Project Cybersyn was meant to be announced to the world on Wednesday 14 February 1973, when Stafford Beer gave the Richard Goodman Memorial Lecture at Brighton Polytechnic under the title “Fanfare for Effective Freedom: Cybernetic Praxis in Government”, and in Santiago the same day. The Observer got there five weeks early, on Sunday 7 January, with a headline, “Chile run by computer”, and a story by its science correspondent Nigel Hawkes which said that “the first computer system designed to control an entire economy has been secretly brought into operation in Chile” and had been “assembled in some secrecy so as to avoid opposition charges of ‘Big Brother’ tactics.” Those two sentences are quoted here from Eden Medina’s book, which cites the article; the Observer of 1973 is not open to automated readers, and the piece was not read for this entry. Beer believed he had been betrayed by someone he had briefed on condition of silence until the lecture.

What follows is what the British technical press made of the project once it was allowed to speak, from the one issue of the New Scientist that can be read in full online.

The draft

The lecture was drafted in Surrey and telexed to Santiago for comment, and the Chilean team did not like all of it. Medina quotes the telexes in Beer’s papers. Raúl Espejo, Fernando Flores and Herman Schwember agreed that cybernetic thinking might help a government increase social stability, but asked whether instability was not sometimes the point: “Historical development is a succession of equilibriums and unequilibriums,” Espejo telexed, and disequilibrium “might be indispensable.” They objected to framing a revolution as a control problem. “The social phenomena goes further than the control problem,” Espejo wrote; “there is for instance the problem of power.” A cybernetics that looked only at control and ignored power, he said, ran “the danger that cybernetics might be used for social repression.” Beer answered: “I cannot write the next book in this one lecture.” He kept his framing, and chose to meet the charge of repression by arguing that the system promoted freedom, in a sense of the word he then had to define. Cybersyn, he argued, computed “effective freedom”: the greatest autonomy an enterprise could have without threatening the viability of the whole. “The polarity between centralization and decentralization—one masquerading as oppression and the other as freedom—is a myth,” he wrote. “Even if the homeostatic balance point turns out not to be always computable, it surely exists.” The warning signals that passed a plant’s trouble upward were a threat to the plant’s freedom, but a necessary one: “The body politic cannot sustain the risk of autonomic inaction any more than we can as human beings.”

Three hundred people came. On 16 February Beer telexed the project’s coordinator in Santiago, Sonia Mordojovich: “The Brighton Lecture was a huge success. 300 people came. Chilean ambassador was delighted. Also his wife.” On 17 February he wrote to Flores that it had “attracted much attention. I have given six radio interviews about it.”

Hanlon, the report

The New Scientist dated 15 February 1973, volume 57, number 833, went to press before the lecture, on the strength of an interview its technology policy editor, Joseph Hanlon, had conducted with Beer at home, and carried two pieces. The first, in the Technology Review pages at 363 to 364, was headed “Chile leaps into cybernetic future” and is the fullest contemporary description of the system in print. Its first sentence: “The dream of operations research — to monitor an entire country’s economy by a real time computer system — is actually happening in Chile. Data on one-quarter of the economy is fed daily into central computers, where it is analysed for significant deviations from normal and reports made to managers and government officials.”

Hanlon’s figures are the ones later writers have leaned on, and they are Beer’s. “Compared to traditional management information systems, Beer’s system is surprisingly simple. Total input is probably less than 5000 numbers per day. The system uses only two medium size computers — an IBM 360/50 and a Burroughs 3500 at Chile’s national computer centre. And it required relatively few people to develop — less than 100 scientists are working on the project for example.” Each firm had “only ten critical numbers which must be Telexed to the central computer each day”; each index had “its own model constructed based on 100 days of data”; each day the computer used “Bayesian statistics to do a forward projection” and picked out the changes of slope and the steps. The program, Beer told him, “is incredibly sensitive to these changes, recognising them long before the human brain would dare to make a judgement.” The manager was “not overwhelmed with information, typically getting only one report per week”, and a branch or sector manager never got “more than 7 numbers a day, on average.” The report described the four levels of models, the exception report that went only to the manager, the time limit that passed an unsolved problem upward “so that even a minor problem left unsolved long enough gets up to the top minister, Flores”, the DYNAMO simulations, and the operations room, which “Beer proudly declares ‘looks like a set for a futuristic film’.” It gave the timetable: “Beer began work on the system in November 1971. By March 1972 data was already flowing to the central computer. Cyberstride was in use by November 1972. The system works, according to Beer, and is beginning to be used.”

