Varn Vlog
Abandon all hope ye who subscribe here. Varn Vlog is the pod of C. Derick Varn. We combine the conversation on philosophy, political economy, art, history, culture, anthropology, and geopolitics from a left-wing and culturally informed perspective. We approach the world from a historical lens with an eye for hard truths and structural analysis.
Varn Vlog
Exploring Language Modeling and Cybernetics: A Conversation with Mark Rainey
Ever wondered how the complexity of Language Modeling (LLMs) and cybernetics can revolutionize the way we communicate and interact? Are you curious about the sociopolitical implications of such advancements? Join us in an enlightening discussion with the well-versed Mark Rainey, as we dissect these technologies and their potential impact on our society.
Our journey begins with an exploration of LLMs, their potential, and inherent limitations. We discuss the nuanced logic of human language, Boolean operators, and their influence on the design of these systems. We then shift gears to delve into the intriguing world of cybernetics, viable systems, and the behavioralist and cognitivist wars. We scrutinize the implications of these advancements and the tests used to measure intelligence, all the while contemplating the potential and pitfalls of Stafford Beer's Viable System Model.
As we navigate further, we probe into the relationship between technology and Marxism, questioning the teleological feedback loop of capital and its effects on the proletariat. We also explore the exciting realm of cybernetic planning and the potential role of LLMs in such systems. Finally, we reflect on the concepts of agency, alienation, class dynamics, and the implications of capitalism on social reproduction. This rich and riveting conversation with Mark Rainey is not to be missed!
Crew:
Host: C. Derick Varn
Intro and Outro Music by Bitter Lake.
Intro Video Design: Jason Myles
Art Design: Corn and C. Derick Varn
Links and Social Media:
twitter: @varnvlog
blue sky: @varnvlog.bsky.social
You can find the additional streams on Youtube
Current Patreon at the Sponsor Tier: Jordan Sheldon, Mark J. Matthews, Lindsay Kimbrough, RedWolf
Hello and welcome to VarmVlog, and today I'm here with Mark Rainey Mark is a long time from the show One of the few people I know who actually read Capital's Volume 1 through 3 and theories of surplus value and read it closely. We talked about this a few years ago but you're one of the few people I know who actually did it. Most of the people I know who talk about Capital get through about volume 1 and get through half of volume 2 and quit.
Speaker 2:It's a nice small task.
Speaker 1:Yeah, it is what like 5,000 pages or something. So, however, I'm here to talk to you about one of your other hobby obsessions. You and I got into cybernetic theory at the same time. I am more of a noted cybernetic friendly skeptic than you, although my reasons for that maybe we can get into later. But we've also talked about LLMs a lot, and that was a big LLM skeptic in the beginning.
Speaker 1:I've read a lot of the work of Duane Monroe and I see his concerns. I've talked to Nicholas Villarreal on this as well. So Duane seems to think LLMs are kind of a scam. That is his object of expertise. But I've talked to a lot of other people who are not normally given towards tech optimism, who have far more complicated views in the last two months. One is Nicholas Villarreal, who's talked to me a lot about what he sees as, kind of like others, there are oversells in what LLMs can do, or at least the threat that they pose, and that itself is a form of marketing, but that they are actually really useful technologies that, in the right hands, could be massively labor saving and, with the right development, really could be a kind of intelligence that approaches something human, although he seems to think that, under capitalist control, that that's not how they're going to be used or what they're going to end up being. You've played with them a lot and done a lot with them, and you've been very optimistic about what they can do.
Speaker 1:I have recently started playing with them in the context of education, where I think they can be both helpful, but how they're going to be utilized in education is actually quite worrying, both from the standpoint of teachers and from the standpoint of students not doing shit, and I can already tell you what I'm receiving is people who don't even know that this is a large language model tag out of the thing that they copied. Oh no, they can't even cheat. Well, I'm like no, I don't have to use the AI checker on this. I know what you're doing.
Speaker 1:But I've come around, though, on this idea that if we take its energy input limitations in the account and we take that this is going to be as good as the kinds of things we put into it, and if we started, say, siloing different ones of these large language models simulations off from each other so that they learn different ways, we actually probably could be doing something quite interesting with these that, under non-capitalist conditions, would actually be a major labor saving bit of software. Now, I don't think that's what's going to actually end up happening, but I wanted to talk to you about that because, like I said, you are probably the most optimistic person on my spectrum of optimism about this. And so what have you been discovering as you've used these?
Speaker 2:So it's funny to say I'm the most optimistic, because, I mean, I would say that my optimism has been maybe tempered a bit. I'm still very optimistic about the technology. I think that it's extremely powerful and useful if it's used correctly. I think one of the cool things about basically neural network technology in general is how well it takes on isomorphisms of systems that it's able to learn from, and basically what that means is it's able to act like that system through the learning process and it's applied to language in this way and it's really convincing and the things that it's able to output. I've used it to automate a lot of my own personal work and I've used it to do a lot, but I've also run into limits. I've been teaching myself some linear algebra and sometimes I'll ask for some help from it and I'll get some good outputs, but when it comes to actual computations, as everybody knows, it's not very good. So I can't rely on it for that and I'll go to things like Wolfram Alpha instead that are specialized towards numbers.
Speaker 1:So what do you think? I don't know. Actually, I shouldn't ask you what do you think, because this is actually an objective question. Why is it that if you train an algorithmic systems model on a language that is qualitative, it loses the ability to do math? It's something I was playing with the school AI program I'm currently being trained on that and they were warning us like oh yeah, this is great for all kinds of stuff except math. Don't do math. And I'm like my first response is math is the one thing that I could in the past, consistently trust kinds of AI, artificial, you know, kinds of kinds of algorithmic models to be able to do. Why is it? If we train it in this, it just loses the ability entirely to do the other thing. Right?
Speaker 2:There's something fuzzy about language, it seems. I was listening to Sam Altman talk and he made a point to where the more human feedback that's been integrated into GPT models, the worse it's actually been at doing numbers and making predictions and things like that. So I mean it's like the better it gets at language, the worse it gets at those things, and I don't know why that is. You say that there's an objective reason. I'd like to know that.
Speaker 1:Well, one of the things that I think a lot about around this is that human logic in general is fuzzy and the assumptions that we've used in going back to Boolean operators way back in the day is that we actually use a far more rigid set of logical simplifiers than we actually do. And then, when we first designed a lot of these systems, we actually took these 18th and 19th century misunderstandings of human cognition and encoded them in. Then they were maintained because you can learn them and get consistent results from them, and search engines don't use Boolean terms anymore, but when they did, they were pretty predictable on what you could figure out if you learned the Boolean system right, gotcha. One of the things I think that we have to deal with this is where I think Wittgenstein helps a lot Like is that a lot of our associations are a lot of our associations are very much limited by the fact that we actually do see links that are not logically totally justified, like if you look at a cluster of a series of words right and you look at how they might get semantically expanded or narrowed. An example that I ran across and recently is meat in old English means all food and then it gets narrowed more and more to just mean flesh, whereas you know, if you look at all the stuff around blood and heart, that's in English and it shows up all these places both from French and Anglo-Saxon right, we see all these expansions of these categories that are rhyming Like. If you think about it you can see how they are associated together and how we got semantic meaning from that and they're definitely like resemblances of things, like there's types of things where you say, okay, I can see that, but it's not. It doesn't occur naturally to me, particularly now, 1,000 years, 2,000 years down the line from the initial association.
Speaker 1:Math doesn't work like that. Math is pretty like it removes all the qualia that we could get all mixed up in that shit in the first place. I think that might be part of the answer to why we have this problem. But it does seem interesting because humans can go back and forth between language-style reasoning and mathematical reasoning. We can teach ourselves to do that, even if, like loosely speaking, they are separate skill sets and one is left brain and one is right brain, although if anyone's actually studied the neurology of this, you know pretty quickly that that breaks down really fast Like language. Syntax is in the same part of the brain. All these things have to be also pretty much put in quotation marks because these regions are not actually that delimited, but it's in the same part of the brain as math. When we do math and we look at MRI, we see the same kinds of regions light up in the head as we see when we do in like sentence construction, but other regions light up what's in its construction and don't light up when we're talking about math. I find that very interesting.
