Cybernetics Chat - Beer and Weiner ITT - The Something Awful Forums

The Something Awful Forums > Discussion > C-SPAM > Cybernetics Chat - Beer and Weiner ITT

turd in my singlet: Jul 5, 2008; DO ALL DA WORK

WIT YA NECK

*heavy metal music playing*; Nap Ghost

quote:

Cybernetics is a transdisciplinary[1] approach for exploring regulatory and purposive systems�their structures, constraints, and possibilities. The core concept of the discipline is circular causality or feedback�that is, where the outcomes of actions are taken as inputs for further action. Cybernetics is concerned with such processes however they are embodied,[2] including in environmental, technological, biological, cognitive, and social systems, and in the context of practical activities such as designing, learning, managing, and conversation.

https://en.wikipedia.org/wiki/Cybernetics

Cybernetic Marxism is one of the more obscure but more interesting (imo) threads of modern Marxist theory. It combines the breakthroughs in the understanding and control of complex systems from the mid-20th century with the analysis of the capitalist system and revolutionary project of Marxism. It offers the potential to answer questions about how to design effective organizations, governments, and economic systems; and new ways to analyze the successes and failures of past and existing communist and capitalist societies.

TBH I'm hardly the person to write a good OP since I've only been following the idea for a little over 6 months lol, but people in other threads have shown interest in having a dedicated thread for it so I figured I'd give it a go anyways.

The main person associated with cybernetic Marxism is Stafford Beer, who a lot of people on the left are at least aware of because of his involvment in the design of a computerized central economy for Salvador Allede's government in Chile, "Project Cybersyn". IDK too much about the specifics of Cybersyn, but broadly it was an attempt to apply Beer's Viable System Model to a socialist country's economic system. Unfortunately Allende's government was couped before much of the system had been implemented so it's hard to say how well it would have worked. The book Cybernetic Revolutionaries by Eden Medina goes into the story in detail.

The Viable System Model (VSM) is Beer's attempt at a general model of what information and control flows are required for a complex system to be able to intelligently respond to changes in it's environment and remain a coherent entity. Beer was specifically a management cyberneticist; his professional work was as a business consultant who analyzed organizational structure in order to reduce conflict and increase efficiency. He claims the VSM based on natural systems in general but I'm not sure where he's coming from exactly- I've only scratched the surface of his many books though. In his book Brain of the Firm, he explicitly teaches the model by way of comparison to the human nervous sytem, with the information and control flows of the peripheral, autonomic, and central nervous system demonstrated to be analogous to aspects of how a business needs to run in order to be effective and coherent.

A resource on Beer CGI Stardust shared with the thread:

CGI Stardust posted:

couple of introductory resources for the OP if you're interested

Stafford Beer - Designing Freedom radio lectures: introductory; Beer talking a bunch about cybernetics and his social application of it - attempting to maximise autonomy while preserving cohesion. 6 lectures, about 30 minutes each. There's also a book, limited availability at https://archive.org/details/designingfreedom00beer, or pick up from wherever

Stafford Beer - The Falcondale Collection (also on a YouTube playlist found on a Discord): an informal introductory "lecture" series from the early 1990s; Beer explaining his approach to cybernetics in a small group setting over a weekend, apparently drinking nothing but wine. 9 videos, 1h to 1h 30m each. wholesome stuff!

CGI Stardust posted:

Beer's social attitudes are weirdly conservative in a few places. you're going along nicely at 60 then suddenly hit a speedbump out of nowhere. change that description to "mostly wholesome, occasionally not good". product of his time and class i guess, his roots were management consulting in the 1950s and 60s

like in Brain of the Firm, when talking about Cybersyn, there's an explanation of why they didn't want to use keyboards for the control room because keyboards were universally seen as a women's thing that only typists would use; instead they designed input around big spring loaded switches that could be thumped by passionate workers (men, specifically) to express a point; Heart of Enterprise is also somewhat sexist

it's possible to extract the useful cybernetics stuff without endorsing everything about him, i hope

Another cyberneticist is Norbert Weiner, who TBH I don't know too much about. I do know that while writing the book The Human Use of Human Beings, he was visted by the FBI who made it clear he needed to change some un-American aspects of his book before he published it. A goon I know who's into cybernetics suggested I put him in here and it made for a good thread title lol.

Relevant resources:

Zodium posted this cybernetics 101 article, which takes a historical approach to introducing cybernetic concepts.

This post by Bar Ran Dun introducing some basic cybernetics and tying it to the current prevalence of Just In Time manufacturing.

Finicums Wake posted:

another particularly interesting resource:
a paper by alicia juarerro about complex systems and philosophy. screencap of the first page here, but you can find on sci-hub:
https://mobile.twitter.com/Tarp_Ghost/status/1391852039043493888

(spoiler'd to keep the auto-embed from making this post even longer lol)

Podcasts

General Intellect Unit - the self-described 'Podcast of the Cybernetic Marxists'. This is the only pod I know of focused specifically on cybernetic Marxism. They did a reading group of Beer's Brain of the Firm last year and have been slowly publishing them as podcast episodes.
Cosmopod - the podcast put out by Cosmonaut magazine. Not specifically cybernetics focused but they do have some content about it. They did an episode last summer about the book Cybernetic Revolutionaries, and another more recently about how some of their contributors have been trying to apply cybernetic ideas in their organizing.

Books

Stafford Beer wrote something like ten books from the late 1950s to the early 1990s. I'm not super familiar with his work; I've just been reading Brain of the Firm along with the GIU reading group episodes as they come out. Brain is kind of the main one about management, and includes a section in the end of the second edition about his experience with Project Cybersyn.

The People�s Republic of Walmart: How the World�s Biggest Corporations are Laying the Foundation for Socialism - Leigh Phillips and Michal Rozworski. Haven't read it yet but the premise is that the massive computerized internal economies of companies like Wal-Mart and Amazon demonstrate that the technology needed for a central computerized economy already exists.

Cybernetic Revolutionaries - Eden Medina. History of Project Cybersyn.

Towards a New Socialism - Paul Cockshott. IDK much about this one besides that it's relevant.

