absenteeism.

This ongoing computer-aided analysis—working off the latest data communicated over the telex network—was supposed to allow managers to make the right interventions at the right time and at the right level of the economic system. Autonomy was very important for Stafford Beer: factories would be given some time to try to resolve emerging problems on their own. Only if they failed to do so would the higher levels of the system—including the ministries—intervene.

Serpe: What is cybernetics? How did it become the framework for dealing with the problems of managing state-owned companies?

Morozov: Cybernetics was an interdisciplinary effort that emerged in the 1940s to make sense of some of the conceptual and technological advances made during the Second World War. At the time, more and more military technologies—think guided missiles or radar systems—were driven by new concepts that, in one way or another, had to do with information.

Take feedback—the notion that you would be able to adjust the performance of the system based on some external moving target or criterion (think of a guided missile, which adjusts its course based on the “feedback” from the target). Soon, it became clear that feedback-powered systems were present everywhere: in society, in the economy, in technology, in our bodies. The insight that these seemingly disparate systems work in similar ways, once analyzed through concepts like feedback and homeostasis, was at the heart of cybernetics.

Beer was too young to be part of these initial discussions in the 1940s. But, starting in the early 1950s, he drew on techniques from cybernetics and operational research, another very important discipline at the time, to develop his own peculiar brand of management theory. It became known as management cybernetics. It celebrated the power of technology-aided simulation as a way to experimentally discover the best course of action under rapidly changing circumstances.

For example, Beer could simulate how a slightly different allocation of resources in a steel factory would affect productivity. Normally, a steel factory is just too big of a place to do expensive experiments in real life. It’s much better to do these experiments using analog machines and, eventually, digital computers. Two decades after his initial pioneering experiments in the steel industry, Beer got a chance to apply them in the context of managing a national economy in Chile.

Serpe: The Allende government was made up of a coalition of parties called Unidad Popular. Many of the Cybersyn participants were affiliated with one of those parties, MAPU, which was more white-collar in terms of the backgrounds of the people involved. How did they see themselves fitting into this coalition of left-wing organizations that was trying to maintain unity in the face of a lot of opposition?

Morozov: While the main story line of The Santiago Boys concerns Cybersyn, there are parallel and occasionally overlapping narratives that illustrate the broader context. One of them is about Allende’s fight for technological sovereignty against the tech giant ITT. Another one examines the long-running debate about the trade-offs between having a classical Leninist party, with its hierarchies and centralized decision-making, and a more bottom-up and decentralized approach, with workers and students leading the way.

MAPU had many currents. It included radicals who studied French theory and Althusser: they were very interested in building a traditional Leninist party. There was another chunk of MAPU that had split from the Christian Democratic Party; they were less interested in Marxist theory and more interested in completing the agrarian reform. And then there was a technocratic layer composed mostly of recent graduates in economics and engineering. All three of these currents, at least initially, opted for the more centralized approach, with its celebration of expertise and discipline.

In the podcast, I contrast MAPU with MIR, a parallel leftist movement that didn’t join Unidad Popular. Its youthful leaders often came from the middle class; many of them studied to become doctors. Yet, for all their credentials and class background, they thought of themselves as enablers of bottom-up organizational activity.

It was interesting to reexamine some of those arguments, because this dilemma of how to strike the right balance between centralization and decentralization resurfaces every decade, with Indymedia, or with Occupy Wall Street, or with the Arab Spring, or with the student protests in Chile. This debate was also at the center of Cybersyn. The project ended up being criticized by some leftists for being too centralized, and not being open enough to worker participation.

Serpe: Many people on the left today are familiar with Cybersyn; it has an almost romantic reputation in some circles. It is often seen in the context of the socialist calculation debate—these big, often technical questions about the feasibility of managing an economy in the absence of markets. But you show that Cybersyn was operating in an environment filled with conflict, including both these tensions within the Allende coalition and forces outside that were militating against Chilean economic performance under Allende.