Speaker 1:So I remember talking to someone I think it was Joseph or no Dave of Nightmare Masterclass who was pretty optimistic and I said well, you know, one of the things I'm beginning to accept is we should accept these large language models as a kind of intelligence. But since we don't understand how human intelligence actually really works, like the precise ways in which, like servo-mechanic mechanisms interplay with cognitive structures and stuff like that, we don't really completely understand that. That we're actually creating an intelligence that can do things that we can do because it's learning from the way we input patterns and whatnot, but it didn't develop it with the same kind of weird, spandrily history that we do, so it probably doesn't actually think the way we do. That doesn't mean that it's not super fucking intelligent, and I think that's the whole rub, and I think I've become convinced that it doesn't matter if it thinks like a human.
Speaker 2:Yeah, I think that that's. I agree with that. It doesn't matter if it's necessarily a human intelligence. It's going to be an intelligence and it's going to be very intelligent and it's going to be able to do operations very quickly. So we're going to have to figure out how we're going to deal with it. But I don't think that it's going to be anything like us, mainly because it's not going to have a body.
Speaker 2:In fact I have those sorts of experiences. It's going to be trained on sort of like an abstract human experience that's aggregated on the internet, and so GPT five or six who knows what we're going to be looking at. But I do think that it's going to change the world similarly to how the internet did, as far as how it interrupts human lifestyles. And you already see and I think it was Wendy's that might be testing replacing draft through tellers with GPT technology, no-transcript. You know that that kind of stuff is gonna hurt the working class. A lot of jobs are in that kind of work and you know it's not just the, the white collar workers that are gonna be feeling it, and you know, when you look at it that way, it's just a matter of time.
Speaker 1:Yeah, I, I Think you're right about that. I think, and one of the things I think it's just a matter of time that we have to look at, is this being rolled out at a time period where we are having a labor shortage we could not build up the human capital for, for mass armies of coders. There was a, there was a liquidity crunch, and tech Because of the increase in cross and debt, because so much tech is not actually that profitable, like it's rent, it's rent sources versus its commodity inputs, versus its R&D costs, are like way out of balance. And so it's only I've ever been able to leverage that by like basically hoping that interest outpaced debt costs to like increase the, the, the Increase to increase nominal wealth, but like like when people ask me, for example Well, what happened all that wealth that like meta and Elon Musk and stuff lost, and I'm like it wasn't valorizable, it wasn't real, they could not have ever spent it into reality, anyway, so it's not like it, it's not like it went to someone else, it's just gone like so how do they cope with that?
Speaker 1:Well, all these technologies that we think have been simmering for a while, um, I think have very clearly been been kind of pushed the forefront. You know, in light of that, I mean I don't think it's any accident I've said this now in three shows, but I don't think it's any accident that this happened when it happened. I don't think it was conspiracy either. I think, like we, they were sitting on all this tech, they, you know chat. We've had the slow development of the like chat dpd1 and chat dpd2 and Then all of a sudden we need to really push something out that can help us off, like coding.
Speaker 1:And do all these other things because, honestly, also one of the questions that I've had that this sort of quasi solves is how do you re industrialize when you start to decouple our friend couple or whatever? Internationally, you start moving, you start breaking down these long-established multinational supply chains and you know the US does a lot of production, but there's a lot of finishing production and now you need it to do a lot of other stuff. You have a demographic crisis. It is politically unviable, for whatever reason, to actually deal with mass immigration. So what do you do? Well, maybe the idea could be as like, let's push this stuff out, it'll help us offside a lot of work. Also, it'll create a lot of people back opened up for blue collar style, are, are, are even skilled industrial work for this kind of Industrial development that we need to do, but we're not gonna have the capacity to do, for you know who knows how many years, like I mean, even even with this, is gonna take what? Probably a decade to start getting these, these physical Systems up. You can't automate them in the same, in this kind of way, like. So I mean, you know, I think, I think there's that and you're right there's. This is pushing back on a On wage growth, I think I think also You've seen the Fed back off on on interest rate rising right, and I know people gonna be like, oh Varn, you sound crazy, but I think it's related, like because I'm like, oh, they were only doing that because I thought it could help on the 1% Of inflation, they could control that 1% being wage growth inflation, which is a very small percentage of like the 5 to 8% inflation that we were dealing with, but like it's something they can do something about. They can't do anything about international supply chains, they can't do anything about geopolitics. You know they can't do that much about our own production stuff right now, but it's hard to do anything about fucking birthrights. I mean, and like I said, for whatever reason, the political establishment not just the Republicans has become hostile to relaxing immigration, which was the last way they fix these sorts of problems which you know. I guess this gets us into, like our interest in cybernetics, like you're interested in cyber.
Speaker 1:Next, and I'm interested in cybernetics, I told you my critique of cybernetics is that Norbert wiener to some degree Ashbury over relied on the server mechanic mechanism and behavioralist assumptions about human beings which were very popular both in the Soviet and the Western world and the American Union. They're both in the Soviet and the Western world and the in the 40s and 50s. I can't, I don't blame them for that, I just think you know. And the behavioralist, cognitivist, wars the cognitivist one. I think we've probably gone too far in the cognitivist direction and, by the way, for people don't know what we mean by cognitivist Cognitivist believe basically that there's a static human being, that like there's cognitive structures in our mind that they don't change. Chomsky is a cognitivist right. If you are a pure behavioralist, you, you either believe in the weak version, which is like we don't know, we can't know what human nature is, we can only know human outcomes Because of social moderation, etc. So we're gonna have to just Deal with. What we can deal with are the strong version. There is no human nature at all, everything is inputs and outputs, which I think the first position is pretty defensible. The second position is not. But we've kind of gone way over in the other direction on cognitivism and I have been sort of frustrated because I've been hoping there's been like in this Particularly, maybe with this LLM model stuff, we could start to answer some of these questions, because Right now we can't really answer what we can.
Speaker 1:We can for, for example, we can say behavioralist assumptions on the strong variety of are wrong, but what we cannot say is how much, because we know that, like, say, pavlovian inputs work on people. So what we can't say is how much Behavioralist assumptions Actually can be utilized and what, what. What cognitive structures can we change, you know, and what can't we? How plastic is human intelligence and and how most? I faceted this human intelligence and it's really hard to measure.
Speaker 1:Like, like we all know, like people know the insufficiency of the IQ test and they, you know that. Thank you, steven J Gould. But, like, almost any language-based test is gonna have some insufficiencies for testing Intelligence because there's all kinds of linguistic assumptions that they're gonna have that other kinds of manifestations of intelligence and abstract reasoning may not show up on right and, and so you know this is the classic thing that's brought up as like, if you take abstract intelligence tests and look at Like, what is it? Like hunter-gatherers, they have an IQ of like 60, but they have skill sets, that and things that I can't imagine. Like, like, literally, they can, they can, they can reason in ways and relationships like inputs and nature and whatnot, that, like, we can't really do anymore. And so calling them less intelligence than us because they're not as good at math in the way that we articulate math even seems to be Racist AF, but are even Even prejudice AF, like, like, because we're also like discounting our own ancestors.
Speaker 1:Conversely, there's a lot of people think something like IQ doesn't measure anything, and that's not true. It does it, you know, within the parameters and limitations that I just talked about. So, yeah, when it comes to cybernetics, there's a lot of brilliant rules that really help you out Recursion, requisite, variety. But one of the things that we have to ask ourselves like, and what ways is recursion in humans? Are recursion in mammals Like, different from recursion in a machine, from the servo mechanic effect? And and I don't think we actually Like people go oh, we know, I don't think we actually know in either direction.
