Future of Intelligence in the Age of Intellectual Scarcity (Session 2)

online hello and welcome to the second session of future of intelligence in the age of intellectual scarcity I'll now hand the mic to Reza thank you very much hello everyone thank you so much for showing up. So first we are going to continue with the criticism of David Rodin's Disconnection thesis. We'll have, we'll open it up to questions. After that I will talk a little bit about the work of Zoyal Malek, particularly his recent work on futurity, contra-contemporaneity

and the risk order to somehow show that some of the main motifs shared among these both thinkers actually do converge. They do converge in a very specific way, in the way that I was talking about in the previous session in the sense that they both try to underline a concept or representation of futurity which is fundamentally fundamentally in a in a

complete epistemological sense but nevertheless radically different than ours now of course both of these thinkers are not you know theologians in the sense that they are not like negative theologies of this scholastic era they say that you know this future is completely unknown and cannot be ever interpreted no they do in fact like the disconnection thesis they leave a room for a certain amount of interpretability or connection. Otherwise, if they don't do that, then that talk becomes just yet another version of negative theology. You might, as I said,

would be an arbitrary choice. Why are we talking about prosumen? We could as well talk about the omnipotent God of the scholastic era. But then, my point is that to show that that room that they have reserved for certain amount of interpretability, computational analysis or epistemological traction in principle, is in fact voided and cancelled out by the very arguments that they make. Ergo, the conclusion of the thesis in fact does relapse back into negative theology,

hence scholasticism. As simple as that. Any question at this point? Valentin, do you want to say something, heckling stuff, challenge? I can't wait to hear you destroy them. I'm pleased, dear friends, the reason that I pick some of you other than the others is precisely because I know the troubles that you can make. Whereas some of you, I'm still

trying to figure out what your potentials for trouble making is after a few sessions I will actually come after you too so don't laugh so Remember we talked about the last thing that we talked about, David Roden's Disconnection thesis. We talked about one of the aspects of Roden's speculative thesis with regard

heuristic biases particularly a very serious brand of heuristic biases which are called judgment under uncertainty judgment under uncertainty you see and we talked about that the kind of probabilistic or heuristic probabilistic judgments that you make under uncertainty usually tends to be disproportionately they are not by any means robustly probabilistic they are actually far more what you might call to be similar to those very probabilistic

pre-critical judgments that people at the time of medieval ages were making about the widgets. You see that's why, sorry, you know that my memory is really really short, but I think that I said something in the previous session that One of the great ways to think about a future AI is not just about refining the probabilistic mechanisms or the probabilistic accounts and systems, the logic of probability, for example,

in making neural networks or developing a statistical account of conceptual activities. You see, I am not one of those people who think that conceptual activities cannot be statistically understood. I do think that they can be understood statistically. In that sense, I am against Chomsky. Chomsky actually believes that semantics cannot be statistically modeled. I think that recent advances in a statistical analysis, probability and computation, show us that every qualitative scale of minding activities can do in fact be

a statistically modeled but not under one single unique rubric of probability logic essentially different qualitative levels of faculties of minding faculties perception conceptualization and hypothesization so on and so forth they require their own statistical models and probably their own paddling a probability logic and that's why probability the concept the modern concept of probability is of utmost importance to think about a future AGI or AI for that matter. It's just that right now we are in the same kind of impasse that I witnessed when I was young

about symbolic logic or syntactical AI in the sense that AI research field are just like bad philosophers. And there are so many bad philosophers out there. They get super overexcited. So they have a new single method, which is so powerful, like the transcendental paradigm in philosophy. And they think that they can talk about every single thing, model every single paradigm of faculties by way of a single paradigm, by way of a restricted set of methods. You can see this back then, it was symbolic syntax,

our Chomsky and his legacy, and now you can see it in terms of neural networking. And what is usually called among probabilistic mathematician or statisticians, what is called greedy Bayesianism, greedy Bayesianism or immodest Bayesianism. Immodest precisely because they think that the Bayesian paradigm of probability is the ultimate paradigm of probability. But the thing is that looking into the history of probabilistic logic, we see that the concept of probability, the take that we are talking about the statistical analysis changes,

has changed quite drastically, even in the past, I would say, three decades, if not for the last 50 or 60 years. One of the first people who actually addresses this problem is Rodolphe Carnap in his Magnum Opus, which is his late work, more mature work. It is called The Foundations of Logical Probability. It's a superb book, superb book. Any person who actually wants into the nitty-gritty of