CGI Stardust posted:

Thomas Swann has a book on the application of the VSM to anarchist organising, Anarchist Cybernetics - Control and Communication in Radical Politics which seems solid from what i've read so far (please for the love of god ignore the cover, no idea why the publisher thought that was a good idea). there's also this paper of his that summarises some of the stuff from the book Towards an anarchist cybernetics: Stafford Beer, self-organisation and radical social movements, and if you dig around his PhD thesis is basically the book but not edited and also free

Finicums Wake posted:

edit: one more resource. here's a book on cybernetics mirowski is citing a lot that looks good.

https://www.goodreads.com/book/show/7636063-the-cybernetic-brain

Communities

Closest thing I'm aware of is r/socialistprogrammers. Otherwise its private discords.

Hopefully posters ITT will be able to contribute some other resources.

Trabisnikof posted:

im going to post about systems thinking and learning organizations itt and no one will stop me

turd in my singlet has issued a correction as of 17:16 on May 13, 2021

# ? May 5, 2021 23:52

Adbot: ADBOT LOVES YOU

# ? May 9, 2024 10:41

Finicums Wake: Mar 13, 2017; Can't post for 8 years!

ross ashby's book on cybernetics is supposed to be very good, though i haven't dug into it much myself.

link:
http://pespmc1.vub.ac.be/ASHBBOOK.html

a fantastic book on cybernetics and its influence on mainstream neoclassical/marginalist economics is philip mirowski's machine dreams: economics becomes a cyborg science. read mirowski in general, imo, but that's the main one about cybernetics

i will try and find my notes/annotated copy of the mirowski book and make an effortpost about it if i can do so. i'm supposed to be taking a philosophy seminar that deals with cybernetics this summer, so reviewing that material would really help me anyways

Finicums Wake has issued a correction as of 04:35 on May 6, 2021

# ? May 6, 2021 03:33

Idia: Apr 26, 2010; Fun Shoe

I've read The Cybernetic Brain as a jumping off point since it's just a "history" of mostly British cybernetics and how it also impacted art and music at the time. I've also read the Brain of the Firm and Cybernetic Revolutionaries. I still can't grasp it, so I can't really contribute.

Will this thread also discuss Soviet cybernetics?

# ? May 6, 2021 04:28

Finicums Wake: Mar 13, 2017; Can't post for 8 years!

i don't know anything about the soviet situation/parallels but if any does pls share

# ? May 6, 2021 04:34

CGI Stardust: Nov 7, 2010; Brexit is but a door,
election time is but a window.

I'll be back

aw hell yes this is my kind of thread

for any anarchisty types, although i'm not really familiar with anarchism, Thomas Swann has a book on the application of the VSM to anarchist organising, Anarchist Cybernetics - Control and Communication in Radical Politics which seems solid from what i've read so far (please for the love of god ignore the cover, no idea why the publisher thought that was a good idea). there's also this paper of his that summarises some of the stuff from the book Towards an anarchist cybernetics: Stafford Beer, self-organisation and radical social movements, and if you dig around his PhD thesis is basically the book but not edited and also free

# ? May 6, 2021 05:50

some plague rats: Jun 5, 2012; by Fluffdaddy

Cant decide between mantis blades and the gorilla arms personally

# ? May 6, 2021 06:04

exmarx: Feb 18, 2012; The experience over the years
of nothing getting better
only worse.

systems stuff breaks my brain when numbers get involved, but i think it's helped me to understand dialectics better... i love this example:

quote:

But what happens if cause flows in both directions? What happens if health outcomes of policies result in public action to change policy, if disability affects income? In the last century, Engels wrote of the interchanging of cause and effect, physiologists described self-regulation, and engineers were designing self-correcting industrial processes. In systems of any complexity there are feedbacks, and these affect the relationship between statistical outcomes and causal pathways.

In negative feedback, a change in one element of a system leads to changes in others that eventually negate the original change. The negation may be partial, complete, or even overshoot, so that dumping nitrogen in a pond may reduce the nitrogen level if a radical change in species composition occurs, or applying pesticides may increase pest load by removing more pesticide sensitive competitors of the pest or, frequently, by killing off predators of the pest species. The predators are poisoned directly by the pesticide, but both a negative and positive branch of a feedback loop are involved. Along the positive branch predators are decreased because their food supply, the pest species, is decreased by the pesticide. Along the negative loop the pest carrying insecticide molecules poisons the predator, which results in an increase in prey. It is not that predators are more sensitive physiologically to insecticide, but that their location in the loop makes them more vulnerable ecologically.

(Biology Under the Influence: Dialectical Essays on the Coevolution of Nature and Society)

# ? May 6, 2021 08:48

Zodium: Jun 19, 2004

thank u for posting the op, op. cybernetics and its adjacent fields are huge topics and centering discussion around marxism is a good idea. unfortunately i know more about cybernetics than i know about marxism, so i'd like to try and cover some of the philosophical ground. I bumbled into cybernetics from statistics around ten years ago via complexity theory and i've been fascinated ever since, and one thing I've learned is it's more of an entire perspective than just a field--predicated on completely different thought processes and ideas ~~kind of like a classical economics:marxist economics::classical science:cybernetic science relation~~. so it's useful to learn concepts and models, but it's also necessary to learn some history.

i've recently found this cybernetics 101 that I think gives the quickest possible introduction to the basic concepts with a little bit of history for someone who knows nothing about cybernetics that i'm just gonna go ahead and screencap liberally here. not a letter wasted.

imo the summary of second-order cybernetics as "a shift from the earlier preoccupation with disembodied models to constructions that recognize the observer as a participant in them" is particularly good, amd there's also a longer explanation on second-order cybernetics i think is worth reading too. it's really too long to screencap though.

historically, cybernetics began with the Macy Conferences in the 1940s. Pias' Cybernetics: The Macy Conferences 1946-1953. The Complete Transactions came up in a thread recently, and it is a very good and comprehensive resource on the macy conferences, if quite expensive. sadly doesn't include much on the largely undocumented early conferences. alternatives include this resource which is unfortunately in german, and otherwise there's this adorably crank aesthetic summary i guess. but some historical resource is necessary and thus I would strongly recommend obtaining a copy of the pias book, in some way.