Speaker 1:There's a whole lot of hard claims being made. Oh, the brain is just a, it's just a machine. Well, yes and no, I don't know what you mean. We machines do a lot of different things. Man why? Yeah, and then there, you know. So. So I Get the optimism. I've talked to you about it. I sometimes feel that, like when, when we talk to people now who are interested in cybernetics, they always go back to like the big old guys like Stanford beer, norbert wiener. If they're smart they'll read Ashby. I think Ashby's actually. I've increasingly decided Ashby's actually the more interesting one.
Speaker 2:Wiener is great though he. He is really interesting. But Ashby's intro to cybernetics Made cybernetics a little bit more real. But I mean, I always do fall back on Stafford beer. His, his viable system model gives a Good framework. A lot of you know just rules of thumbs. A lot of you know just different ways to look at systems and and consider why they may not be functioning as well as they should be. And you know recursion and Variety is at the heart of how those analyses take place. But there's also a way that you can misuse the Bible system model and that is to not see it in relation to variety, flows and Recursion. A lot of people look at the Bible system model and want to see it as like, as a corporate structure, like a way to to assign blame to different elements of a system. But that's not. That's not how it's met.
Speaker 1:Yeah, so here's one I find. Here's what I was listening to. You know what's that podcast? Auxiliary materials.
Speaker 1:Okay, it's a podcast out of the UK okay and and they were talking about their big Stanford beer heads. You can tell they're in contact with the general intellect unit. They have that podcast and. But they found a guy, roger Avalos, who'd like studied Worker systems and and he was talking about beer and he's like well, the viable systems model was fine, but if you try to design everything off of it you're gonna come into problems. One he says like you don't really need all systems to be viable. Some systems need to just survive immediately and to.
Speaker 1:Sometimes it's so abstract and it's deliberately abstract for useful mistake, right, it doesn't actually give you clear guidance for the reasons that you're hinting at. Like you can. You can treat it as a corporate structure. You can treat it as a corporate check and not deal with the fact that that, like beer beer system is a good way to check if all the systems are working in In a system is supposed to survive. I absolutely agree there. In fact I was gonna do a video on that. Like I can't think of a better way to conceive of it to check if things are flowing. But you can have a very despotic system that meets the viability Mechanic. That does it like all parts of it are operating fairly correctly, but it's still a very top-down model, I think you can accept.
Speaker 2:There's one aspect of Of the viable system model that I think really blows the ability for an authoritarian regime to be successful in the long run and.
Speaker 2:So Stafford beer talks about I don't want to get into the system numbers, but the components. The components of the system, the primary workers. You know, in a body it may be the cells, in a society it may be the people, and in a workplace it's the workers. So One, one thing that's required for a system to be viable is for those component systems To be what beer said is maximally free and mentally constrained. And their day-to-day activity and that's a requirement because they have an awareness of their environment and their needs, the needs of their jobs and the system around them that you know, a Higher system, say, the manager or the boss or the dictator can't see, and the dictator knows that, the CEO knows that that wants to control everything. So what has to happen is there has to be extreme constraint from above.
Speaker 2:But that makes for a brittle system because the, the systems below the systems, one has to make, they have to make decisions in real time in order to be Viable. You know you have to to react to a constantly changing environment and if you're so constrained that you can't react, then you're not going to survive. And I think that a lot of that is Is is key to sort of understanding why is why capitalism isn't working very well. The working class is so constrained that the Viability is just not a question anymore. We don't have the freedom to react in ways that keep our social systems at this recursive level functioning. And you know, like stacking dolls, you know if one recursive little fails, and they all fail above it as well. And basically, what I'm trying to say is variety. The Flow of variety has to be structured in such a way that you have you have autonomy, and in the lower systems, because it's impossible to make those decisions from above and I'm having a hard time- no, that makes sense.
Speaker 1:For those that you don't know, the systems model. These are feedback loops and One of the interesting things I think about capital and Roger Abelos, I've got podcast mention this is like you will see capital occasionally innovate ways that are more humane for workers, that are actually even more efficient, bought back for reasons that are not obvious unless you reframe it back into the terms of pure class power, not even of pure efficiency with class power, because one of the things Roger Abelow pointed out is like oh, in Japan and Norway and stuff, they actually did come up with systems to incorporate workers in ways that made both profit more efficient and make workers more happy by incorporating more worker inputs into the system, but they don't get maintained, even when they're profitable Right, and part of that makes me you know. Part of that is why I'm like, it's why my class war exists honestly, it's why when we talk about stuff like I can talk about the problems of class analysis all day long and limitations to it and how, like, when people try to reduce everything here, it is not always explanatory. I'm definitely with you, but that tendency of like even subverting like the quote rational law and value so that class power can be maintained actually does indicate something to me. Even this even more, this is probably even more aggressive than what Marks thought.
Speaker 1:I mean, marks, it's hard to say what Marks thought on this.
Speaker 1:If you read like, if you read a capital volume one through three, you definitely get the idea that Marks is trying to prove that this would be the case even if the capitalists weren't like this.
Speaker 1:Like you know, even if you have a good boss, this is going to go this way basically, but the likelihood you're going to have a good boss is still really low and that's because of like these power dynamics and what I think, and I think something like a viable systems model when you mix it with something like Michelle Rebears iron law of oligarchy, which I don't think is an iron law, right. I mean I think about like one of the reasons why I like cybernetic theory, even if I'm critical parts of it, is it gives you a way to start checking on how to keep oligarchy from developing, because one of the things that I think and I also don't want to get into the system numbers, but this middle system is a place that is right for skills hoarding, for deliberate miss sharing. I mean there is a sense like the central management of any organization, even as much as the head of the organization has incentives to like fuck with things and hoard.
Speaker 2:So the control room is what Deere would call that, the three and four mix.
Speaker 2:Yeah, I think so. But tying back into the LLMs. I think that a lot of that can be solved through this diffusion of knowledge that is offering. Imagine an LLM stronger than GPT-4, that's able to guide managers of some future socialist society. That may rotate, but it's still trained into the general ethos of our society. I think that stuff like that can be powerful in curtailing the skills capture, because that all the knowledge wouldn't have to be contained in our heads, we could just ask questions.
Speaker 1:Yeah. So I want to play this with you for the internet, because this is a great point, mark Sure. The internet has become a nightmare. It's always been commodified to some degree. I don't have this naivete about the decommodatized internet, but one thing I push it's always been equally commoditized. That's not true. Early internet development, because of its relationship to DARPA, because the World Wide Web was not patented deliberately because of international communication enabling and etc. Was actually a very hard space to commodify for a long time. I think people like to the extent that it is today. Yes, it was still commodified in the sense that we had to buy computers and maintain all this equipment and do all this stuff that's often hidden, to access it, to deal with it, etc. But one of the reasons I think the internet has become shit is that because of its commodification tendencies, and particularly because it's commodification in particular, so depending on kinds of rents and cartailing and coordinating off what was a way to increase communication, increase knowledge, share and if you want to, you can still do it has actually become a way to One of the few points where Nathan Robinson made a valid point to me to we can distribute, renown bad information, but bad information is decommodified or it has less.
Speaker 1:Many websites don't usually make you pay for them. Sensational news where you get the headline but don't actually get the details. Well, you'll get the headline from the New York Times, but you're going to have to pay for the details. This encourages an environment that could have been massively useful for any skill sharing. If anyone's ever played with YouTube, holy shit man, I have learned how to fix so much stuff in my house. At the same time, there's all kinds of knowledge that people want to squash or put behind walls or keep within certain silos. This is made the internet I think it's turned it plus. Social signaling, being a substitute for other kinds of communities, has turned the internet into a signifying hellscape where it's really easy to signal and it's really easy to do critique. But it's actually really hard to know how to get information because there's too much of it available to you and the search filters are gotten shittier. Google, for example, increasingly just gives you paid affiliate stuff, despite all this indexing.
