Issue 6 / Play

January 01, 2019
The operations room of Cybersyn.

The operations room of Cybersyn.

Network Effects: Raul Espejo on Cybernetic Socialism in Salvador Allende’s Chile

How do you plan production in a socialist society? From 1970 to 1973, Salvador Allende’s government tried to answer this question with Project Cybersyn, known as Project Synco in Spanish. Over the course of three years, an international team of scientists and engineers built a communication and control system for Chilean factories using one central IBM computer and hundreds of Telex machines dispersed throughout the country.

The project was among the casualties of the 1973 coup against Allende. Members of the team were imprisoned or exiled. Raul Espejo, operations director of Project Cybersyn, fled to England, where he became a professor at the University of Lincoln.

Kristen Alfaro, a PhD Candidate in the Department of Cinema Studies at New York University, spoke with Raul in London and Liverpool about the history of Project Cybersyn and its lessons for today.

How did Project Cybersyn begin?

After Salvador Allende was elected in 1970, a friend of mine, Fernando Flores, was appointed technical director of the State Development Corporation (CORFO). CORFO is a Chilean government agency that was created in 1939 to promote economic growth, and was responsible for many nationalized companies.

I joined CORFO as an operations research scientist. We were extremely young, and this was an extraordinary opportunity to do things that people who were much older than us would find difficult. Fernando and I started to talk about how to proceed with the policies of Allende's government, which included a plan to expand the nationalization of Chile's industries with the aim of creating a strategy of economic development that prioritized the production of affordable goods. We wanted to create simple, accessible, and efficient industries that would produce products that everyone could afford.

One way to do that was to centralize the economy. Now, Fernando and I didn’t want centralized planning. But, for at least the first year and a half of Allende's government, that's what CORFO was doing. Still, we thought there were more imaginative ways to bring together government policy with current developments in knowledge. It was in this period that we started to work with Stafford Beer, a British management consultant known for his work on cybernetics. Through Beer, we found how to achieve decentralization through devolving power. We needed to create the conditions where people could express their wishes and interests.

How did you first come across Beer?

It was quite serendipitous that Stafford Beer had published his book Decision and Control in 1966, and that Fernando found it in a bookshop while he was visiting New York City. He brought it back to Chile, and a group of us read and discussed the book. At that time, I was a student of engineering.

The cybernetician W. Ross Ashby’s “law of requisite variety” is central to Beer’s method for controlling complex systems. According to Ashby, variety is the only thing that can control variety. To control a complex system like an industrial economy in order to get it to perform a certain way, you must ensure that the relevant people, such as workers and managers, can respond to all of the possible scenarios that might prevent the system from performing that way.

With great elegance, Beer developed these ideas about how to manage complexity. They appealed to us, because we were not intending to have a centrally planned economy. We were intending to develop organizations that had the capacity to make decisions within themselves to support the development of the economy.

In 1971, Flores wrote to Beer and requested his help to apply cybernetics to the management of Chile’s increasingly nationalized economy. A few months later, Beer arrived in Chile.

What were some of Beer’s specific recommendations as you began developing Cybersyn?

Beer said that we needed to move from a model of running the economy that was based on information to a model that was based on communications. That meant that you weren’t going to ask companies to send reports and have them sit on desk trays until someone had the time to read them. That isn't the way the world works. The world works through communication and on-the-spot interactions.

In one of the earliest ideas he produced for us, he proposed having Telex machines distributed throughout the country. Telex machines were teleprinters that sent text-based messages over a network. The previous government in Chile had bought about 500 Telex machines and didn't know what to do with them. They were in a warehouse without any purpose. It was very serendipitous that someone in the engineering team informed us that we had Telex machines available and he knew how to install them. What would have taken years ended up taking four months to put into place.

At the beginning, the idea was to transmit real-time figures from the factories to CORFO through the network of Telex machines, which we called “Cybernet.” Then the figures would be sent to a computer that ran a software program called "Cyberstride" that analyzed the data.