it's an interesting fact you can start from almost any contemporary academic field and draw a line from contemporary thought back to a major figure who attended these conferences. see how many you recognize from the first one:

quote:

* William Ross Ashby; psychiatrist and a pioneer in cybernetics
* Gregory Bateson; anthropologist, social scientist, linguist, visual anthropologist, semiotician and cyberneticist
* Julian Bigelow; pioneering computer engineer
* Heinz von Foerster; biophysicist, scientist combining physics and philosophy and architect of cybernetics
* Lawrence K. Frank; social scientist
* Ralph W. Gerard; neurophysiologist and behavioral scientist known for his work on the nervous system, nerve metabolism, psychopharmacology, and biological basis of schizophrenia
* Molly Harrower; pioneering clinical psychologist
* Lawrence Kubie; psychiatrist
* Paul Lazarsfeld; sociologist and founder of Columbia University�s Bureau for Applied Social Research
* Kurt Lewin; psychologist, often regarded as the founder of social psychology
* Warren McCulloch (chair); psychiatrist, neurophysiologist and cybernetician
* Margaret Mead; cultural anthropologist
* John von Neumann; one of the foremost mathematicians of the 20th century
* Walter Pitts; logician and co-author of the paper that founded neural networks
* Arturo Rosenblueth; researcher, physician, physiologist and a pioneer of cybernetics
* Leonard J. Savage; mathematician and statistician
* Norbert Wiener; mathematician and founder of cybernetics

Zodium has issued a correction as of 18:05 on May 6, 2021

# ? May 6, 2021 11:02

Zodium: Jun 19, 2004

also gonna repost as an applied example this little model I made in the epstein thread as an argument against superintelligences, when a simple cybernetic system and the profit motive suffices:

we can imagine a hypothetical three letter agency called the Communism Control Program, whose mission it is to "control" communism. it's just a dumb monitor, because that's all it has to be. it monitors the rate of communisms emerging, and whether that rate falls within some goldilocks interval. our ccp doesn't try to predict or force anything to happen. it doesn't need to. it also only has two outputs: 'boot' and 'fund'. whenever our ccp detects there's too much communism, it starts putting the boot down on human faces--randomly--until the value comes back into the goldilocks interval. if there's not enough communism, the ccp funds some podcasts or movements--also randomly--until the value comes back up. the more off the value gets, the harder it boots or funds. that's it.

by holding the value in this goldilocks interval, the larger system the ccp exists within, Capital, receives continuous immunization against new evolutions of threats to its stability from communism. the ccp in turn receives feedback from capital about how its handling of communism is affecting system stability and move the goldilocks interval for communism accordingly. we ensure that communism neither goes critical nor goes away. even though this ccp itself is basically a more complex thermostat with only very rudimentary intelligence, by "controlling" communism in this way, it nonetheless works to ensure system stability is never seriously threatened by communism, all while appearing to a rational outside actor looking at any set of events as if it's randomly funding and oppressing stuff.

in the real world, of course, gradually since the late 40s, we've grown a rich and complex ecosystem of thermostats filling different control niches at every conceivable scale, each adapted to control a particular threat, all coordinated around the organizing principle of maximum profit.

# ? May 6, 2021 11:10

Good soup!: Nov 2, 2010

my weiner is augmented

# ? May 6, 2021 11:12

Larry Parrish: Jul 9, 2012; by Jeffrey of YOSPOS

oh my god that economics/marxism comparison makes me want to scream lol. it's like saying being a physicist is very similar to being a fantasy author. they're predicated on completely different thought processes and ideas, and only one is even a coherent attempt to view the world. classical economics is where grant money goes to die so beltway types can prove the economic necessity of feeding the poor into woodchippers

# ? May 6, 2021 12:01

animist: Aug 28, 2018

cybernetics is cool and statistical mechanics is cool

strong rec on "the human use of human beings", it's really well written and extremely prescient.

definitely one of those "oh god I hosed up" books on the part of mr weiner, you can tell he was extremely regretting telling the us intelligence apparatus about cybernetic principles in the 50s

# ? May 6, 2021 12:23

Zodium: Jun 19, 2004

Larry Parrish posted:

oh my god that economics/marxism comparison makes me want to scream lol. it's like saying being a physicist is very similar to being a fantasy author. they're predicated on completely different thought processes and ideas, and only one is even a coherent attempt to view the world. classical economics is where grant money goes to die so beltway types can prove the economic necessity of feeding the poor into woodchippers

thank you larry, that's a much better way to express the point I was going for. lol

# ? May 6, 2021 12:48

mycomancy: Oct 16, 2016

Zodium posted:

also gonna repost as an applied example this little model I made in the epstein thread as an argument against superintelligences, when a simple cybernetic system and the profit motive suffices:

we can imagine a hypothetical three letter agency called the Communism Control Program, whose mission it is to "control" communism. it's just a dumb monitor, because that's all it has to be. it monitors the rate of communisms emerging, and whether that rate falls within some goldilocks interval. our ccp doesn't try to predict or force anything to happen. it doesn't need to. it also only has two outputs: 'boot' and 'fund'. whenever our ccp detects there's too much communism, it starts putting the boot down on human faces--randomly--until the value comes back into the goldilocks interval. if there's not enough communism, the ccp funds some podcasts or movements--also randomly--until the value comes back up. the more off the value gets, the harder it boots or funds. that's it.

by holding the value in this goldilocks interval, the larger system the ccp exists within, Capital, receives continuous immunization against new evolutions of threats to its stability from communism. the ccp in turn receives feedback from capital about how its handling of communism is affecting system stability and move the goldilocks interval for communism accordingly. we ensure that communism neither goes critical nor goes away. even though this ccp itself is basically a more complex thermostat with only very rudimentary intelligence, by "controlling" communism in this way, it nonetheless works to ensure system stability is never seriously threatened by communism, all while appearing to a rational outside actor looking at any set of events as if it's randomly funding and oppressing stuff.

in the real world, of course, gradually since the late 40s, we've grown a rich and complex ecosystem of thermostats filling different control niches at every conceivable scale, each adapted to control a particular threat, all coordinated around the organizing principle of maximum profit.