Morozov: I knew about Project Cybersyn years before I began making the podcast. I wrote a review essay of Eden Medina’s book Cybernetic Revolutionaries in the New Yorker in 2014. I interviewed Fernando Flores, and I had examined Beer’s archive in Liverpool. I thought I had a good understanding of the project’s goals and history. But the two years of working on the podcast have changed my perception of it quite significantly.

In the end, Cybersyn was a tragedy—and a drama. This project started in an optimistic, even utopian political environment. The Santiago Boys worked off the assumption that Allende would be allowed to govern, and they would be able to build a different economy in Chile. These assumptions were quite unrealistic. If you know anything about how ITT, the CIA, local industrialists, the government of Brazil, and other forces were trying to prevent Allende from even coming to office, you would never think that such optimism was warranted—especially when Allende won the election with only one-third of the popular vote and relied on a very unstable coalition of six parties.

Cybersyn unwittingly became a victim of its own early success. The junior technocrats around Flores embraced Beer’s cybernetic thinking. It’s a very technical vocabulary, but it’s also, in Foucauldian terms, an episteme. It allows you to look at realities through a different perspective—one that sees the world through the lens of “variety reduction” or “levels of recursion” or “algedonic signals.” This mélange of cybernetic concepts helped them tackle one of the worst crises to hit Chile at the time: a truck owners’ strike in October 1972, which sought to paralyze the country and create conditions for an anti-Allende coup.

Flores and his Santiago Boys applied cybernetic ways of thinking to manage the strike—and not without some success. This led to Flores occupying ministerial positions in the government. But it also took him away from the project. Initially, the Cybersyn team interpreted this positively, because they thought it would put cybernetics at the very heart of the government.

However, the political, economic, and geopolitical situation was clearly unfavorable to Allende, so the government went on the offensive. There was no space to implement the agenda they were dreaming of in 1971. They continued to accomplish parts of it; they modeled factories, they analyzed the data that was coming in, and they built the Operations Room, a futuristic space that became the calling card of the project. Alas, that room was never actually used as intended.

As the worst of the crisis hit, nobody at the upper levels of the post-Flores CORFO [the government agency overseeing state-owned companies] cared about this cybernetic apparatus for managing the Chilean economy. This is the paradoxical element of the whole endeavor: had the coup not occurred, Project Cybersyn would probably have been shut down by Chile’s own bureaucracy. In fact, at some point its Santiago-based administrator even threatened to pull the plug if the government didn’t change its attitude and commit more resources. Had that happened, would we even remember its glamorous Ops Room today?

If you read the more celebratory accounts of Cybersyn, you tend to get a different picture. But it’s a much more complex and tragic story than the tempting tale of the first instance of fully automated luxury communism being cut short by the coup.

Serpe: Given this track record, what is the relevance of Cybersyn today?

Morozov: What I think is relevant about the project today is how the peculiar Chilean political and economic situation forced them to work in the paradigm of management rather than planning. Planning, as it is discussed in the socialist calculation debate, normally has a very particular focus: the allocation of goods. You want to produce and allocate goods in a way that’s more rational than the market. The tendency, naturally, is to simplify and rationalize. You want to use computers to establish control, and the cybernetic paradigm naturally lends itself to this search for control, one of its key concepts.

But Beer saw that the world was becoming increasingly complex; the simplicity and rationality that is usually prized by socialists was neither possible nor desirable. For Beer, complexity was not something to be eliminated; it was something to be embraced and managed. Thus, instead of using computers to streamline the distribution of goods and services, you can use them to learn how to live with more complexity without letting it overwhelm your administrative system.

Serpe: The tension between technocracy and democracy runs through the entire podcast series. One component of this is that there were many bottom-up projects around the same time that were attempting to deal with the issues of allocation, production, and consumption, and these efforts sometimes outstripped the government’s position. We see this with factory takeovers by workers, and also with the JAPs, which are local councils that are trying to allocate goods in places where there are shortages, frequently caused intentionally by forces opposed to the government. How did someone like Beer think about this problem, or the criticisms he received from some on the left around that time? Could these developments that were sprouting up from the bottom be assimilated into the state-led project, or were they in conflict with it?