AI development, Bayesianism, so on and so forth, should definitely read this book precisely because it's the first book ever, ever, in which the logic of probability is proposed as the logic sufficient for constructing something called a universal learning machine. What is a universal learning machine? It is essentially a machine given a set of arbitrary data model is capable of singling out among all these arbitrary sets

of data the most optimal model of data namely a pattern okay this is what is called a universal learning machine and there are of course different you know updates on this machine to the point that today they are not it's not called universal learning machine it's called an optimal learning machine optimal in the sense that you have pool of data models and you have also a pool of inductive formal inferences and through such means you can single out among an arbitrary set of data an optimal pattern like for example the pattern of something being

red that's for example the Sun or the earth revolves around the Sun so on and so forth, these kinds of stuff, you know. So Carnap wrote this book. When, you know, that when Carnap, I mean, to give you a little bit of a historical background, Carnap began with a very positivistic account of philosophy. He nevertheless abandoned that positivistic account, the so-called logical empiricism. After his friend was murdered on the footstep

of the same university he was teaching, he moved to the US. During his time in America, began to teach a number of fundamental courses. Yes, Shalik, yes, Maurice Shalik was his friend, who was killed by a fanatic Nazi, yes. So during the time that Karlak was in the US, he gave a series of seminars about formal logic, about probability, about the logical structure of theoretical, scientific theoretical systems.

During some of his seminars, one of the greatest mathematicians and computer scientists of all time attended his seminars. That person was Raymond Solominov. Ray Solominov. Ray Solominov, if those of you who are coming from the computer background know that Solominov-Kolmogorov thesis thesis is the utmost importance in information theory and computer science. So Solomonev actually adapted this idea of Carnap and developed what is today called

Solomonev induction or formal account of induction. It is called formal account of induction or computational account of induction precisely because it's different from the human account of induction. It is not based primarily on observation. So you see, in Hume problem of induction, it is something like that. a well-watered grass time t1 looks green okay time t2 well-watered grass looks green t3 looks green

So, after a hypothetical timestamps of such observations, we can, according to this rudimentary account of induction, we can say that after hypothetical time, TN, any well-watered grass in such and such conditions will be green. okay this is an observation so-called frequentist accounts of induction frequent is precisely because it's based on the frequency of your observations okay now this is the first account of probability something that

today is no longer used in its purest form. Yes, parts of it are used, but not in that kind of purest form. Now, in his famous revolutionary essay called Two Concepts of Probability, If I'm not mistaken, the first version of it is written around late 1940s by Carnap. As a counterpart to this frequentist or observation-based account of induction, he proposes a new paradigm of probability and induction, which is called the logical induction.

In the sense that the observation of well watered green grass is no longer taken to be a reference, the primary reference. It is the logical connections, couched in terms of the probabilistic logic, which is now the primary reference of your induction. So, Solomanov used this logic of probability rather than the observation itself and turned this into a computational formulation, which is called the Solomanov computational account

or formal account of induction. Without going too much length, what I was trying to say is that in this book, The Foundations of Logical Probability, or Probability Logic, Karnap, in the first 20 pages, you don't even need to read the whole book, just read like the first 30 pages. quite clear, you don't need any sort of technical knowledge. He talks about a very, very serious problem which is actually the problem of philosophy. He talks about a term called explication.

Does anyone know what explication is in the Carnapian sense? Anyone? Anything? It's making a concept more effect. Can you repeat, Valentin? Sorry, I couldn't hear you. I think it's about making a concept or replacing a concept more exactly. Yes, yes, something like that. Yes, yes, yes, yes, absolutely. So, you see, every concept that we use in an ordinary language, like probability, possibility, I don't know, the concept of life, the concept of soul, the concept of man, the concept of life, whatever, anything that you can imagine, okay?

These are concepts which are cluttered, they are fuzzy, they are vague, they are bloatwares. In the very sense that when you get a Dell computer it comes with a bloatware. You know, all those apps have so many crap bunched into the main app. These are bloatwares, metaphysical bloatwares. So when we are asking say, what is life? What the hell does this question even mean? What kind of life are you talking about? You see, unfortunately we philosophers have the very bad habit that when we look at a concept, we always look at the bloatware concept, the fuzzy and vague concept.

concept. And we think that the polysamy of that concept is actually, you know, a source of richness that we can do some sort of poetic crap around this concept. But that's just bad philosophy. That's just bad philosophy. Unfortunately this is not just a curse of continental philosophers. Analytic philosophers do that too. So this is what I just said is not by don't say to it Reza that Reza said that continental philosophy is just some sort of poetic mononism. No that's not what I said. All I am saying is that

that if philosophy begins with concepts, concept of life, concept of happiness, concept of probability, concept of this, concept of that, so on and so forth. insofar as these concepts are couched in an ordinary language and we all know that ordinary language is extremely vague and fuzzy in the sense of when I'm talking you know like at the beginning of a conversation like me and Alejandro talk about the question of life at the beginning of our conversation

we actually don't know what we are talking about unless and until we contextualize our conversation and triangulate a concept of life which is more exact less vague, less fuzzy we know that this life is not life of the cosmos. It's not the life as an existential thing. It's just simply we mean biology. Biology. Simply that we are both made meat of meat and flesh. Okay? But even that is fuzzy. So this is the whole point of the Carnap