Then came Beer’s own list of what remained, which is the most useful paragraph in the piece and the least quoted. Five steps: debugging and model correction; expansion “to the entire economy by the end of Allende’s presidential term in 1976”; third, “because the system was imposed from the top in relative secrecy, there is little understanding of it at lower levels. So the government must get managements to use the exception reports and to add their own indices to the system”; fourth, “an entrenched and probably angry bureaucracy which has been bypassed by the system”; and fifth, “a feedback system for the peasants and workers, so far ignored by the system.” Beer closed on the power he knew he was handing over: “like guns, we are surrounded by things that can be perverted.” But “the worst abuses of freedom come from an absence of tools”, and “Cybernetics can do the job better than bureaucracy.”

The report also carried the strike as Beer told it. The system’s power “is best illustrated by the October 1972 ‘bosses strike’, when owners of transport companies nearly brought the country to a halt by withdrawing services. It has been reported that using data coming in on the special communications system, Chilean leaders had an accurate and instantaneous picture of the state of the country. Computer models were reportedly used as an aid to selecting the best strategy to break the strike. As Beer commented, ‘communication is indeed control’.” The reporting verbs are Hanlon’s; the claim about computer models is Beer’s, and the Chilean participants’ accounts, gathered by Medina, do not support it. What the network carried during the strike was messages.

Hanlon, the comment

The second piece was on the Comment page, at 347, under the heading “The technological power broker”, and it is where the word technocrat entered the story. It opened with Beer “sitting next to a goldfish pond in the centre of his spacious study in his large house in the Surrey stockbroker belt”, speaking “glowingly of the Marxist revolution in Chile” and explaining that “cybernetics is mappable onto Marxist-Leninist theory”; he had also said that “liberty must be a computable function.” If the system succeeded, Hanlon wrote, “Beer will have created one of the most powerful weapons in history.” The operations room “looks as though it belongs in the film 2001”. The system “reportedly was an important element in breaking the ‘bosses strike’ last October.”

Then the argument. Beer said the system devolved power, because a manager learned of a problem before his superior did and could act “without his superior even finding out.” But he had to act on what the computer thought was a problem, “because if he does not, his superiors will be told. And Beer admits that it would be a simple reprogramming job to take out this safeguard. It also seems likely that managers will soon act to make the computer respond correctly, independent of the actual factory.” Beer’s second devolution, the feedback for the peasants, was a two-way link for responding to television programmes. “Thus we have Beer’s view of Chile: a computer system monitoring all of the economy, designed by a small group of scientists and imposed from the top in secret. Managers have autonomy so long as they do what the computer expects. The peasants and workers will be permitted to respond to TV programmes.” Beer, Hanlon said, had seen the main distortion in government as delay and proposed a technical cure, “but he has missed the real distortion — when you make decisions for someone else, your own problems and values intrude.” The system gave the most power to Flores, “Minister of Finance, who has risen through four government posts in two years.” Beer’s answer, that he had only “provided tools” and was not a technocrat because he did not govern, Hanlon reported and did not accept: “Many people, however, will think Beer the super-technocrat of them all.”

Two things about the comment are worth noticing fifty years on. It took Beer’s claim that the system ran a quarter of industry at face value, as every paper did. And its prediction about the managers was not a prediction about Chile’s politics but about any instrument of this kind: that the people measured will learn to satisfy the measurement. Nothing in the record shows a Chilean manager doing so, because nothing in the record shows most of them reading the reports at all.

