Chapter 141 23.1 What happened to cybernetics?

"Good morning, Organization System!" The cheery speaker adjusted his tie gracefully and smiled: "On this carefully chosen day, the Office of Naval Research and the Armor Research Foundation are sponsoring this symposium to discuss what I personally consider to be very important subject, I am very happy." It was a day in mid-spring in May 1959. 400 scholars from a wide variety of disciplines gathered in Chicago for an event that promises to shock the scientific community.The guests at the meeting covered almost the world's major scientific branches: psychology, linguistics, engineering technology, embryology, physics, information theory, mathematics, astronomy and social sciences, etc.Never before has there been a conference that has brought together so many top scientists from different fields to spend two days on one topic.Certainly no large-scale conference has ever been held on this particular subject.

Only young and prosperous countries, when they are confident in their role in the world structure, will think about such a question: self-organizing system-how does the organization bootstrap itself.Bootstrap image storage!This is the American dream put into the equation. "The timing of the meeting was also of particular importance to my personal life," the speaker continued: "The Department of Defense has been going all-in on organizing for the past nine months, and it's clear that This clearly shows that we are still a long way from properly understanding what makes self-organizing systems possible."

Knowing laughter came from the crowd who entered the venue early in the morning.Dr. Joachim Weil, Director of Naval Research, who spoke on the podium, continued with a smile: "I would like to draw your attention to three basic elements that deserve to be studied. In the long run, we have a great understanding of the memory element in the computer field. Absolutely and inevitably it will be applied to the 'self-organizing system'. When it comes to computers, you may think of them as I do these days as nothing more than a tool, a way to help memory transfer from one state to another. a state tool."

"The second element is what biologists call differentiation. It is clear that any system capable of evolution cannot do without what geneticists call mutations that are essentially random events. Pushing a population in one direction , pushing the other group in the other direction, requires some initial trigger. In other words, for long-term natural selection to work, it must rely on a noisy environment to provide the trigger." "The third element, perhaps, manifests itself in the purest and most intelligible form when we deal with large social organizations. For the purposes of this conference, let me call it subordination, if you will , which can also be called executive function.”

Look at these terms: signal-to-noise, mutation, executive function, self-organization.When these words were uttered, the DNA model had not yet been established, digital technology had not yet been applied, the specialty of information management systems had not yet appeared, and the complexity theory had not yet been born.It's hard to imagine how outrageous and innovative these ideas were at the time. And how right it is. In the blink of an eye 35 years ago, Dr. Will outlined an entire book I published in 1994.In that book I discussed adaptability, the breakthrough science of distributed systems, and the emergent phenomena that this science led to.

As remarkable as the predictions at this meeting in 1959 were, I see another aspect worth mentioning: how little our understanding of the system as a whole has improved in thirty-five years.Despite the great recent achievements mentioned in this book, many fundamental questions about the self-control, variation and subordination of the whole system are still foggy. The all-star lineup of papers presented at the 1959 meeting brought together scientists who had been gathering in small meetings since 1942.These intimate, invitation-only gatherings were initiated and organized by the Macy Foundation and became known as the Macy Conferences.In the tense wartime atmosphere at that time, most of the participants were interdisciplinary academic elites, focusing on important organizational issues.In the past 9 years, dozens of artificial intelligence researchers invited by the conference include Gregory Bateson, Norbert Wiener, Margaret Mead, Lawrence Frank, John von Noy The star-studded gathering of Mann, Warren McCulloch, and Rosenbluth became known as the cybernetic community for its pioneering idea that cybernetics was the art and science of control.

Some things don't stand out initially; not this time.From the very first meeting, the participants of the Massey meeting could imagine the magnificent scenery behind the door of heresy they opened.Despite their strong scientific backgrounds and natural skeptics, they realized right away that this innovative perspective could change the rest of their academic careers.The anthropologist Margaret Mead later recalled that she was so excited by the ideas that came out of her first meeting that “I didn’t notice until the meeting was over that I had bitten a tooth out. " Members of this core group included leading thinkers in fields such as biology, social science, and what we now know as computer science, even though the group was only then beginning to create the concept of computers.Their most important achievement is to clearly describe the control and design language, so as to work in biology, social science and computing.The remarkable results of these conferences benefited from an alternative approach at the time: treating organisms strictly as machines and machines as living beings.Von Neumann quantitatively compared the computing speed of brain neurons and vacuum tubes, boldly suggesting that the two can be compared.Wiener reviews the evolution of automatically controlled machines into the human anatomy.Physician Rosenbluth predicted homeostatic circuits in the human body and cells.Steven Hymes, who recounts the story of this influential group of thinkers in Cybernetic Crowds, says of the Massey meeting: "Even anthropologists like Meade and Frank They have also become proponents of understanding things from a mechanical perspective. In this concept, they describe the living body as an entropy decay device, give the characteristics of human automatic control devices, regard the human mind as a computer, and use mathematical game theory Let’s look at social conflict.”