Speaker 2:I'll have the Facebook algorithm feeding me legit conspiracy stuff all the time.
Speaker 1:Actually, and Twitter's worse. I get a lot of Some of it's mass, some of it's the fact. I think I live in Utah and it has a very broad algorithm like, oh, utah must be Republican LDS, whatever. I don't see any of that on Twitter and I'm like look, I get wire ring trigger stuff all the time and I do not follow white wingers on Twitter. I don't hate follow people. It's not me giving it that input. I think when you look at all that you see this increasing. We see what capitalism over the course of 40 years has done to a system that I remember how much stuff you could find Seven years ago.
Speaker 1:I have trouble finding stuff now. I'll give you an example. I've had people who've wanted to argue me about Alexander Dugan and I know certain things about him because I've been involved in this stuff for so long. I want to find an article where he said something that I know he said and it's almost impossible to find. I put in anything that I think is going to be a search stuff to lead it to me on Google. That worked in 2017.
Speaker 1:I'm going to get a ton of liberal propaganda, argues, conspiracy theories. I might get patriotic, socialist stuff the actual material I'm going to find if it's on the internet still at all is usually like I got to look for like 100 pages to get it. Now. That's with me knowing what it's about and vaguely knowing what I'm looking for and I still can't find it now. Then I'll get all these affiliate links and this, that and the other. Some of that's worse now that I'm back in the US. It wasn't quite this bad when I was outside of the States. Some of that, I think, is also just the algorithm getting shittier, because I've heard other people just tell me that that's the case.
Speaker 2:Sometimes it might not be in Google's best interest to send you to the best information as far as their programs go Right.
Speaker 1:Well, now, for example, I might put up find my birth certificate, just something random. I had to do this 10 years ago. It's going to give me 10 to 15 paid affiliate links before I get to a real search item. All the paid affiliate links are like scams. I'm going to pay someone to go look at this, that I could totally do myself. I'm going to pay someone who's promising something that I actually know they can't deliver on because I know enough about the law. You can't actually even do that for me. What are you even doing? You'll just get tons of that. When I first used Amazon back in the day, it would give me great book suggestions. That's not true anymore. It's trying to force me into very broad base categories. It's actually trying, instead of giving me what I want, it's trying to force me into what it wants to sell. Let's go over to these large language models. How much do you see heading that way with these large language models? I'm already seeing being feeling like it might go that way since I started using LLM tech.
Speaker 2:I think that a lot of it will go that way. Just in general, the average person is going to feel more comfortable speaking conversationally and getting information conversationally. That's one of the biggest complaints I get whenever I'm talking about. Something I'm interested in is don't use the technical words explained simply. I have trouble with that, but chatGPD doesn't. I'll take a quote that I'll find to be very inspiring. I'll try to talk to my mom about it, for example. She has a hard time understanding it. She has a high school education. If I'll put it into GPT and say explain this to an average person, it'll give me a very good. I don't want to say translation, maybe transduction of what that quote is trying to say in a way that my mom can understand it. That's not a skill that I have, but I have access to it with LLM technology For explanatory power for people that may want to change the world. I think that that is an extremely useful tool.
Speaker 1:Yeah, I was thinking of it the other day. Actually, this is one thing where I'm hopeful. I now use AI to write the not on my YouTube channel yet, but on my podcast to write the description of my shows, because it actually understands what I said. That was important probably better than me, unless I wrote it the day I did it. It can do it faster than me because if I was going to do that, I'd have to re-listen to my show after I've already edited it. If you've ever edited a podcast, you know now my podcast is lightly edited these days, but you know that by the end of that your own voice just sounds like garbled nothing to you. By the end of that process, I ran it through some of these AIs and I'm like oh no, that's good at telling me what my key points were. It's good at that. However, if I ask it to write something, that would be the kind of thing I would write. It gives me gobbity-gook, frankly.
Speaker 2:Yeah, it's highly sterilized, appealing.
Speaker 1:Yeah, it has that same feature. I used to play with this. This was a kind of AI technology way back that would like aggregate faces, right. So they did this aggregation of like, taking all the faces that we thought were attractive and putting them together and then coming up with an aggregate face of like the or attractive person. And they were boring, like that image was boring and I mean the person was fine, it wasn't anything wrong with the or attractive person, but they weren't actually like, they wouldn't stand out to you either.
Speaker 1:And that's the problem with these large language model writing samples. It's terrible at like, particularly really complicated writing or very idiosyncratic writing. It's not good at it, but it's so amazing at like. Oh no, it can figure out how to explain something to a person that I would. You know, even I, as a teacher, who I do I feel like I kind of have that skill right To explain stuff to normal people and it can do it way better than me. Like it kind of knows what the median person is gonna get. Yeah, and that's great. Also, man, does it make coding easier? As a person who's always hated coding in the past, like I'm just like, hey, write this code for me Like okay and it's like accurate, like 90% of the time. It's actually pretty useful.
Speaker 2:Yeah, you can go and fix the obvious errors.
Speaker 2:One thing that I found to be really useful as far as coding and this goes back to the information hoarding, the knowledge hoarding problem. So one of the biggest problems I found with learning how to code and this has been a pet project of mine for a minute since getting into cybernetics is learning about which packages, which libraries I can use and how to use them. And what that basically is is pre-written code that people have made available that do specific tasks. So I can have a library that reads word documents, for example, and it does a bunch of operations on that word document. But learning how to implement that in code can be.
Speaker 2:It can be a challenge for me, not for GPT, as far as knowing about the packages, knowing how to implement them. It just it provides examples. If I give it the code that I want to integrate it into, it'll do it for me and I'll just plug it right in. So where before I may be searching two or three hours, I may be asking on forums which packages, how to use it? Why is my code not working? It just doesn't, and you know that's useful. You know it's like it's knowledge I don't have and would take a long time to get.
Speaker 1:Yeah, it's also great for something like. I just speak shit up like I know some. I have some friends who work in coding and used to do a lot of like coding for social media advertising. You know, working for the devil, but they would. They told me how much faster, how much less time they had to spend on coding, how much more time they could spend on design with the chat GPT. So that stuff, I think, is really amazing. Now do I think we're gonna run Cybersyn off of a chat GPT model and it's just gonna like we'll have fully automated luxury space communism? No, I don't think so and I'm really worried about how, like you said, I'm really worried about how this is gonna be used. We've already seen what automation was used in the Americas towards the industrial working class. That's why it's now 16% of the population or less we. I always find that interesting.
Speaker 1:There's this tendency right now to wanna shrink the working class down to the blue collar working class and on one hand, I get it because, yes, college educated people are annoying. On the other hand, I'm like that's a disaster, because 80% of us work in like 80% of us work in services and sub capacity, and like pretending that everyone has a degree is actually some kind of management, is literally nuts. That's not remotely demographically justifiable. And do you really? What could the working class do with only 16% of the population? And it'd be like saying we're gonna use the peasants in the first world to be a political subject and like, okay, all of agriculture's like 2% of the working economy. And yet I mean, I'll admit, if there was no food, everything would shut down. But it's really easy to find scabs in a field that small. So yeah, I think that's gonna happen now with this and we don't. It doesn't feel like there's a whole lot of largesse in the state, even for the removal of debt, but it's much less for like something like UBI, which I'm not even necessarily for anyway.
Speaker 1:But I mean, like one of the things that I think people have to like ask themselves why was there so much focus on both college and prison in the United States? And that's like, well, you had a lot of people that you had to do something with before. Now, admittedly, we've seen prison reforms and college's capacity over like the poor people into it is declining. But I want people to like think about this as a system, like why would that be happening now. Oh, maybe we need raw labor again, like, but depending on how these things, like these LLMs, are used, it could make this a lot harder to do.
