I started reading “The Unaccountability Machine” which is actually quite interesting. I don’t know what I think about the subject matter, but it touches on a lot of things I’ve thought about over the years. It seems to describe a different economic model, not capitalism and not communism, but a systems theory field called “Cybernetics”. For some reason it’s like I asked someone: explain economics to me like I’m a programmer. And now my brain is slowly reworking itself to understand economics like I would understand a program. It feels very unnatural as a human having grown up in this *waves vaguely at the world* but it is the first time I am starting to feel that it makes sense. It’s still completely wrong, of course. I fundamentally disagree with some extremely axiomic parts of capitalism, but I suddenly might not end up with a parse error.
https://social.vivaldi.net/@Patricia/112626658364328553
Vivaldi SocialThanks to
@vaurora for pointing me in this direction 🤓
I guess everyone probably needs to get stuff explained in a way that ties in with the things they already know. I’m sure this probably is a whole field in pedagogy.
I have been sidetracked into what is a side quest in the book. A strange electronic system set up in Chile under Allende before the dictatorship. I was looking on Wikipedia and found this podcast:
https://the-santiago-boys.com/ imagine grafana dashboards and data driven decisions meets socialism and… Star Trek aesthetics.
The Santiago Boys
This is a podcast about a continent that dared to dream big - and challenge the power of Big Tech before Big Tech was a thing.
Project Cybersyn - Wikipedia
Also, even if I kind of grew up in this, I hadn’t really thought about it for years. But one of the big things Latin America saw in this period were dictatorships brought about by large international corporations (often backed by the US) that preferred a “strong leader” so that they could extract the country’s resources in peace without being bothered by the pesky people who lived there. Ref the expression Banana Republic.
This scholarship might be progressively more relevant for the rest of the world going forward.
The US had their hand in pretty much every single military dictatorship in Latin America, this included money (remember Iran-Contras?), military training and supplies (everything to equip your friendly military dictatorship) and running disinformation campaigns.
Finished first read, but now I need to reread the beginning. Current impression is that I find the arguments individually interesting. I’m not sure the underlying logical argument they are a part of is logical or if the conclusion is a logical result. But I still found the book very interesting. Especially the first 3/4ths. The basic logical argument he seems to present is that current capitalism has constructed a mechanism that forces managers of publicly traded companies to serve shortterm goals. And that this makes it impossible for companies to adapt sufficiently to accommodate looming crises like climate change. When leadership has to deliver on quarterly numbers they cannot steer according to any other metrics. I think his idea is that if we can somehow break that mechanism then… and this part I’m unclear on and I’m not sure I find it very interesting either.
Basically I don’t know how I feel about the broader logical argument, but I found the discussion to get there very interesting.
He also seems to think that the international tendency to elect populist wannabe dictators is a cry for help. That seems a bit… infantilizing?
I think I need to have lunch with an economist. I have questions.
I think the people for which Cybernetics will make the most sense are people who work in vulnerability research and exploit development. This systems theory fits very well with a lot of their findings and methodology.
Possibly also folks working in machine learning
And people in distributed systems tbh
Basically any field where you have acknowledged that the systems you are concerned with are complex and might never be fully understood, but where you still have stuff to do and you work with the knowledge that these systems are adaptive and might have emergent behavior. And that poking in one place might have unknown consequences. Where the primary tools are around monitoring and feedback to detect and diagnose new behaviors. And to manipulate these complex networks of stuff to do the thing you want, all the while knowing that it might very well also do something completely different and probably not desired.
Oh shit, this fits with medicine too, doesn’t it,
@siljelb ?
I guess the fundamental difference from modeling is that even though you might try on occasion to reduce things to a formal model and reason about things in that formal model, you don’t really believe in it. You know the real thing is “alive” and that the information that is removed to model whatever process you are concerned with, will contain data that is essential and even more importantly: you know that there is essential data that has been removed that you don’t even know is essential.
As a programmer, it’s funny to read about people believing in their models. It’s a very junior developer thing to do. Senior folks don’t. They do their best and then they observe. And even more interestingly: if they don’t see unintended consequences of significant changes, they become suspicious.
The more I think about this, the more I need an economist to tell me this isn’t how they actually work and what they actually believe. Because that’s just bananas. First let’s “simplify” a complex system into a formal model and then you reason in it, maybe run some simulations and then… you believe it? And this idea of believing in a “market” as some sort of entity that will not only stabilize into a state, but that that state is somehow “good”, or even “optimal”. That’s bananas. That’s a fucking religion.
I’m sorry, but that’s on level with worshiping kubernetes. That’s fucking mind boggling.
I guess I have to read Piketty’s Capital book now. Because this is blowing my mind.
https://social.vivaldi.net/@Patricia/112644094554478633@[email protected] the “market” isn’t an entity. It’s a distributed system, and a distributed system isn’t a moral thing. It’s a thing that is serving the purpose it’s supposed to serve or not, and usually somewhere between those. And we try to keep it on that path, we don’t sit back and… wait for it to do its thing? And believe it’s Good because… I don’t even know. This is a religion. This isn’t something that can be taken seriously.