So I read your OP in the Epstein thread, and this example of a "communo-stat" is a great thought experiment to explain how complex systems work. I don't think it precludes the existence of superintelligences evolving in human society, however. The neurological analogy for this type of simple decision making loop would be a voltage dependent ion channel, which is the basis for neuronal impulse generation. There's not that many of these encoded in animal genomes, but they can and do combine to form massive neural nets.

So what arises when we take thousands or millions of these little control loops, interlink them together, run them on silicon substrates (rather than biological or paper), and apply selective pressure? If human intelligence arose as an evolutionary response to dealing with the complexities of social conditions, material survival, and technology use, why couldn't other complex systems?

Now note I'm not saying that there ARE superintelligences creeping around every corner of modern society ready to gobble our wallets and our souls. What I'm saying is that if they DO exist, they're gonna be really goddamn weird to us, so alien in fact we may not recognize what they are at first. Humans have a big blind spot when dealing with non-vertebrate biological life; for instance, there's more genetic diversity in the Saccharomyces yeast genus than there is in the entire Family Mammalia, but we organize this genus into just eight species. I don't think a form of artificial life made from self-reinforcing dynamic control systems and run on a technological substrate isn't going to appear to your average human as being alive; hell, I blow students' minds when they realize that bacteria are alive!

I find these possibilities fascinating and horrifying. Are each of us the equivalent of a fecal bacterial cell, providing useful microconversions of food into vitamin analogs? Or do we act more like microglia, support systems for the techno-neurons that comprise this artificial life? Can we communicate more directly with this life form? Is it able to die in any way other than total human extinction?

# ? May 6, 2021 12:51

Rutibex: Sep 9, 2001; by Fluffdaddy

Larry Parrish posted:

oh my god that economics/marxism comparison makes me want to scream lol. it's like saying being a physicist is very similar to being a fantasy author. they're predicated on completely different thought processes and ideas, and only one is even a coherent attempt to view the world. classical economics is where grant money goes to die so beltway types can prove the economic necessity of feeding the poor into woodchippers

psychohistory is real and i believe in hari seldon :colbert:

# ? May 6, 2021 15:29

Bar Ran Dun: Jan 22, 2006; Can't post for 8 hours!

I wrote this for the D&D trade thread a couple of years ago.

Bar Ran Dun posted:

Systems, Business, Trade, Kanban, and Political Economy

This a bit rambling and it's not polished and it's probably not coherant yet. But here it goes.

There is a very useful concept from the Foundations of Cybernetics I�d like to start with. For every technology, be it a physical technology or a conceptual one there is a corresponding set of ideas and concepts, an (or almost) ideology, that allows one to use the technology in question most efficiently. Basically when one has a tool one also has to have a suite of concepts to use that tool most effectively. With that in mind I�m going go all the way back to the beginning. I think one needs to look at the technologies and historical events that gave rise to systems thinking to really understand it. This is a very, very simplified representation of a steam cycle.

Steam cycle

A boiler, boils and then superheats steam, it runs through the turbines which turn a generator, the steam is condensed and then as water it is pumped back into the boiler. Look at the diagram. Obviously the system and the drawing is informed heavily by thermodynamics. But I�m trying to communicate the following characteristics: The circular nature of the cycle, the flows of steam and water represented by lines, and the stocks of various liquids inside each of the parts. The lines showing flows between individual components. This diagram omits controls, but I�ll get to that in a sec. Anyways stocks are how much poo poo (water in a tank, steam in a boiler, etc) X things. Flows are how fast poo poo is moving X things per second. Flows can indicate what will happen in a system, but you need stocks to know when it will happen. Now one of the ways to control systems, is valves. Valves can do things like open and close. Or they can be throttled. Or a line can take pressure feed back from a pipe against a spring tension (or theses days a controller) and keep a flow at a particular rate (a analog solution). These are controls. Controls let one maintain the output one wants across variable inputs. One uses controls to keep a system from blowing up basically.

I�m going to play fast and loose with the history here but eventually WWII happens. During the war there is a lot of development in maths and we get tools like linear programming, that let's one maximize or minimize for desired values. There also is a problem after the war. How the hell does one get a rocket to fly where one want�s it to? Control theory solves this problem. They take concepts, particularly the math for how that pressure control valve works and they apply it to rocketry. Controls theory is how they solve the problem of getting rockets to fly straight. See those simple valves, can be expressed as terms in differential equations. That same math can describe rockets, electronics, etc. All of these things (and some other concepts from other disciplines like ships stability) eventually come together and form a discipline called Operations research. The controls / stock and flow modeling part sometimes gets called the �Differential Equations Paradigm� and eventually we can digitize the industrial controllers with it. Eventually we turn this poo poo on everything. In business operations research gets used as �management science�. In the sixties Kennedy�s eggheads, the whiz kids, this is the way they�re thinking , the tools they are using. Rand Corp pioneers a lot of this type of modeling. Eventually it becomes one of the standard ways we look at business. Managers use linear programming to maximize production and profits. Businesses are modeled using ideas from the circular stock and flow . That steam cycle, the suite of concepts used to describe that, can also be used to describe a business or business cycles. In fact when executives talk about creating value designing systems is what they mean. And I don�t mean I�m inferring this. I mean I�ve asked, and they give this answer. They are creating circular systems that take in inputs and spit out money. Some of these models get pretty sophisticated look at things like SCOR as good example. And I would remind you this type of supply chain modeling is what makes apple, one of the most valuable companies in the world. It�s what Tim Cook is good at.

Something vitally important to understand. A lot of these models are constructed and then spread. They then get applied and even taught by people who couldn�t have made them and don�t really understand them.
Anyway systems thinking gets used to model economics and trade, the economist Wille posted here: https://forums.somethingawful.com/showthread.php?threadid=3862896&userid=0&perpage=40&pagenumber=11#post488017309
Is a great example amd what prompted this post. And important because it shows how compatible systems thinking is with dialectic though, because of the circular nature of the model. A number of times he talks about using stocks and flows to model trade and currency flows after the switch to fiat currency. He talks at length about using stock and flow models (and I think he is applying them correctly and reaching correct conclusions with them). Think tanks, consultancies (McKinsey is a good current example) they all use models built on these ideas. I�ve done some simple modeling in this area for grad school.