Speaker 1:Like you're saying, this could be a great empowered network with some other things you could do. Decentralize, lee. You could build a whole lot with it. You could do a whole lot of stuff with it. There's already stuff that we've seen we could be doing, like you start implementing this. You start implementing stuff like create spaces and 3D printing. We could be making tons of components right now. If we really worked on it with relatively, we could cut down anyway the kind of scale of industrial pollution that we're doing dramatically if we did the smartly right we have the capacity under current technology to do all that.
Speaker 1:Yet what do I see? 3d printing, mostly used for Stupid digits. Like are making cheap D&D tokens Are like it's just not used for anything particularly useful. And I've just come to the conclusion it's not the tech, it's the telios of the tech under capital. Like it's being used for shit. It's dumb because the stuff that would be useful would not be particularly profitable if we used it. That way.
Speaker 2:There's a concept from Wiener called teleological feedback, and I use it a lot when I'm thinking about these sorts of things. There's certain goals that are just impossible to reach with the telios of capitalism. The feedback loops are too powerful to overcome and without some radical restructuring of how we do our social reproduction. You mentioned that some people want to shrink the working class down to like 17%, and as a college educated person, I push back against that Because I feel like I'm working class too and I actually think that it's sort of the opposite.
Speaker 2:I think it's been generalized, and Marx this time, obviously it was small. The industrial working class, the Polateria, was still tiny. But as Marx described it, as far as the relationship to ownership and production, I think most people are proletarian at this point. But when you generalize something, it sort of stops mattering, and so we're finding new ways to sort of to understand what the proletariat is. But as far as we're related to production, we're still wage laborers, we still have limited capacity to develop our individuality as mediated by market relations. We don't have any ownership of the things that we use to reproduce our lives, and all that stuff still matters.
Speaker 2:Which is why I'm still a Marxist, but I think a lot of the people, a lot of Marxists, are stuck in the 19th and 20th century and it's because of the language that we use to talk about Marxism. And I think Marx would be disappointed in us for that, because he wasn't trying to make an eternal system of how to think about society. He was trying to model societies, but he was doing it with the languages that he had, and the language he had came from the romantic 19th century revolutions, and now we have better language to describe those things and it may give us more insight on how to describe those things, as long as we stick to his general ethos that striving towards liberation and looking at how labor is distributed in society. And those are things that I think that cybernetics and systems theory are much better suited for than people arguing about dialectics or the Soviet Union.
Speaker 1:Absolutely. There's things we could learn from the Soviet Union. For example, if we took a cybernetics model and applied it to Gazplan and like because I'll be honest, gazplan was not that I think we should replicate it or anything. For those of you who don't know, that's the Soviet planning system. It actually surprises me that it took as long as they did the breakdown, like I'm surprised it ever worked at all, and which tells me a lot about how close we really were on a lot of things.
Speaker 2:And yeah.
Speaker 2:I think that planning has to be totally different now. Cybersum was on the right track, but one thing that Beer points out is that planning it can't be five-year plans. That's too much of a time delay between active processes that are happening constantly and the planning process. And one thing that I think that LLM technology and neural network technology can do really well, like I said, is take on isomorphisms of complex systems and I think that if we applaud them, maybe in a more developed, sophisticated form, to production change and distribution chains and how they're related to environmental impacts and stuff like that, it would be able to model those systems well enough to simulate them coarsely at this point, but coarsely enough to where we could steer our economy more than plan it out in five-year chunks but actively steer it, which is the goal of cybernetic planning. Is that steering of active systems, which is not what the Soviets had? I don't think that's what their goal was necessarily.
Speaker 1:No, I think that's kind of an understatement. One thing I would say about, about like even a lot of what I consider like more like Anarcho-Communist theories, about this like a participatory economy which, right, like we do plans and you have a year plan and a few back, like, I'm still like that's too brittle. Like all it takes is one rather for a year planning to go to shit. We see that with capitalism, which is more flexible. What do you like? So this idea that we're just gonna have like one giant participatory meeting and then we're gonna figure out how to do all this through negotiations and then we have stuck with it for a year, is just that doesn't make sense to me. For one, I'm just not that into that many like even though I'm pro feedback of the idea of us having that many meetings forever is enough of a time for me to be turned off by it. One thing I will say, though one of the things with Cybersyn that I have pushback on is, like part of the Chilean effectiveness was actually just as much the Codones and instigating workers into that as it was Cybersyn. I mean, cybersyn was never fully used, but it's always interesting to me what we could have learned if we looked at like what it would mean to run that kind of information system with the Codones and empowered workers, who are also very much trained in a kind of cybernetic model of like roles, not positions, so like anyone can step into the role if they have the skills to do it, and we wanna make sure as many people have the skills as possible so we can have fairly flexible teams and they're planning for like a whole district in an area. These are kinds of things that we do need to think about. That's why I've been interested in like localism and translocalism and stuff like that, because I'm like well, what we don't want is what we should have learned from like sectional bargaining, and one hand, sectional bargaining is great because it gives workers a lot of power. On the other hand, it washes out workers regional concerns.
Speaker 1:If we know this from Italy, there was sectional bargaining going all the way to the fascist union system and the syndicate system under the fascist. That was actually maintained after the fascists were gotten rid of and it was supported by the Italian Communist Party, but it treated everything as a wash and that led to operaismo and autonomia, actually even emerging out of Marxist Leninism. People think it's this far left thing and it kind of is, but it actually emerged from people who we associate with Stalinists today, and because they were realizing there was problems with sectional bargaining and all the sectional rules being just drawn up for everybody that wasn't resilient enough on the kind of things that we're talking about, like it couldn't take in what the specific needs were of, say, female workers and say Milan, not Milan, that's in Spain, and like some specific part of Italy and like shoemaking, for example. And so that's when they started doing workers inquiry and really pushing that kind of stuff. Workers' inquiry is a lot easier to do with these kinds of technologies because we'd be like, okay, give us the inputs and this will automate so much of this.
Speaker 1:And because this is automating a lot of the skills capture areas and giving it to you very easily, which that means anyone with moderate intelligence can now learn it, particularly because it's not hidden in jargon or in cartel off other systems, then it really is quasi-meritocratic and you're gonna want as many people in your team to have the skills, because if somebody's sick that day, if a couple people have the skills, it's fine.
Speaker 1:On someone sick, you have cascade failure, right Like which as an individual worker in a capitalist system you might want because you want to be indispensable, but as somebody who just wants their society to work, you don't want that right Like. And so I do think these kinds of planning apparatuses and taking a kind of like, okay, we realize that LLMs under capitalism are probably gonna be a nightmare. Let's think about what we could use them for if they're not, or even dare I say it as a Marxist let's think about what we can prefigure to use them for now, while they're still, you know, freely abated. Be some before some fear monger makes this only providence of, like, large corporations who can keep everything away from us.
Speaker 2:Luckily that's been subverted a bit.
Speaker 2:Mata released one of their models called Lama and they got leaked and you can pretty easily go and download and run an LLM on your personal computer. It'll be slow but it's possible. So I mean, these technologies are gonna be around, they can try to regulate them, but the cats sort of out of the bag. The Muzhwazi may get better ones, but we'll always have something. But like, I'm really a fan of this concept of looking at all of the control systems that are in place in our society to manage everything going on, all of the new technologies, and sort of detangling them from their bourgeois nature, their bourgeois roots, and seeing what they may be useful for. I think Marx was into that too, because he brought up things like the stock market, which was crazy to me in the beginning whenever I read Capital Volume Three and he suggested that the stock market may be useful in planning. But now, looking at it cybernetically, the stock market is an intelligence that I think is greater than LLM technology. Right now it's not being used in its potential.
Speaker 2:Stafford Beer has this little quote the nature of a system is what it does basically, and the nature of our system is constraining working class freedom to produce profits, of extracting surplus value, and so everything that comes from that is gonna be in line with that primary goal.