This world needs a whole lot of intellectuals who are not privileged white men with way too much time on their hands.
By the way, I wouldn’t expect me to be more impressed by communism.
Gah… I have to read Marx too now?
I’m rereading the beginning again, and he talks about (in my words, field and terminology) how complex distributed systems will, when changed, either enter a new “steady state” (“homeostasis” in his terminology) or crash. This is because the system will respond to a major shift to it. Either by adapting to it or bringing it down. The adaption doesn’t have to be good however, which is something he also emphasizes. What surprised me about Twitter was how resilient a system can be to destructive forces. But that doesn’t mean it comes out in a better state. The new state can be what we in tech call a “Graceful degradation”, some things might be worse or down, but it isn’t all down. He also describes what might be this, as a state that is more vulnerable for future changes, fragile if you will.
Or in his terminology where the systems responses can be both negative and positive feedback, one counteracting the other amplifying (my brain went to waves and the mythbuster episodes) - purely tech and concrete my thoughts went to the Natanz centrifuges in Iran and Stuxnet.
But as for the amplification I thought of stories told by Audun Ytterdal on incident response at the Norwegian newspaper VG, and how they DDOSed themselves. (Which I think twitter also did a few months ago?)
https://youtu.be/gxfkw99k4Js?si=3-S-AcbGBdzu6oB_
Incidents and incident handling @ VG.no - Audun Ytterdal - NDC Security 2024
The point I’m getting to is that these systems are complex, they might not even be trivial to bring up again if they go down because they end up having cyclical dependencies, so some subsystems might not come up because another isn’t up yet (which is also waiting for the first). So we can have “cold boot” deadlocks.
Trying to “simplify” these systems to a model, draw conclusions and assert them with any confidence *about the real system*, is naive at best.
This behavior of complex distributed systems has led to a whole subfield in testing often referred to as Chaos Engineering. “Let’s kill this thing and see what breaks” 😈
Lol, now I remembered the financial crisis mantra “Too Big To Fail” 💀😂
Ok, I’m still re-reading the beginning which is the part about Cybernetics and tbh it sounds like a kubernetes cluster 😂
What if we designed kubernetes in the 70s?
I digressed into the more “political” offshoot of cybernetics, clearly springing out inspiration from “Project Cybersyn” and it seems intellectually thin so far. It seems to boil it all down to: “of course any modern socialism* would have to include technology”
And tbh that is both extremely shallow and broadly uninteresting.
The rest seems to be a lot of “capitalism is bad” ranting which… yeah sure, but that’s not a thought, that’s more of a feeling.
*they mean communism, I think
@Patricia I may take exception with "broadly uninteresting": I'd say the only interesting thing about any concept, system or idea is how one might actually make it work, how it would be implemented in actual detail, how one might deal with the unforeseen problems and diversions and emergent behaviors.
Socialism as a "simplified model" was a solved problem by the 1920ies, but to have a truly deliberate, planned economy requires the ability to track, foresee, plan, orchestrate literally hundreds of thousands of parameters, all interacting in sometimes nonlinear ways with each other.
This is the 𝑠𝑜𝑐𝑖𝑎𝑙𝑖𝑠𝑡 𝑐𝑜𝑚𝑝𝑢𝑡𝑎𝑡𝑖𝑜𝑛 𝑝𝑟𝑜𝑏𝑙𝑒𝑚 and we're only now maybe slowly getting to the point where human computational resources are even vaguely up to actually solving even the easier strata of that problem.
It is possible that we will find in the end that an economy complex enough to produce the computational infrastructure to run that economy is too complex to be modeled by the infrastructure. Like the old adage that if the universe was simple enough to be fully understood by humans then humans would be too simple to fully understand it. But the only way to know is to persistently try - kudos to the soviet engineers of the 1950ies who tried to model a national economy with what counted as "computers" at the time - they stood no chance, of course. But if a modern attempt was started right now it would of course build on much better technological tools - and might just pull off what the first generation attempted (and was in fact successful in - for about a decade or so).
Since you seem to be up for reading suggestions, there's no better intro to socialism as a computational problem than
https://crookedtimber.org/2012/05/30/in-soviet-union-optimization-problem-solves-you/
In Soviet Union, Optimization Problem Solves You — Crooked Timber
You all know that I wrote this with economics in mind. But read it again, now with tech in mind. And hopefully you’ll see why I am infinitely skeptical of tech (Project Cybersyn version 2.0 etc) running society too. Yeah we can handle distributed complex adaptive ✨computer✨ systems. But what is our track record with humans?
https://social.vivaldi.net/@Patricia/112694261543358964
@Patricia
Fair enough.
Might it pique your curiosity when I note that that link up there does in fact go to a (long) blog post, not a book-that-could-have-been-a-blog-post? It uses a book as a springboard to look into the actual computational complexity challenges, it has real equations and links to sources, many of them even real scientific publications (mostly in computer science, not economics). It's what opened my eyes to the idea that socialism is largely a computational problem...
Patricia Aas
Σ(i³) = (Σi)²