In an era when popular science fiction was just emerging, before it became the influential element of modern science today, the attendees at Macy's conferences often used wildly exaggerated metaphors, much like science fiction writers do today.At a conference, McCloch said something like this: "I don't particularly like humans, never have. In my opinion, humans are the most despicable and destructive of all animals. If humans can evolve to outlive humans I can't think of any reason why machines shouldn't be quite happy to replace us and enslave us. They might have a lot more fun and find more fun." Humanists are astonished to hear such speculation. But behind this nightmarish and dehumanizing plot lurks some very important ideas: that it is possible for machines to evolve, that they may indeed be able to do everyday tasks better than us, and that we share the same privileges as good machines. operating principle.These ideas are good metaphors for the next millennium.

As Meade wrote after the Massey meeting: "What the cybernetic community fails to take into account is the succession of new inventions of high order and success." In particular, feedback control, circular causality, machine Concepts such as dynamic equilibrium and political game theory have gradually entered the mainstream, and until today, they have become basic concepts that are almost overwhelming. The cybernetic community did not find the answers according to the timetable it arranged for solving problems.Decades from now, scientists studying chaos, complexity, artificial life, containment architecture, artificial evolution, simulation, ecosystems, and bioinspired machines will provide a framework for problems in cybernetics.Those who give a one-sided overview may say that this book is an up-to-date source on the state of cybernetics research.

But this book is also quite confusing.If it really deals with cybernetics, why is the term "cybernetics" so rare throughout the book?Where are the early pioneers of cutting-edge scientific research today?Why aren't the academic authorities of older generations and their brilliant ideas at the center of their naturally extended research work?What happened to cybernetics? This was an incomprehensible thing that bothered me when I first dealt with the younger generation of system developers.These more learned people certainly knew about the early cybernetic work, but hardly any of them had a cybernetic background.It was as if in the process of knowledge transmission, that whole generation disappeared and a gap appeared.

There are three speculations as to the reasons for the demise of the cybernetic movement: Cybernetics research was halted as funding dried up, as the then-hot but defunct field of artificial intelligence research was siphoned off massive amounts of funding.The failure of artificial intelligence is that utility was developed at the expense of cybernetics.Artificial intelligence is only one aspect of cybernetics research, but when it gets most of the government and university funding, the rest of cybernetics' vast remaining research topics disappear.Fresh graduates have entered the field of artificial intelligence research one after another, so there are few successors in other fields.After that, AI research itself ground to a halt. Cybernetics was a victim of the batch computing paradigm.Information transmission is the most important strategy of cybernetics.The kind of experiments that required testing his ideas required the computer to run at full speed many times in a fully examined mode.Such a requirement is obviously out of date with the strict laws protecting the mainframe.As a result, cybernetic theory has done very little experimentation on this.Then cheap personal computers became popular, but adoption in universities was notoriously slow.Even high school students took Apple IIs home, and colleges still used punched cards.Chris Langton made the first artificial life experiment in his life on the Apple computer.Don Farmer and friends discovered chaos theory by building computers.Real-time control of a complete general-purpose computer is something that traditional cybernetics needs but never does. The phrase "put the observer in the box" killed cybernetics. In 1960, Forrester wisely proposed that the system observer can be added as a component to a larger meta-system to obtain innovative views on social systems.He set up a framework for his observations called secondary command control, or a system called the observation system.This insight is pertinent in areas such as family psychotherapy, where the therapist theoretically has to integrate himself into the family for therapeutic effects.But "putting the observer in the box" runs into infinity when the therapist videotapes the patient, then the sociologist shows the therapist a videotape of the patient's video, and then watches the therapist video for himself... return.By the 1980s, the American social cybernetic roster was full of clinical therapists, sociologists, and political scientists whose primary interest was to observe the effects of the system. These three reasons acted in concert, so that by the end of the 1970s, cybernetics withered and died.The vast majority of cybernetic research stays at the level described in this book: unrealistically stitching together a grand picture.The real researchers either suffered setbacks in artificial intelligence labs, or continued to work in remote scientific institutions in Russia, where cybernetics research as a branch of mathematics did continue.In my opinion, there is not a single official cybernetics textbook written in English.