Speaker 2:But it doesn't have to be. And there is a way to look at these things cybernetically and sort of detangle that web of mess of private ownership and the exchange of goods from those very intelligent systems that are managing that flow right now. And I think that a lot of the technology that's being produced right now is doing that for us. In a way, it's working against what Capital may find useful at this moment because it's given us the tools to plan our economy in ways that were impossible before. I think that these are extremely powerful planning tools If we can look at them that way and if we explore those possibilities and if socialists get serious about understanding them. I think that the capitalists are so focused on these LLM technologies right now is more than enough reason that communists should be exploring why they may be potential to our project, and a lot of it is the control problem. It's model building in control is what these technologies are great at.
Speaker 1:I think you're absolutely right. I think there's a tendency to have a moralistic view of capitalism and, look, I have a moral critique of capitalism too, but that's not really what I'm concerned with here that these technologies could be further developed and utilized in ways that would be empowering to huge swaths of people If we didn't buy their hype and really thought about what they were being used for. I think the stock market thing is an interesting one, like how much capital is managed through that? Another thing that we can always point out yeah, people plan all the time. We know that capital is planning.
Speaker 1:Do we need to convert everything into it? Like there's this weird fetish that we have in Marxism to try to like well, we're gonna convert everything into dollars and I'm like, yeah, but there's plenty of knowledge and planning. We don't convert into dollars now. So value yes, it's the primary thing we do, but we shouldn't be trying to replicate that. And what I mean by this is like we also don't need to convert everything into, like labor tokens or whatever. Like, sure, that might be helpful for certain things, but it's not gonna be. Like I don't need to figure out like the unit of labor for every single thing we do, and that would not necessarily be a useful way to plan everything either. But there are ways in which these systems that already exist actually do account for a lot of these things and we just don't deal with them. We don't use them, we don't wanna use them because of their origins. And while I don't think technology is neutral, I want people to understand that, like technology is teleologically motivated, but it's teleologically motivated by the people implementing it, right.
Speaker 1:Like an LLM ran by workers would not do what an LLM run by Google's gonna do. An LLM ran the stock market that we're using, the technologies that we use behind the top market, the kind of aggregation that we see in the stock market could be used for all kinds of other than what we use it for. We should not be turning away from this because of all kinds of different things Turning away from this because it's capitalist, turning away from this because it's difficult. I also think there's been a lot of critical theory and psychoanalytic critiques and stuff like cybernetics that they freak out about the name. Now I admit I've read a lot of cybernetics sociology in the last couple of months, so that stuff is bad and I also think we do have to deal with this turn in cybernetics that I think kind of happened in the 1980s, when it really gets limited to computer computing technology and all the other stuff is dropped just almost entirely.
Speaker 2:Well, there is a good, a new sort of movement and psychology that's integrating a lot of the old cybernetic insights along with sort of insights that were gained from AI development.
Speaker 2:It's called predictive processing and I think it's pretty interesting. It's fairly convincing. What it argues is I'm waiting for the guy to cite Stafford Beer, because it kind of seems like he's just implementing the viable system model into psychology a bit more. But so basically they're saying the brain is a prediction machine that's running simulations that have been fine tuned by our experiences. You first have just the basic human brain, which is overloaded with connections and through our experiences it fine tunes its model of the world, its model of its own body, the proprioception, and basically what our senses do is send error signals into that simulation that's running and so like, if you're seeing something, what you're seeing is your brain hallucinating, but the simulation believes it should be seeing, and the error signals are what gives the updates to the system and what gives attention. That allows you to assign attention to things.
Speaker 2:And I'm not totally convinced, but it's very interesting. And the concept of the model in the head is very interesting to me as well. I think that is extremely powerful and for my own experiences it's how I've sort of rearranged the model of the world I've held in my head. I've seen profound impacts on how I see the world as a result of that and how I see people and how I can have compassion with them, and I sort of wanna believe that this predictive processing thing is true, because I just out like it. But I also think that it's really good and I think that there's some good evidence for it as well.
Speaker 1:This is interesting. The advantage is you can computers. If you mention something, I can look it up. I know a little bit about predictive processing but I didn't actually think the link at the cyber.
Speaker 2:Mm-hmm the Gagi. He quotes Shannon a bit, but Shannon's quoted pretty much everywhere, not just in cybernetics, but he also talks about sort of the cybernetic revolution in the 50s and its limitations and Well, this actually would actually combine like a servomechanic and a cognitivist viewpoint, right?
Speaker 1:Yeah, that's what I was talking about. Like to say, like okay, we got all these random structures in the brain that are basically spandrels, they just evolved, and we do know from neurology that we're opening up neuro pathways and closing them off all the time, particularly in certain periods of our life like childhood and early adulthood. And if it's working in this kind of feedback, we're like basically like okay, we have all this random architecture we're dealing with, but we can turn it on and off and we're getting feedback and simulating, processing this, and then like so we have this simulation that we're running, we have this theory of the world, this theory of mine that we're running and we refine it by feedback and that's literally also leading the stuff like neuroplasticity and, yeah, that actually is helpful. That's a much more robust theory of what's going on than like even the pure neurology ones, because some of them are like well, you know, your brain lines up afterwards. So either A there's no such thing as agency at all and we're uttered the term as going back to the big bang or B neurology is shit, because we can't explain anything by it and this probably would. Actually, I don't know that it was a help, but my inclination just thinking about it right now, but go now, this might actually explain some things. Like we might actually be seeing why processing is happening the way it's happening and ways that we will experience differently than what we do, and why these systems might not be related, why agency may still be somewhat real. I mean, I don't mean counter causal agency, that's absurd but like people might actually be running enough models, that choice actually is something that you're actually encoding and naturally, et cetera. So like, yeah, that's interesting.
Speaker 1:I mean, you know about me, mark, you know that I've been ranting about this for years. But one thing I will say that I was always frustrated with Marks about is Marks because he's kind of dealing with the Hegeon context, because he's dealing with science in a very early form. I mean he's reading in anthropology, he's reading in soil. I mean I've been fascinated recently finding out from mega two stuff, all the stuff we've discovered Marks was reading. He had weird interest. But did I be like?
Speaker 1:Well, marxism seems to imply a theory of mind, but class consciousness is kind of a way to hand wave that away. Alienation is getting to something like we're alienated from ourselves, we're alienated from each other's workers, we're alienated from products of our labor. He focuses on the products of our labor because that's the part of capitalism is super important. But he talks about other things, like you talked about, for example, or the good-backed point you made earlier, mark.
Speaker 1:You said that right now we have this we see massive amounts of social alienation right now and skills hoarding and stuff like that, and there's all these competition and so, while there's more proletarian now than there's ever been, we don't think to divide ourselves up that way for sectional, regional, all kinds of actually all kinds of reasons, but also because we experience like status class amongst us as proletarians. Interestingly, when you take this theory of alienation that Marx is playing with in early Marx, there's an explanation for that. Like, the competition of workers against workers is an alienation of workers between what would normally be a cooperative thing, do something as a competitive because of the stresses of labor, and so status mongering within that would also become highly incentivized and so we wouldn't experience class as proletarians versus capital. I mean, most of us do not actually meet the capitalist in our lives unless they're petty bourgeoisie, like when I worked at GEICO it was very hard for me to meet the owners of GEICO. I met managers, even met high-up managers, but I didn't meet the owners Like I met.
Speaker 1:I kind of I met Warren Buffett once, but he's not the only owner. Like, structuralized capital removes that from you to such an extent as it's not something you experience. And then all these skill capture games and everything, and I think this is why there's all this anger at the PMC whatever right, it's part of it's. Liberal, educated people are annoying. You've already established that Part of it is. There are these skill capture games we see in life that like there's some random arbitrary oligarch fucking up things in our lives that we can't really say they're a capitalist. They probably are college educated, although they might not be, depending on the kind of job we have, and we're in competition with them.