The outcome of this analysis was displayed in the Cybersyn “operations room.” We worked with industrial designers at CORFO’s Committee for Technological Research (INTEC) to create a room conducive to non-hierarchical management. The chairs, which each had slide-control panels, were placed in a circular arrangement. The room did not have tables; data and graphics were displayed on a panel in the main wall. Everything was designed to facilitate a relaxed environment to produce ideas for the future of Chilean socialism.

The operations room epitomized the Cybersyn project. It was an extraordinary design produced by a transnational team. Many people thought it was too flashy, too technological, and not something that was connected to the workers. I think it was the opposite—the room was designed to facilitate creativity and imagination. It was offering something that couldn't be obtained by doing something trivial like sitting in front of a computer. Today, after all these years, I am amazed. I really think Beer was an imaginative man.

In hindsight, what do you think were some of the limitations of your team?

We had more to learn about the cybernetics of organizations. Stafford Beer knew that subject well, but we were only learning. And when you’re only learning, your bias is towards the technology. So our approach tended to be technocratic. We wanted to design good indices of economic performance. And those indices would be designed by specialists, experts in operational research who could do all the technical aspects of economic modeling in mathematical terms.

However, while that work is necessary and useful, it is not sufficient in the social sense. We needed to get workers much more involved in the meaning of these indices. We needed to give them far more opportunities to influence the design of Cybersyn. To do that, we had to build up rapport with workers—but this wasn’t always possible.

Second, we had a limited understanding of Beer’s concept of the Viable System Model (VSM). The VSM is a management model that works on the premise that managing the complexity of factories, enterprises, sectors, and even whole industries is isomorphic to the way complexity is managed by the human nervous system. Autonomy was the essence of the model. What we struggled with was, how do you recognize autonomous systems within autonomous systems? How do you recognize the capacity of the government to organize the activities of state-owned industry and implement its policies while connecting these activities and policies to the specific dynamics of various factories?

Today, we would do it with more sophisticated methodologies. We would have a better understanding of how to promote autonomy. That is very important, because if you are going to devolve power, then workers, managers, and administrators must understand how to close the loops locally. They must understand how to produce feedback locally. We shouldn’t insist on each unit sending information to the higher echelons and receiving decisions in return from above. Rather, we should explore ways of providing each unit with enough resources to match the complexity of its environment through local decisions.

What are the lessons of Project Cybersyn for today, in the era of Google and Facebook?

Perhaps the most crucial thing that has changed is the abundance of data, and the possibilities for the peer-to-peer coordination of activities. Data is overwhelming us. Companies like Google and Facebook are greatly increasing the variety and the complexity of networks. But we hear very little about how to bring the data they are producing to a manageable level, one that can be managed directly by the people.

The VSM suggested the need to chop up a situational complexity into different autonomous units at a whole range of levels, each with their own decision-making capacity.  A simple system can proliferate a huge number of possible states. We need to learn how to "chunk" those possible states. The more we recognize constraint in the world in which we function, the more we can produce this chunking, and the more we can manage situations that appear to be totally out of hand. Constraint is where we need to focus our attention, because it increases our adaptability to complex situations.

And we need to carry out this chunking and adapt to those situations in real time. One of the most interesting aspects of Project Cybersyn was its attempt to perform economic management in real time. We weren't interested so much in the past as we were interested in the current situation. How do we recognize what needs to be changed today? We need to do cybernetics in time zero rather than trying to respond to past situations.  

What is the role of artificial intelligence in a cybernetic socialist project?

One thing that we need to understand with artificial intelligence is that, in one form or another, a person is responsible for the code. Whose values are then embedded in the code? Developers of artificial intelligence must communicate with citizens to understand their values, so that we are not living with technology created to support a set of values that only meet the needs of a small fraction of society. To build a system that benefits all people, experts must communicate with all different levels of society.

This piece appears in Logic's issue 6, "Play". To order the issue, head on over to our store. To receive future issues, subscribe.