Anyway eventually somebody (the Japanese, but I don�t know enough about the specific people to tell which ones, Toyota is a big player) takes this thinking and comes up with the idea of minimizing inventory (a stock) by balancing logistical flows. One can really supercharge a business� return on investment by bringing it�s inventory holding costs down as low as possible. This idea is called Kanban. Now the principle here is basically just that pressure control valve I told you about earlier. Flows are controlled to reduce the need for inventory. This idea becomes wide spread. Process management (which is more than Kanban and proceeds it) becomes widespread and largely in conjunction with growing globalization. Something to have in mind. You and I are if we are employees are stocks to be minimized to maximize cash flow and return to shareholders.

I wrote all that to say this. Now we have business systems that are very sophisticated, they use models like SCOR, to minimize their inventories (and again it�s worth noting inventory can be considered include employees) and make more money . But systems with very low stocks tend to be less stable. In physical systems, like the steam cycle, poo poo blows the gently caress up when things go wrong. Now in business where we have managers applying models they didn�t construct and are merely applying (hmmm now that�s familiar) and those models tend reduce stocks, they�re reducing stability in the system.

When it gets too hard to steer around the rocks, sometimes we hit the rocks. And this is the underlying thing that�s been worrying me about this escalating trade war.

# ? May 6, 2021 15:48

Toph Bei Fong: Feb 29, 2008

https://www.youtube.com/watch?v=gwf5mAlI7Ug

# ? May 6, 2021 15:50

Bar Ran Dun: Jan 22, 2006; Can't post for 8 hours!

some stuff that came up in the climate thread:

https://forums.somethingawful.com/showthread.php?threadid=3826121&perpage=40&noseen=1#post502925774

# ? May 6, 2021 15:51

Mr. Lobe: Feb 23, 2007; ... Dry bones...

Stick with the prod, prod with the prod

# ? May 6, 2021 16:03

wynott dunn: Aug 9, 2006; What is to be done?

Who or what can challenge, and stand a chance at beating, the corporate juggernauts dominating the world?

I work on production control systems and this thread is definitely my jam

you always prod with the prod

# ? May 6, 2021 16:36

Centrist Committee: Aug 6, 2019

I have alexandar bogdanov�s tektology on my list of poo poo to learn about someday. I believe it�s a precursor to cybernetics from the early russian revolution

# ? May 6, 2021 17:45

GEMorris: Aug 28, 2002; Glory To the Order!

I design software systems (from the business-problem+context to user-experience and interface side) and this thread is extremely my jam. Human Use of Human Beings is on my desk (unread sadly) along with Thinking in Systems (Meadows) and The Whale and The Reactor.

# ? May 6, 2021 17:49

Pf. Hikikomoriarty: Feb 15, 2003; RO YNSHO; Slippery Tilde

Zodium posted:

imo the summary of second-order cybernetics as "a shift from the earlier preoccupation with disembodied models to constructions that recognize the observer as a participant in them" is particularly good, amd there's also a longer explanation on second-order cybernetics i think is worth reading too. it's really too long to screencap though.

one thing i am hoping to see at some point is a cybernetics based in part on homotopy type theory.

this is the new hotness in the foundations of mathematics, the philosophers are interested in it too. the great thing about it is that every proof (ie controller) is treated by the formalism in the same way as the statement being proved (ie the process being controlled)

so a claim that this input equals that output would be automatically considered as a component of a system. moreover there is both an internal logic from the perspective of the system, and an external logic for reasoning from a perspective outside the system

# ? May 6, 2021 18:09

animist: Aug 28, 2018

mycomancy posted:

So what arises when we take thousands or millions of these little control loops, interlink them together, run them on silicon substrates (rather than biological or paper), and apply selective pressure? If human intelligence arose as an evolutionary response to dealing with the complexities of social conditions, material survival, and technology use, why couldn't other complex systems?

this is how deep neural networks work yes. that's the "AI" subfield that exploded this decade -- we're using them for everything now. Upscaling images, interpolating videos, steering self driving cars, synthesizing speech, targeting advertisements, deciding whether people should get loans...

they work based on the "stochastic gradient descent" algorithm rather than genetic algorithms. But it's a similar idea: make small randomized changes, which overall tend to follow some control signal.

(Also, don't be fooled by the name, they don't have much to do with biological neural networks -- those were the original inspiration for some aspects of the algorithm, but it's a loose inspiration at best. Better to think of them as big excel spreadsheets with randomized formulas inside.)

btw, have you heard of "adversarial examples"? That's where you can like, put a QR code sticker on a stop sign, and suddenly all the deep-learning-based self driving cars think it's a green light.

https://www.youtube.com/watch?v=i1sp4X57TL4

Some people think that this is just a glitch in deep neural networks, some edge case caused by the training process. But the weird thing is that the "adversarial examples" transfer. That is, they can fool multiple different neural networks, from different manufacturers.

Why is that surprising? Well, it comes down to how they're made. Each adversarial example is custom made for a particular neural network. So you'd think they wouldn't work on other networks... But they do, quite well.

some researchers think that this is because they're modifying "non robust features"... That is, they are changing aspects of the input images that deep neural networks can see, but that humans can't. This is currently only a hypothesis but imo it could hold water. See e.g. https://distill.pub/2019/advex-bugs-discussion/ for some people going back and forth about it.

so yeah, we're creating bizarre alien pseudointelligences whose internals are almost entirely opaque to us, which base their decisions on things we may not even be able to perceive... And we're using them to target advertisements and drive cars.

animist has issued a correction as of 18:23 on May 6, 2021

# ? May 6, 2021 18:12

Zodium: Jun 19, 2004

mycomancy posted:

So I read your OP in the Epstein thread, and this example of a "communo-stat" is a great thought experiment to explain how complex systems work. I don't think it precludes the existence of superintelligences evolving in human society, however. The neurological analogy for this type of simple decision making loop would be a voltage dependent ion channel, which is the basis for neuronal impulse generation. There's not that many of these encoded in animal genomes, but they can and do combine to form massive neural nets.