Speaker 1:Well, that's a status game that's incurred from the kind of alienation of work right, and the universalization of the proletarian would make that more the case, particularly if the capitalists are so far removed from us that they don't even really care about the data and the operations of the bourgeoisie anymore. I mean, one of the things I've kind of thought about a lot is capitalism is so abstractified now and capitalists have gotten so into like rents and rent commodities. Not that we've developed into neo feudalism or anything stupid like that, but like rentier. Capitalism really has meant that a lot of capitalists aren't even that interested in like running their businesses anymore. They're outsourcing that. They're just kind of backing and hoping for ways to collect capital, even if the capital they're collecting is not valorizable unless effectively fictitious right.
Speaker 2:That's been a pretty long standing process. Marx talked about how the capitalists handed over control of the factories to the managers. That wasn't a skill they had to have anymore. All they had to do was collect profits, and that was back in the 1860s.
Speaker 1:Right.
Speaker 2:They usually had to actually manage profits.
Speaker 1:so, and now Right.
Speaker 2:Well over this time they've developed such intricate meta systems to do all of this stuff in this necessary because the division of labor, these supply chains, have grown so complex that we have to figure out ways to manage all of these, all of this variety, as the cyber net assist would say. And the way you do that is through meta systemic control, so systems that are built over and above the systems that can constrain them or direct them. And, you know, going back, I think that these meta systems that they've built are extremely useful If we learn how to use them, because they are. They're built to control product flows, to control distribution of labor, to control. I mean, these are things that we're going to have to think about in planning and the. These systems didn't exist in Mark's time and they allow that, that pulling back of the capitalist. So to the degree they are, yeah, yeah, I think that's a fair point.
Speaker 1:I mean, I would say that. So there's a kind of Michael Hudson tendency to like posit industrial capital versus financial capital and that's like a socialist country doing industrial capitalism like China is but smart, valorist. And you know, and my point to them is always like yeah, but why did if? If financial capital is always parasitic which actually I don't think it was doing industrial capitalism ascendancy and it definitely existed, then Mark's talks about it Then then why did it turn into this parasitic form of financial capital? Just because people got greedy? That's not really viable. Like, like, so we have to look at the systems. So people get all like all industrially and like, oh, if we just had industrial capitalism, we had our workers movement back, etc. I'm like, but you might just restart this process.
Speaker 2:Right Like, but I think that they're ignoring the reason that that happened in the first place.
Speaker 1:Right.
Speaker 2:And I think a lot of it is is profitability is falling. If, if, if your interest rates are higher than profitability, you know, if you, if a capitalist, can expect to get more back from dividends on speculative investments than they can on industrial production, then it only makes business sense to invest it in that direction.
Speaker 1:Right. However, what this does create is a problem eventually where there's nobody in the system, because we got sourced it to meta systems etc. That can actually, like people ask right now. I was reading this palladium article recently that blamed it on diversity. It was very kind of race realist actually, but I was like no, the reason why Maritor-Craxie is broken down isn't because just diversity quotas. In fact, if you look at a lot of this, we have more nepotism than we've had in the past too. It's not just diversity quotas. The reason why a lot of this meritocracy is broken down is, though, is like there's no one who can actually do it, like like we, we.
Speaker 1:The system is so complicated now, and the meta systems are kind of been running on their own, that no individual part of the system can actually reign any part of it in. They don't understand all the feedback loops. I don't know anyone who does Right, and so this ironically leads to highly specialized, highly educated people, workforce, even bourgeoisie, who are subsequently, however, less competent because it's too much to do Right and, and so when people are like oh well, you know, profit, weight decline don't exist, I'm like well, there's countervailing tendencies. We're in right now, actually, but we got rid of a lot of fictitious capital, and stuff looks to be flowing again. Also, a lot of people died, yeah, like. All kinds of stuff happened during COVID. In fact, I've pointed out that I think COVID actually ended up from a mixture of social democracy, getting rid of excess stocks, realigning world capital, etc. Actually avoided a recession, even though it hit one. Hit one in a way that, like forced us to reorganize, in a way that, like probably starved off the recession for a little while, and it might be coming in the next couple of years, I don't know.
Speaker 1:The one thing we can see, though, is even a country like China is experiencing unemployment in the business cycle now, so we know that it still exists. Sorry, techno, feudalist people. If we're in a feudal economy, there's no business cycle. Sorry, it just wouldn't make any sense. So I don't know business, but this is something that I think we're going to have to look at, when people think we can just turn back the clock on our industrial policy, or we, you know, basically let's create industrial capital again so we can have the workers movement that we had in the 20th century. That's not going to happen.
Speaker 1:It's not like the US is unproductive now, right, like, and it's not like I see what's going to restore those profit rates for a huge period of time. I don't see it, in fact, right now I really I don't know how you know, which is not to say that something you know, capitalism always surprises me, man, but like it's a dastardly motherfucker, but it's going to get out of these current like trends. Like I don't know where the viable capital is. It seems like no one can do anything about any of the systems. There's too much prior investment. Even something as simple as student loans seems impossible for this or something that we know.
Speaker 1:One of the things I tell people like why is it so hard for capitalists to institute socialized medicine or even socialized insurance in America, when it was even good for capitalists in the states that they did it to in Europe and Canada and Asia, et cetera? And I'm like, well, because we have to destroy all the shit we've invested in the private system, like, and you have to figure out what to do with that, whereas if you instituted this earlier, you would never have that problem, right, so like, the longer these systems go on, the harder they are to change, the harder reform becomes, et cetera. And you do kind of you end up in a situation where we are going to have to break parts of it down, even for capitalism survival, and why would we want to? If we're going to have to do that and keep capitalism up, why would we want to stop there?
Speaker 2:Why Exactly?
Speaker 1:Yeah, Like we're going to have to do it Go ahead.
Speaker 2:That old vision of workers' movements and revolution is dead, and I don't think that that's necessarily what it's going to look like I mean this we don't live in the world of a massive industrial proletariat and we probably never will again, and in the limit of how the trends are going right now, I mean, I see the industrial proletariat was famine, and industrial proletariat, in specific, was never more than 50% of the population anywhere Like we.
Speaker 1:it never hit the point of being the majority of the population, which which that is something that Marxist freaked out about in the early 20th century because they were like wait, this isn't this. We thought that we'd be and we're not. Like, and you know, at the time there are a lot of debates like are are people in the service sector lumping right and stuff like that. And I do think there is like a lumpenization of lots of the working class, particularly the industrial working class, because because of de skilling, because of precarity, etc. Totally think that's real. But but the idea that like, oh well, we shouldn't deal with them because the service sector is lumpen, when, like now, even in the developing world, it's an increasingly huge part of the economy, because it's not automatable, right Like, or it's hard to, it's very hard to automate, that's a, that's the reality we have to deal with, right.
Speaker 2:Yeah.
Speaker 1:And so I just don't think. I guess my point is not only can we not turn it back to what we think the 1940s and 50s were, we also have to be honest that the 1940s and 50s weren't what we thought they were in the first place, or we would never be in this scenario now.
Speaker 2:Right, and also I think that it's sort of tangential to the actual problem. So the identification of the proletarian is important for Marxism, but I think that it's important for Marxism because it's just a technical fact that there's really nowhere else to go below it. Once you, once you've reached proletarian you are, you are alienated from society. Unless you can sell your labor to a money owner, you do not have access to develop your individuality, much less to eat or shelter yourself as a proletarian, unless you convince someone to buy your later power as a commodity, and that commodity is mediated through a market price that's ultimately mediated through its value, which is its basic reproduction, and so I mean that severely limits us as workers. And so I think that what the real point for Marx is is that we're at the point to where we have no other incentive but to take control of social reproduction.
Speaker 2:We don't have any incentive of a ownership to create a new form of social reproduction, that that sequesters surplus value or surplus labor as private property.