Morozov: Cybersyn was subject to the same tensions as the broader project of Unidad Popular. The coalition was torn between acting in a traditional, centralized way, where the political party sets the direction and enforces discipline, and a more bottom-up approach, which rejects this hierarchical leadership in favor of a more deliberative process, with policy and its execution rooted in the ideals of radical democracy.

In the case of Allende’s Chile, the Communist Party was very disciplined and the most moderate of the big parties in the coalition. It was very much in line with what Moscow was preaching—that there were hard limits to what Allende could achieve. On the other end of the spectrum was MIR (not in the coalition); the radical wing of MAPU (it eventually split into two because of these tensions); and the Socialist Party.

The Communist position was that Allende shouldn’t have flirted with the radicals; that he should have been much more cautious not to provoke the military and a little less provocative in his remarks. The position of the radicals was that Allende should have given more power to the people, created more horizontal forms of power, and maybe even distributed weapons, so that the people would be able to resist the coup and defend democracy. The debate as to which of these positions could have saved Allende’s project rages to this day.

Project Cybersyn embodied those tensions and debates. Its team was rather homogeneous: these were mostly political moderates of a rather technocratic bent. They did not see expertise itself as political, and that was one of the more interesting critiques made of Project Cybersyn in Britain. Critics argued that it wasn’t enough to involve workers in designing the indicators of what happens in their factories because you would just end up with a more sophisticated form of Taylorism: extracting the workers’ knowledge, packaging it into models, and then using those models to better manage the workers and the factories. But shouldn’t workers participate in designing the processes for improving productivity? That activity is deeply political, not just scientific.

The more radical elements in MIR and parts of MAPU were open to the idea that expertise is political. But there were many other problems for them to tackle; contesting the authority of technocrats in the scientific and engineering sphere was not at the top of the agenda. For British radicals, who were more concerned with the responsibility of engineers and scientists as the Vietnam War was raging, this was priority number one. And that’s why Beer found himself in these heated debates in Britain and elsewhere.

Neither Beer nor his colleagues seemed to grasp what these radical critics were complaining about. This shows how a celebration of the neutrality of technical expertise can become an ideology. But it does matter who decides how many chairs and what kind of chairs and what kind of screens are put in that Operations Room. For Beer, all those decisions about the design of the managerial system were in the realm of science. They were his decisions to make, for he was the expert and the scientist.

Funnily enough, Beer never even finished college, so he was not really a card-carrying scientist, even though he was a very erudite and well-read man. How much of his arrogance was a product of preaching cybernetics and how much of it came from his upper-class background?

Serpe: It’s striking how many instances there are throughout the podcast of attempts by Chilean elites, the military, far-right groups, and of course the U.S. government, from the Nixon White House to the CIA, to undermine this government. Over the course of your research, did you see paths that weren’t taken? Or did you leave with more of a sense of how overwhelming the odds were against the Allende government? 

Morozov: A country’s political trajectory can be affected by how well it understands the destabilizing and anti-government uses of seemingly innocent technologies. Could Allende, for example, have made it harder for his enemies to monitor and surveil what he was doing? Could he have taken more steps to eliminate the possibilities for spying on the government’s own networks? At one point the government asked ITT to check Allende’s office for bugs—here we are talking about letting a company that was fighting Allende into rooms and spaces where they definitely shouldn’t have been present.

I think the radicals in Chile’s political debate—people from MIR, Allende’s daughter, Beatriz Allende, some parts of MAPU—were more sophisticated about such matters. They understood how amoral their enemies were and knew that they would not stop at anything to overthrow Allende. The same applied to their analysis of the military. Allende, to the last day, was a big believer in the constitutionality of the military.