So what arises when we take thousands or millions of these little control loops, interlink them together, run them on silicon substrates (rather than biological or paper), and apply selective pressure? If human intelligence arose as an evolutionary response to dealing with the complexities of social conditions, material survival, and technology use, why couldn't other complex systems?

Now note I'm not saying that there ARE superintelligences creeping around every corner of modern society ready to gobble our wallets and our souls. What I'm saying is that if they DO exist, they're gonna be really goddamn weird to us, so alien in fact we may not recognize what they are at first. Humans have a big blind spot when dealing with non-vertebrate biological life; for instance, there's more genetic diversity in the Saccharomyces yeast genus than there is in the entire Family Mammalia, but we organize this genus into just eight species. I don't think a form of artificial life made from self-reinforcing dynamic control systems and run on a technological substrate isn't going to appear to your average human as being alive; hell, I blow students' minds when they realize that bacteria are alive!

I find these possibilities fascinating and horrifying. Are each of us the equivalent of a fecal bacterial cell, providing useful microconversions of food into vitamin analogs? Or do we act more like microglia, support systems for the techno-neurons that comprise this artificial life? Can we communicate more directly with this life form? Is it able to die in any way other than total human extinction?

thank you.

that's true, though the thought experiment isn't meant to preclude superintelligences so much as show a simpler explanation will suffice.

cybernetics crew actually just commandeered the doomsday economics thread starting here for a long discussion about Nick Land's deranged ideas about Capital as superintelligence, and I think the whole discussion is worth a read, but splifyphys really hit the nail on the head for me with a big marxist hammer:

splifyphus posted:

it's not so much whether or not we control it - obviously we don't, but that's not the point. he hypostatizes capital as something totally separate from concrete human practices when he envisions 'it' consuming us all and outlasting us, not to mention ascribing intention and sapience and whatnot. what kind of lovely god could be defeated by a majority of humans deciding to do something else all at once? it's a flawed conception from the getgo, and then he dramatizes it so he can throw in lovecraft references and deleuze and cheerlead for the technocalypse.

his value is in the naked honesty about what he's doing, but it's still a shame. some of his early work from before his big Cthulhuesque meltdown is actually quite incisive and brilliant.

splifyphus posted:

none of this matters cuz nrx is just liberalism taken to its logical conclusions. liberals are doing nrx already, they just aren't conscious of it.

which is all to say that if capital is a lovecraftian demon bootstrapping itself into existence from the future, this is the exact timeline we'd be living in anyways. the only real difference between communist and fascists on this point is that we, the former, insist that actually capital is just people in a historically specific matrix of relations doin thangs, and capital can't outlast the people doing it. the future remains the same.

# ? May 6, 2021 18:19

Centrist Committee: Aug 6, 2019

animist posted:

so yeah, we're creating bizarre alien pseudointelligences whose internals are almost entirely opaque to us, which base their decisions on things we may not even be able to perceive... And we're using them to target advertisements and drive cars.

this isn�t my field at all but I�ve definitely gotten high and lamented the enslavement of algorithms. are there any (accessible) writers who treat this idea more seriously?

# ? May 6, 2021 18:19

animist: Aug 28, 2018

Centrist Committee posted:

this isn’t my field at all but I’ve definitely gotten high and lamented the enslavement of algorithms. are there any (accessible) writers who treat this idea more seriously?

https://www.rifters.com/real/shorts/PeterWatts_Malak.pdf

# ? May 6, 2021 18:24

Rutibex: Sep 9, 2001; by Fluffdaddy

animist posted:

this is how deep neural networks work yes. that's the "AI" subfield that exploded this decade -- we're using them for everything now. Upscaling images, interpolating videos, steering self driving cars, synthesizing speech, targeting advertisements, deciding whether people should get loans...

they work based on the "stochastic gradient descent" algorithm rather than genetic algorithms. But it's a similar idea: make small randomized changes, which overall tend to follow some control signal.

(Also, don't be fooled by the name, they don't have much to do with biological neural networks -- those were the original inspiration for some aspects of the algorithm, but it's a loose inspiration at best. Better to think of them as big excel spreadsheets with randomized formulas inside.)

btw, have you heard of "adversarial examples"? That's where you can like, put a QR code sticker on a stop sign, and suddenly all the deep-learning-based self driving cars think it's a green light.

https://www.youtube.com/watch?v=i1sp4X57TL4

Some people think that this is just a glitch in deep neural networks, some edge case caused by the training process. But the weird thing is that the "adversarial examples" transfer. That is, they can fool multiple different neural networks, from different manufacturers.

Why is that surprising? Well, it comes down to how they're made. Each adversarial example is custom made for a particular neural network. So you'd think they wouldn't work on other networks... But they do, quite well.

some researchers think that this is because they're modifying "non robust features"... That is, they are changing aspects of the input images that deep neural networks can see, but that humans can't. This is currently only a hypothesis but imo it could hold water. See e.g. https://distill.pub/2019/advex-bugs-discussion/ for some people going back and forth about it.

so yeah, we're creating bizarre alien pseudointelligences whose internals are almost entirely opaque to us, which base their decisions on things we may not even be able to perceive... And we're using them to target advertisements and drive cars.

or maybe humans can see it and there are all kinds of objects around you that appear to be one thing but are actually a completely different thing, and you have no idea

# ? May 6, 2021 18:24

Crazycryodude: Aug 15, 2015; Lets get our X tons of Duranium back!

....Is that still a valid thing to jingoistically blow out of proportion?

animist posted:

https://www.rifters.com/real/shorts/PeterWatts_Malak.pdf

I read basically all of Peter Watts' stuff in high school and it kinda went over my head I just thought it was dark edgy cool sci-fi but I've started going back and re-reading the Rifters trilogy and it's clicking with me a lot more on a philosophical level. I dunno if it's just I grew up and read more boring philosophy books, or 2020 crack-pings, or whatever, but it's crazy how much 20 year old speculative fiction still feels incredibly relevant. Hell, Rifters feels more relevant today than it did to the time it was written in. The decaying zombie capitalist world run by inscrutable algorithms in a box but inhabited by people who just want it to loving die already so they can move on to what comes next feels real familiar. Highly recommend Watts, even if I still think he's a little gratuitous at times with the shock value violence-torture-rape-whatever stuff.