Grosch

On 15 March 1973 the New Scientist printed, under the heading “Chilean economic controls”, a letter from Herbert Grosch, the American computer scientist whose name is on Grosch’s law of mainframe economics and who had advised the previous Chilean government on computing in 1969. Medina quotes it. From what he had seen then, Grosch wrote, “it is absolutely not possible for Stafford Beer, Minister Flores or the Chilean government or industrial computer users to have since then implemented what is described.” A new model of that scale could not be commissioned in a few months in “a strange and primitive hardware and software environment.” “I call the whole concept beastly,” he concluded, and thought it as well for humanity and for Chile that it was so far only a nightmare. Grosch’s objection was the opposite of Hanlon’s: not that the thing was dangerous but that it did not exist. Measured against what had actually been built, a hundred telexes, thirty indices in July 1972, a filter running a day or two behind on a borrowed Burroughs, he was nearer the mark than he knew, and for the wrong reason: the hardware was adequate to the job that was done, and the job was smaller than the lecture.

Science for People

The British Society for Social Responsibility in Science, founded in 1969, counted Beer among its sympathisers, and he had talked openly with some of its members. Its magazine Science for People, number 21, carried an article by John Adams titled “Everything under control”, with a cover drawing of a fat Chilean manager in an armchair pressing a button on a mainframe to move the limbs of a robot worker. Adams read the algedonic signals as a device for keeping individual workers inside their “physiological limits”: anyone in the lower reaches of the system with different objectives who strayed beyond them, and so threatened “the well ordered production machinery”, would be “communicated with ‘algedonically’”, he wrote, taking the phrases from the Goodman lecture. The system, in this reading, destroyed “the inner autonomous man”.

Beer answered in a letter to the editor that Medina quotes: “Take an arbitrary selection of any man’s writings. . . . Note down any odd passages that look a bit weird on their own. Disingenuously string these gobbets on a thread of guilt. Could you not make a necklace to hang the man?” Jonathan Rosenhead, a founder of the society, told Medina that Beer must have felt it as “a stab in the back from people who he thought were his friends”, and that the society would have taken a different line had it known what was coming in September. “Then it would have been clear we were on the same side.”

What the ten weeks show

Between 7 January and 15 March 1973 the project was read in Britain as secret and total by a Sunday paper, as a weapon and a technocracy by a science weekly, as technically impossible by an American expert in the same weekly, and as an instrument of discipline by the scientific left. In Chile in the same weeks it was Big Brother in Ercilla and a secret plan in Qué Pasa. Every one of these readings was of the system as Beer described it in the lecture and the interview: a quarter of the economy, daily, filtered, displayed in a futuristic room. None was of the system as it ran, which was ninety-nine telex machines and a program whose output went to a telex room at CORFO. The lecture was the instrument the press read, and the lecture was a fanfare.

Sources

• Joseph Hanlon, “Chile leaps into cybernetic future”, Technology Review, and “The technological power broker”, Comment, New Scientist 57, no. 833, 15 February 1973, pp. 363 to 364 and 347. All quotations from both pieces. Read from the Google Books scan of the issue mirrored on archive.org. archive.org
• Eden Medina, Cybernetic Revolutionaries: Technology and Politics in Allende’s Chile, MIT Press, 2011, chapter 6, “Cybersyn Goes Public”, for the Observer’s sentences, the telexes about the draft lecture, Beer’s words on effective freedom, the telexes of 16 and 17 February, Grosch’s letter (New Scientist 57, no. 837, 15 March 1973, pp. 626 to 627), the Science for People article (no. 21, 1973, pp. 4 to 6), Beer’s reply and Rosenhead’s recollections. In copyright; quoted in short phrases; the Observer, Grosch and Adams are quoted at second hand from it. mitpress.mit.edu
• Eden Medina, “Designing Freedom, Regulating a Nation”, Journal of Latin American Studies 38, 2006, in the Spanish translation in Redes 20, no. 38, 2014, for the references to Adams and Hanlon and the Grosch quotation. ridaa.unq.edu.ar
• Stafford Beer, “Fanfare for Effective Freedom: Cybernetic Praxis in Government”, the Richard Goodman Memorial Lecture, Brighton Polytechnic, 14 February 1973, reprinted in Platform for Change, Wiley, 1975. Not read for this entry; quoted only as Hanlon and Medina report it.