As for Cybersyn, with the October Strike over, Beer did get this idea that maybe they should be using cybernetics to empower the base, i.e., to have a more democratic way for people to react to what’s happening in the community and in the country. But what was needed was an effort by the government to mobilize the masses to support the broader political project and defend it from the coup and foreign adversaries. Cybernetic thinking could have clearly helped in that. As some of the British leftists were saying, if you’re building a project like this, you shouldn’t be doing it secretly; you should be doing it in public, and you should be mobilizing people to help you.

The lesson we have learned in Latin America is that military coups fail only if you manage to mobilize the masses. Allende, I think, ultimately thought that it was still a fair fight where the success of the project would be judged by its economic performance. But the game was much dirtier. Perhaps they should have created some department of popular mobilization and used Cybersyn and cybernetic concepts to fire up the base to resist the coup.

Serpe: To describe the opposition Allende faced, you use the concept of dark tech—a sort of stylized distinction between something like Cybersyn, what Beer called “liberty machines,” and the tech used by other side. At one point, you do an almost forensic investigation to figure out if the literal infrastructure used by Cybersyn was appropriated by the Pinochet government. But it seems to me that the real question about the dark side of these technologies is much bigger. Do you make yourself vulnerable by using certain technologies, or is it just about being outmatched by the technical capacity of your enemies?

Morozov: The concept of dark tech has to do with the need to always be aware that people who work for the police, the military, and big corporations are usually five steps ahead of the leftist activists. On everything, from strategy to technology.

The traditional mythology around Project Cybersyn sends an almost opposite message, presenting it as a socialist internet, or a proto-internet, far ahead of its times. But one can only accept that myth if one knows nothing about the tech apparatus of repression that had existed in the region way before Allende became president, much of it build with American money from institutions like USAID. And it leveraged the latest technologies to track down and eliminate leftist guerrillas.

If you know something about the use of Telex networks, computers, databases, and operation rooms in that earlier repressive project, you see Project Cybersyn in a different light. In the pre-Allende Chile, the police had already interlinked some of its departments across the country using Telex machines, even building a functional operations room to rival that of Cybersyn. If we really want to apply the concept of the internet to the Telex network, it all started with a repressive internet used for policing—the socialist internet came later.

In that sense, my invocation of “dark tech” is supposed to sensitize utopian leftists to the importance of counter-intelligence work—they need to understand what their enemies are doing. This requires a knowledge of history, and more sophistication about how power operates.

I’ve just spent a week in Brazil, where, during Lula’s second term in office, there was a lot of excitement about the promise of free software. Everybody was celebrating free software as a big revolution. In the end, they lost that battle to the companies. They lost that battle because the issue was not framed geopolitically. Instead of linking it to questions of national autonomy and technological sovereignty, they framed it using the tropes of American debates about software licenses, morality, and ethics. In the Latin American context, you needed a very different set of arguments.

I’m not trying to force parallels. All I’m saying is that there is no shortage of utopian tech projects in the region. The concept of dark tech points to the importance of geopolitical and ideological counter-power that needs to be built.

Serpe: While listening to The Santiago Boys, I revisited Patricio Guzmán’s trilogy The Battle of Chile—his documentaries about the Allende years and how things fell apart. One of the most remarkable things about the films is their depiction of the high level of sophistication, organization, and mobilization of working-class people, of peasants, across the country—which still wasn’t enough to avert a coup. It’s good for us to think today about what an alternate economic system, an alternate technological system, would look like, but we often do so in the absence of political organization strong enough to contest corporate power, to challenge international capital and empire. What kind of living legacy does Cybersyn offer under contemporary conditions?

Morozov: We have to face reality, and the reality is that Allende came to power in 1970, and he had almost three years to avoid a coup. He failed. We have to understand why he failed, and we have to learn from those experiences. So, I don’t think that the coup was pre-ordained in some sense. There were errors both among the moderates and among the radicals, but there are also false memories and mythologies that prevent us from seeing those errors.