Speaker 2:It's in our interest to control it socially, and what Marx also saw is that capitalism is producing the needs to control social reproduction in a social manner. So I think that if you look at it that way, who's proletarian, who's not proletarian, it doesn't matter. What matters is who is working on the project of how do we control social reproduction in a way that is democratic and in a way that maintains economic and social viability, in a way that that keeps working class people with a healthy model of their world and their society in their head, so that they don't become what we call lumpen, which I think is just people with a model of the world that reflects the highly competitive and vicious and violent existence. And so of course we're going to interact with the rest of the world in that way, but that's just the result of the fact that they can only mediate that model, that creation of individuality, through what they can purchase, and they can't purchase anything because they're broke. So what my point is is Marx wants us to take control of social reproduction.
Speaker 2:Capitalism has given us the tools to understand it scientifically, and I think when Marx says scientifically, what he means is in a systems theory point of view, look at inputs and outputs, look at the model he was creating, a model of capitalism and how do we redesign what we need to redesign and how do we recreate a world to where we can control what happens to us in the future? We can predict it, we can make it happen. That's what his goal was, and the language of proletarian in class is what he had to describe that, yeah, I think you're absolutely correct there.
Speaker 1:I also think the idea of, okay, who's proletarian? We figure out their productiveness and, yes, I do admit exploitation as part of proletariness, but some of that guys was figuring out where the tax income came from and that's what he's. There's all kinds of stuff in Marx that I'm just like. The argument about the proletarian wasn't just that they were exploited or oppressed, it was also because they had skills. Trying to limit yourself and actually exclude people who don't have certain skills is kind of crazy to me. However, I will also say there's an over valorization of what on the left in general, like academic skills, particularly in the humanities, sociology, etc. Absolutely true, completely correct. If you want to call that annoying PMC tendencies, that's fine, although I always point out that managers don't tend to be humanists. When most people are complaining about the PMC, they're complaining about HR, they're complaining about academics. They're not actually complaining about managers, which seems to be the most crucial part of that whole fucking thing. We do have to ask ourselves about all that. I also think a lot of people do damage to the analytical categories. Proletarian and capitalist is about your relationship to production and your ownership of the means of production. There are other ways of talking about class that are perfectly comprehensible and they're real. We can talk about upper-lower middle class. We can talk about precarious, non-procure, etc. These are real ways. I know what they're describing and I can understand them. They're not the same thing.
Speaker 1:I think there's this Marxist tendency to go well, the only class analysis we have is this one, but if it doesn't work on things we want it to work for, we got to do all these kinds of things to make it. Let's talk about this other theory, let's add this theory, let's pull this out. Let's only focus on productive workers, let's only focus on industrially productive workers, which is different than productive workers, by the way, etc. I just want to go like, like. It makes the one clear category we have in Marxism. Marx talks about like what? 15 or 16 from classes, I'll admit, though, except for, except for, bourgeois and proletariat, most of them are, kind of it, not as clearly or analytically strictly defined, right, like I have. Still, you know, we talk about love, fetishization. I know what that is. It's very real.
Speaker 1:But if you ask me like, and it's kind of a useful social category to talk about people who are for the world. But if you ask me to tell me what its class characteristics are, I'm like, I don't know actually. Like, because there's stuff in like well, the lump in exploit the workers through through, you know, question, I'm like, but some of the stuff Marx lists in that group are not actually exploiting the workers, it's just socially unpleasant shit, like. So I don't, I don't really see what you. You know their petite bourgeois. Another one where it's like okay, like I get it, it's useful, but also like it's not strictly defined. In the same way it's probably more clear than lump in. But reserve army of labor, that's another one that I'm like up, like that that could be anybody. Honestly, this is anyone who could be in the position of a worker. So I say this because I don't wanna do damage to the analytic from Marx that we have. That's clear. Like, and I admit that like no, marx doesn't ever spell out. Like we've talked about this.
Speaker 1:You've read the volume three, I've read it to you. It Marx, you know when he gets to the strict definition of worker he cuts off before he gives it to us, right, like it's not finished in volume three, but we got. We got enough of the description to kind of know what we're dealing with here and it's people who are dependent on the general, on the general wage fund and do not own their own means of production. And Marx does understand for people who don't get it that people are a multiple. I mean his example as a teacher. People are often in multiple class positions depending on what they're doing at what time, like his example about the teacher is like a teacher when working for a private school is productive laborer who's subsisting off this, and then when they're working on, you know, for the public, they're actually living off the surplus. So you can't call them productive but they're not, you know. But the implications of what they're doing is socially necessary in both cases, when it's productive, to capital accumulation, one is not.
Speaker 2:I noticed that, whatever Marx uses that language of productive and unproductive, he's sort of talking about, you know, the relation to the generation of surplus value, right, which makes me wonder what Marxists really care about. That, for, you know, is it's not really our concern whether our labor is productive of surplus value or not. That's something that Marx was sort of discussing, because that's what the people he was reading was discussing.
Speaker 2:You know he was sort of trying to synthesize their different viewpoints or models of capitalism through, you know, sniff and Ricardo the physiocrats, and take what they were talking about and figure out you know how to turn it into this, into capital, and so I mean, I think a lot of that, a lot of that is just a little bit of misdirection and we're falling for it.
Speaker 1:I just yeah, I find that we're falling for it. I also find that we're falling for it at a time when, even though, like, broadly speaking, culturally and maybe even more economically left wing than in the past, the far left and the social Democrats have lost a lot of cultural credibility and a lot of people are going into this stuff, but more people are probably just becoming basically either liberals or conservatives, because they're fed up and we don't monitor that, because they're not talking to us, they're just disappearing, right. But to me, all of this is a sign of like. Well, this is a sign of political and social weakness and that's why we're attracted to. This is like, in some ways, we're trying to justify our alienation by one blaming it on a class PMC.
Speaker 1:And again, I know I have a lot of people who think I'm being soft on them. I do think I'm a big believer in like just met different minds. I think university does something that makes it harder for them to talk to people. You talked about this too, and sometimes we need Like it limits your social world as much as it expands it. But to me that's not a class problem, that's a socialization problem.
Speaker 2:Right, and those people still hold vital knowledge of meta systems that have deemed to be socially necessary by capitalism. By virtue of them existing, being funded, they have some sort of social necessity to them in a capitalist society. And the people who are in those PMC positions hold social knowledge that we've developed over hundreds of thousands of years of human existence, and I think that that group of people is gonna be extremely important in some hypothetical moment to where we can start to consciously manage our social reproduction. Not every decision can be made by the general idea of what a working class person is gonna be. There's gonna have to be people that engineer the new society just as much as there's gonna be people that participate in the discussion of how to engineer it. I guess Absolutely.
Speaker 1:Well, on that point, I guess we've kind of been talking broadly. We talked a little bit on LLMs and the art of the limitations, and we've talked a lot about the intersection of cyber Marxism and we've gotten We've ended on pure Marxology. Maybe it's a good place to end. Is there anything you wanna share as you go?
Speaker 2:No, I don't believe. So I mean I appreciate you having me on and letting me talk about the stuff that I'm interested in.
Speaker 1:Me too. It's always nice to have you on. I am glad that you came on and hope to have you again sometime. So thank you so much for coming on. I remember when I brought you on. Just be like it must have been almost five years ago now when I brought you on and be like hey, read Capital Volume Three. Please, before you start talking about political economy, don't just read the chapters you like. Don't stop at the intro class. It's real stuff. Some stuff in those other stuff is really important. Even if you don't agree with it, even if you think parts of it are wrong, it's really important.
Speaker 2:Yeah, I think about that interview a lot and it was in the beginning of my sort of self-education and I cringe to think about some of the things I said, which I'm sure I'll do the same thing with this one too. But no, I mean, I really appreciate it the whole time. I don't know if people know, but a long time ago I found Doug Lane and you were talking to him and you sort of inspired me to start read Capital and all that. So I appreciate that. And yeah, you've got it.
Speaker 1:Yeah, I'm glad, I'm glad. All right, well, I know. Have a good evening.
Speaker 2:Yep absolutely.