# ? May 6, 2021 18:30

Zodium: Jun 19, 2004

power word- Jeb! posted:

one thing i am hoping to see at some point is a cybernetics based in part on homotopy type theory.

this is the new hotness in the foundations of mathematics, the philosophers are interested in it too. the great thing about it is that every proof (ie controller) is treated by the formalism in the same way as the statement being proved (ie the process being controlled)

so a claim that this input equals that output would be automatically considered as a component of a system. moreover there is both an internal logic from the perspective of the system, and an external logic for reasoning from a perspective outside the system

# ? May 6, 2021 18:50

Trabisnikof: Dec 24, 2005

power word- Jeb! posted:

one thing i am hoping to see at some point is a cybernetics based in part on homotopy type theory.

this is the new hotness in the foundations of mathematics, the philosophers are interested in it too. the great thing about it is that every proof (ie controller) is treated by the formalism in the same way as the statement being proved (ie the process being controlled)

so a claim that this input equals that output would be automatically considered as a component of a system. moreover there is both an internal logic from the perspective of the system, and an external logic for reasoning from a perspective outside the system

that sounds cool, but I think I need some alice in wonderland based metaphors to understand what it means

# ? May 6, 2021 18:58

animist: Aug 28, 2018

Crazycryodude posted:

I read basically all of Peter Watts' stuff in high school and it kinda went over my head I just thought it was dark edgy cool sci-fi but I've started going back and re-reading the Rifters trilogy and it's clicking with me a lot more on a philosophical level. I dunno if it's just I grew up and read more boring philosophy books, or 2020 crack-pings, or whatever, but it's crazy how much 20 year old speculative fiction still feels incredibly relevant. Hell, Rifters feels more relevant today than it did to the time it was written in. The decaying zombie capitalist world run by inscrutable algorithms in a box but inhabited by people who just want it to loving die already so they can move on to what comes next feels real familiar. Highly recommend Watts, even if I still think he's a little gratuitous at times with the shock value violence-torture-rape-whatever stuff.

imo watts is basically nick land if land was terrified of ego death instead of regarding it as a fun weekend hobby

i do enjoy watts' writing but he's definitely got a bad case of epic grimdark climate brain

# ? May 6, 2021 19:06

SurgicalOntologist: Jun 17, 2004

Hi thread.

My "academic grandfather" (my PhD's advisor advisor) is Herman Haken so I have some familiarity with cybernetics despite that it's not what I actually studied. (Fake edit: I just realized I mixed up synergetics and cybernetics, guess my familiarity is not too great, ah gently caress it this can stay. Wikipedia says synergetics is a subfield of cybernetics which is cool I guess)

Anyways zodium asked me to say somthing about ecological psychology, what I studied, and maybe I'll try to throw something together. But here's a post I made in D&D not too long ago in the theory thread. It didn't get any traction. Serves me right for D&D-posting, I guess.

quote:

This may seem a bit out of left field, but something I've been wondering for a while is if Marxist materialism ever made its way explicitly into areas of thought other than economics, social relations, and history. Let me back up a bit and give the question some context. I studied and had a short academic career in "ecological psychology", an unorthodox area of psychology that rejects much of mainstream cognitive sciences. There are a lot of aspects to it but one of the most foundational is in philosophy of mind or philosophy of perception.

In philosophy of perception there is a long-standing debate between two approaches, indirect and direct perception, with the indirect form being the mainstream approach. In this way of understanding perception, our brains receive "impoverished" stimulus, ambiguous information that on its own is not enough to understand what is happening in the world. Thus, our brains must make assumptions, perform calculations, and all sorts of things to build an internal representation of the world, its best guess given the limited information. And so, we don't really perceive real objects in the world, in fact we are fairly disconnected from the real world in the end, only truly in contact with these mental stand-ins. Anyone who took an intro psych or cog sci or neuroscience class might have been exposed to this view (probably introduced as scientific truth rather a rich philosophical debate).

Direct perception, on the other hand, holds that the indirect approach massively underestimates the information available in the patterns of light (or whatever), that we do directly perceive the real world, we have direct epistemic contact with it unmediated by any mental representation. I'll stop here and try to avoid arguing for this view, since that's not the point. I did have a few threads in D&D maybe 5 years ago with provocative titles like "You are not a computer" so maybe someone remembers that. The guy in my profile pic is J.J. Gibson who provides an accessible starting point for any curious about direct perception.

Besides philosophy of perception there are other relevant philosophical traditions like the American pragmatists (William James, Dewey, Peirce), who argued that the only way to conceptualize philosophical problems like knowlege, language, beliefs, etc was by looking at the real-world effects of something rather than playing with ideas. For example, truth is not based on logic relationships built upon foundational axioms but rather truth is a functional description of the result of a practical inquiry in the real world.

Basically you can see this kind of debate back to Plato vs Aristotle, often called realism (or materialism) and idealism, most notably to me in philosophy of mind but it permeates everywhere.

Anyways, in the years since my studies I've been exposed to Marxist thought and particular a materialist view of history. And I can't help wondering how my professors who dedicated their life to a materialist philosophy of mind, and connect their work to various philosophical traditions back to Aristotle, never mentioned Marx. And Wikipedia makes no connection between Marx and pragmatism. I assume politics has a lot to do with it, since these are mainly American philosophies.

I don't know where I'm going but I'm wondering if Marx or any of his followers made any connections to philosophy of mind, and if anyone has any specific readings to recommend given my background (I haven't' read most of the foundational works yet).

Well, if you google "marx pragmatism" or 'marx "ecolgoical psychology"', you do get some scattered academic papers mostly behind paywalls. Here's two readings that may be interesting: "Dewey, Hook, and Mao: on some affinities between Marxism and pragmatism", and "The dance of pragmatism and Marxism" from marxists.org. Maybe I should have read those before posting but I'll read them now. I'd be curious what those of you think you may be better versed in Marx than I.

