(untitled)

Has anyone read Cybernetics by Norbert Wiener? I've gotten curious about the field of Cybernetics and am wonder what are some other books I could read that explore this idea. Norbert Wiener defined cybernetics in 1948 as "the scientific study of control and communication in the animal and the machine." Wikipedia states: "Concepts studied by cybernetician include, but are not limited to: learning, cognition, adaptation, social control, emergence, convergence, communication, efficiency, efficacy, and connectivity. In cybernetics these concepts (otherwise already objects of study in other disciplines such as biology and engineering) are abstracted from the context of the specific organism or device." From my interpretation there are clear ties to Bionics, Mathematical Biology, AI, and automata theory. I don't really know if this belongs here or not, just wondering what lain thinks.

https://www.hooktube.com/watch?v=As4gzzCQJ84

Marvin Minsky wrote a couple of books on approaching AI by starting at the processes of the human mind

I came upon it from the mathematics/physics approach, so I always tie cybernetics in with statistical mechanics, thermodynamics and control theory. A lot of this stuff got re-phrased as "complex systems theory" during the 70s when chaos became sexy. Stuart Kauffmann, Ilya Prigogine and Steven Strogatz are some of the big names in the field if you wanna check them out. (Strogatz wrote a great textbook on dynamical systems, and his whole lecture series on the subject is available on YT. Pre-reqs are basically just differential equations.) From memory, The Human Use of Human Beings is the more approachable of Wiener's work than Cybernetics, since he took down the dickish more-genius-than-thou tone a notch or two. It would probably be worth reading some of his contemporaries at the same time: Koestler's The Ghost in the Machine and von Neumann's The Computer and the Brain come to mind. Jay Forrester was another contemporary, and worked on similar feedback-loop stuff, but he took it in a totally different direction by applying it to business processes and economics. He was really big at the Sloan school for a while; pretty sure MIT Tech Review did a good write-up on him. Obviously this is just one perspective, though, and someone with a robotics or neuroscience background will have totally different impressions. Cybernetics is weird because it never became a discrete discipline, but diffused into a wide range of sciences. I'd be interested to hear others' opinions.

Yes, I came across Wiener's work in graduate school. Its no surprise that it seems to be connected to Bionics, Ai, Automata, etc; Cybernetics was an interdisciplinary precursor to a variety of fields focused on how information is exchanged between entities as a governance mechanism. I would look into systems theory etc etc to continue the study. Check out a great podcast -https://newbooksnetwork.com/category/science-technology/systems-and-cybernetics/

read tiqqun - the cybernetic hypothesis for a critique of the application of cybernetical concepts to politics and economy

https://www.youtube.com/watch?v=LQucESRF3Sg

Check this out. With modern computers (even your desktop) it is now completely feasible to calculate labor values and five year plans for the WHOLE economy (IE, millions of commodities, as opposed to only low thousands in current planned economies), as interior point methods allow for nlogn estimation of optimal industry outputs. You would only need a central database for everyone's inventory and productivity.

>>12541 Going to read it, thanks. Though I think the problem with planned economy is not only computational resources. You need to collect the data, decide on the objective function, make the process of calculation transparent and verifiable etc.

>>12552 >. You need to collect the data, True, this is a political challenge more than anything. Most of the data (manufacturing dependencies, labor time per piece, specialization of labor) already exists in private databases, you would have to seize the data along with the factories. It could take less than a month for most of it to be entered. >decide on the objective function If you mean the social need, then yeah. TANS proposes a voting scheme where planning teams put together competing five year plans, and voters decide which one they like most. For example, one plan might prioritize military, one might prioritize food and healthcare, one could prioritize consumer electronics, etc. And the planning algorithms should be free software and interchangeable, along with the database, making it easy to verify.

I think most people look at cybernetics and think of it as a precursor to neural networks and machine learning. But after reading pic related (and doing the exercises), I think cybernetics has a wealth of information that could be fruitful today now that we have more capable abstractions, better equipment, and a better understanding of the brain, the limitations of artificial neural networks, and the stupidity of GOFAI language-esque tree structures, i.e. "expert systems". We've created mechanisms that mimic functions of a human -- vision, hearing, classification, sequence generation, etc -- but we haven't put these together with any sort of intelligence. Cybernetics is entirely about finding that fullcrum at which emergent properties arise from a system of systems and we absolutely need to revisit the area.

>>12541 I believe the book mentions the efforts of Soviet cyberneticians such as Kitov and Glushkov to change how the Soviet economy was planned (correct me if I'm wrong), but I don't think Cockshott's approach has anything to do with cybernetics itself. An economic planning effort more connected to cybernetics was Cybersyn, Allende's program to democratically plan Chile's economy by means of a feedback network of telex machines. https://en.wikipedia.org/wiki/Project_Cybersyn It was overseen by the British cybernetician Stafford Beer, and based on an implementation of his Viable System Model. It (attempts to) formalize the dynamics of an organizational structure which can replicate itself. https://en.wikipedia.org/wiki/Viable_system_model VSM in Squeak: http://ss3.gemstone.com/ss/VSA.html in Ruby: https://github.com/rbcarleton/VSA A part of Beer's VSM is that when a subsystem cannot decide what action to take, a metasystem is generated which conjoins it to sources of more information so it can decide. This metasystem generation can go on infinitely. This sounds a lot like Turchin's work on metasystem transitions as part of supercompilation, wherein compiling a program consists of generating an abstracted "metasystem" which mimics the existing program and generalizing it to a new version without indirections (unnecessary list traversals, etc). >>11430 I took a gander at it. They have no idea what they're talking about. It's a word salad. >>12533 What is it about the word "cybernetics" which attracts crackpots left and right?

>>12563 >I don't think Cockshott's approach has anything to do with cybernetics itself. He calls it cybernetics and talks about Beer a lot.