This also applies to initiatives like Project Cybersyn. We don’t know, for example, if there was any real interest on Allende’s part in moving the operations room that was developed through Project Cybersyn into the presidential palace. Some believe that Allende thought that it would have helped to manage the crisis. This is hard to believe, because the operations room that did exist wasn’t connected to anything. It was a very nice prototype, but moving that prototype to the presidential palace on the eve of the coup would not have helped anything. Nor did it help Allende survive the October Strike, because in October 1972 the operations room didn’t even exist. But not everyone—including those who had a direct or indirect role in the project—remembers the exact sequence.

At a higher level of abstraction that would take us beyond Cybersyn, what remains valid in the Chilean experience is the idea that technological sovereignty is a prerequisite to developing economic and national sovereignty. That argument is very clear in Allende’s speeches; it’s very clear in his actions against ITT.

For me, Cybersyn is interesting as an entry point into that broader universe of what I call the “Santiago school of technology”—a counterpart to the Chicago school of economics (hence the Santiago Boys to match the Chicago Boys). That’s why Chile’s struggle for technological sovereignty is so important; so is the legacy of CEPAL, the Economic Commission for Latin America and the Caribbean, which is still a United Nations institution in Santiago, and their critique of orthodox economics and trade theory, and their fruitful exchanges with dependency theorists. To me, the Santiago school furnishes a framework for thinking about technology geopolitically and geo-economically—and we need just such a framework today.

The chance to go and read the texts of dependency theory of the 1960s, some of them in Spanish and Portuguese and some of them in English, has been a revelation for me. They helped me grasp the challenges of artificial intelligence and quantum computing, in an international scene where the global economy is dependent on the United States, in a way that I couldn’t grasp with any other theoretical or analytical framework. Cybersyn emerged in that context, with economists and engineers pushing for reclaiming chunks of the economy and decision-making from the market and trying to manage them for the purpose of building national industries and national infrastructures, as a way to promote national development.

I hope that my podcast will help recover the importance of the Santiago school and its radical political and economic project—a project that got smashed by the military and the Chicago Boys. Had the political situation in Chile been less turbulent (with less pressure from the United States, with less domestic terrorism, etc.), and had Project Cybersyn delivered, Chile may have ended up as a fully industrialized socialist Latin American democracy. It could have developed what South Korea or Taiwan achieved later on: a robust, state-steered, heavily industrialized, technologically savvy, national economy.

Allende’s predecessor, [Christian Democrat] Eduardo Frei, had planted the seeds for it. But Allende accelerated the process by giving CORFO, the state development agency, a more strategic direction, as well as bringing a lot of radical economists on board. They built, for example, a semiconductor factory in the north of the country, in Arica. They had an agenda, which was rolling against the logic of the market; rolling against the idea that you should just specialize in wine or shrimp because you are a country that doesn’t have industry and you have good weather and you have an ocean. You should actually build things to move up the industrialization ladder. And you could have done it in a socialist country run by a socialist president with a lot of technological, managerial expertise, furnished by the likes of Stafford Beer, and with a lot of radical economists and engineers.

We could have lived in a world of the Santiago Boys, and not of the Chicago Boys. If anything, this would have spared us a false dichotomy between the Soviet experience and the Wall Street experience as the only two possibilities. It’s either the Central Plan and Stalin and gulags, or it’s full-blown neoliberalism, and there’s nothing in between, right? That’s just false, and the Chilean experience could have shown us that something else is possible. The lesson to draw here is not to abandon the project of the Santiago Boys; it’s to understand its limitations and try to infuse its managerial and cybernetic logics with the spirit of radical, bottom-up democracy. And to do so without losing sight of geopolitics and dark tech.

Evgeny Morozov is the author of The Santiago Boys, a new podcast that explores Salvador Allende’s tech legacy. He is also the founder and publisher of the Syllabus, a nonprofit knowledge curation initiative.

Nick Serpe is senior editor of Dissent.

On Thursday, September 21 at 12:30 p.m., Evgeny Morozov will be speaking about The Santiago Boys at the Open Society Foundations in New York City. Click here for more info.