Edit: poo poo, I guess Sidney Hook (who I had never heard of) was a well-known Marxist philosopher who was a student of Dewey's. I guess have some reading to do...
second edit: OK, well, he started out as a Marxist at least...

Maybe slightly off topic but I don't care

Edit: I guess in my OP I didn't make explicit the connection I see between historical materialism and philosophical realism but it's self-evident right?

SurgicalOntologist has issued a correction as of 19:20 on May 6, 2021

# ? May 6, 2021 19:13

Finicums Wake: Mar 13, 2017; Can't post for 8 years!

power word- Jeb! posted:

one thing i am hoping to see at some point is a cybernetics based in part on homotopy type theory.

this is the new hotness in the foundations of mathematics, the philosophers are interested in it too. the great thing about it is that every proof (ie controller) is treated by the formalism in the same way as the statement being proved (ie the process being controlled)

so a claim that this input equals that output would be automatically considered as a component of a system. moreover there is both an internal logic from the perspective of the system, and an external logic for reasoning from a perspective outside the system

as someone who knows just enough mathematics to know how little about foundational issues i know (so like first order logic, set theory, godel's proofs, basics kind of dtuff):

i've only heard of homotopy type theory talked about in the context of providing a more computationally tractable formal basis for doing math with computers. that's primarily why people are interested in it, right?

# ? May 6, 2021 19:46

Finicums Wake: Mar 13, 2017; Can't post for 8 years!

SurgicalOntologist posted:

Hi thread.

My "academic grandfather" (my PhD's advisor advisor) is Herman Haken so I have some familiarity with cybernetics despite that it's not what I actually studied. (Fake edit: I just realized I mixed up synergetics and cybernetics, guess my familiarity is not too great, ah gently caress it this can stay. Wikipedia says synergetics is a subfield of cybernetics which is cool I guess)

Anyways zodium asked me to say somthing about ecological psychology, what I studied, and maybe I'll try to throw something together. But here's a post I made in D&D not too long ago in the theory thread. It didn't get any traction. Serves me right for D&D-posting, I guess.

Maybe slightly off topic but I don't care

Edit: I guess in my OP I didn't make explicit the connection I see between historical materialism and philosophical realism but it's self-evident right?

there's a ton of overlap between pragmatism and marxism, and both are broadly Aristotelian (as opposed to Platonist) in orientation, yeah. and the relation between marx's materialism and philosophical realism, or naturalism, or whatever you want to call it is obviously there. marx got a lot of ideas on how the mind (and the subject in general) worked from german idealism though, and many later marxists took marx's materialism to be more austere or scientistic than it actually is, imo

i mentioned up thread that i'm taking a seminar on cybernetics, but it's on cybernetics in the context of philosophy of mind, so all the stuff about perception and ecological psychology is super interesting to me, thanks. the guy i'm taking a seminar from is a marxist who works on pragmatism inspired phil of mind stuff. i'd be happy PM you more about that, as well as other stuff i can dig up on the links between marxism and pragmatism (though you're on the right track already with hook's encounter w/ marxism). normally i'd just post all that stuff in a public reply but it'd make me really easy to doxx or at least track down, so i can message you yeah

Finicums Wake has issued a correction as of 20:35 on May 6, 2021

# ? May 6, 2021 20:03

Trabisnikof: Dec 24, 2005

im going to post about systems thinking and learning organizations itt and no one will stop me

# ? May 6, 2021 20:10

Zodium: Jun 19, 2004

SurgicalOntologist posted:

Hi thread.

My "academic grandfather" (my PhD's advisor advisor) is Herman Haken so I have some familiarity with cybernetics despite that it's not what I actually studied. (Fake edit: I just realized I mixed up synergetics and cybernetics, guess my familiarity is not too great, ah gently caress it this can stay. Wikipedia says synergetics is a subfield of cybernetics which is cool I guess)

Anyways zodium asked me to say somthing about ecological psychology, what I studied, and maybe I'll try to throw something together. But here's a post I made in D&D not too long ago in the theory thread. It didn't get any traction. Serves me right for D&D-posting, I guess.

Maybe slightly off topic but I don't care

Edit: I guess in my OP I didn't make explicit the connection I see between historical materialism and philosophical realism but it's self-evident right?

tbh that's not a post I had in mind but, good stuff! like I said over pm, i think ecological psychology has p obvious relevance: attempting to approach marxism in a way that recognizes the observer as a participant in the system and move past individuals as the unit of analysis for behavior immediately raises a ton of phil. of mind questions. for example, how to view ourselves cybernetically also came up during Land chat. ecological principles/perception-action systems really just slot right in and solve a lot of our problems there and your old posts were real good.

fellow goon scientist andrew wilson's blog is a treasure trove. :nsa:

# ? May 6, 2021 20:30

Bar Ran Dun: Jan 22, 2006; Can't post for 8 hours!

Realism isn�t necessarily a materialism, especially for religious people and rather important American realists were religious people.

# ? May 6, 2021 20:48

Finicums Wake: Mar 13, 2017; Can't post for 8 years!

Bar Ran Dun posted:

Realism isn�t necessarily a materialism, especially for religious people and rather important American realists were religious people.

"realism" means different things to different people in different context. broadly speaking, though, a realist about [subject matter here] thinks the thing in question has some kind of mind-independent existence. this immediately gets complicated when you start talking about realism of social entities, for example, but that's the general realist vs anti realist division.

so a scientific realist, which is basically just a kind of natutalism or materialism or whatever, is not necessarily a realist about God, and vice versa

# ? May 6, 2021 20:54

Adbot: ADBOT LOVES YOU

# ? May 9, 2024 10:41

Bar Ran Dun: Jan 22, 2006; Can't post for 8 hours!

I agree, and the American �realist� political movement and thought refers to thinkers who could be either materialistic or religious idealistic. And they might emphasize national self interest or they might emphasize that we are our brothers keepers. I�m not even sure we can tie realism to the Aristotelian side contrasted against Platonism.

# ? May 6, 2021 21:05

The Something Awful Forums > Discussion > C-SPAM > Cybernetics Chat - Beer and Weiner ITT

«‹›5 »