at the beginning of the logic of probability. One of the reasons that AI, and this is not something the Carnap says, But after I have read Karnak for quite a great deal, I think that this can actually be a fundamental lesson for artificial intelligence, or for that matter, the entire canon of philosophy. The task is to explicate concepts. So, explication means to replace a concept which is cluttered, which is bloatware, which is vague and hard to actually even agree what we mean by life at that point,

to replace this with a range, a progressive range of ever more refined concepts, Like life one, life two, life three, life four, each of these stand for a specific aspect of life which we had already bunched them together as the word life. the same thing about probability. When we are talking about probability or a statistical analysis, we are talking about so many different kinds of probability. The whole point is that we have to refine these concepts. Once we refine these concepts,

move from probability, all in capital, the bloatware to probability one, probability two, to probability n, ever more refined concepts. Then at that point we cannot only develop new paradigms of probability, but also appropriate methods of probabilistic logic and analysis for each and every single of these sets of probabilities. And here, that's why I compared neural networking at this point to symbolic logic. The only difference between the two is that symbolic logic tries to see intelligence from a top-down view,

you know, that basically symbols are basically the controlling mechanisms and so on and so forth. and the abstraction is top-down. In neural networks we have a bottom-up abstraction in the sense that we have a single paradigm or not one single or a few handful of paradigms of probability and we try to overextend their paradigms and their corresponding methods to top faculties like conceptualization, hypothesization and so on and so forth. And of course this is not going to work. These are just bad pre-critical, pre-scientific use of

methods and concepts. To do that we have to develop different patterns of probabilities different methods appropriate to them and apply these appropriate paradigms and methods of probability to their corresponding appropriate appropriate target system for example you cannot i really genuinely want if any of you who are coming from a computer science background I genuinely want to tell me that conceptual activities, particularly conceptual material inferences, can at this point be coherently modeled and replicated by neural network paradigm.

Can anyone object to me at this point, please? Hi Reza, I have a question regarding something that you said just recently. Sorry for the background noise. But how do you identify the basic unit for to start to refine, to group these sets of meanings? Yes, yes, yes. Well, this is why I actually really, really, really, I mean, I will try to extract those 20 pages from the first part of Carnap's book. So Carnap already knows

about this. Essentially whatever we talk about, about anything, is always going to be talked about from within the ordinary language that we have and through which we perceive the world. This is what Wilfred Sellars calls the manifest image on the world. It's manifest precisely because it might not be real but that appears to us thus and so, right? So the resources of our ordinary language is our primary or initial point. Now of course here a problem arises so and this is where the role of science as a social

enterprise also come to the picture. So you see, I say cloud, okay? We are talking about cloud. And we start a conversation about cloud, okay? And we know that we are just talking, this dialogue from within the very ordinary language that we do share in fact. Right? But then, even at that level, we just don't know that your cloud is my cloud. As if we are talking, even though we are using the same word, we might actually mean differently

by the cloud. by the cloud you might actually think of the cloud as the cloud in the internet I might actually think of the cloud as some sort of you know vaporware in the sky right okay so this is the this is where science becomes a you know a central point you see this is what community of scientists becomes stronger, not stronger, but more accurate, that the very kind of dialogue that two ordinary people like you and me having at this point, essentially scientists, try to first, when they are talking even in the

ordinary language about the cloud they try to actually analyze whether they're talking about the same thing or not so there is a scientific agreement about the term which are used and after that this after this agreement upon the shared term then they begin the task of explication replacing the fuzzy term with ever more refined concepts that makes sense I'm gonna look into it thank you any question question question heckling

I can insult. Don't worry. Nothing. No one. I guess I have a small question. Yes, please. I can hear a Wittgensteinian sort of intentionality, and that we have a set of practices that we have, such as pointing, these are cultural shared practices in which we derive a shared meaning within the linguistic framework or the behavioral framework we inhabit. But once we pass out of it, those signals or those intentions, the pointing, reading a set of directions, no longer has the same valence.

So to go back to the previous question, which was very good, our starting point seems to be, or at least it appears to me, might be non-linguistic, that we're building upon a set of shared practices that aren't necessarily, we can't pin them down to the type of, to the level that Carnet might desire. You know, the more refined we get with these concepts, the farther we get away from the practices that generated the concept building in the first place. Right. I know that's right. just tie in here with random that I can't I can't fly on my own head but no yeah you are you are you are y'all God Jesus Christ you are baiting me here this is superb question fantastic question utterly I love this question okay okay here I mean please do disagree with me this is this is

something that I'm still thinking about it. So my answer to you is not going to be by any means some sort of as if I have, this is my thesis. It's just how I think at this point, at this phase of my research. You see, first of all, let's get back to the main Wittgensteinian thesis that the limit of the language or the limit is the limit of the world okay I fundamentally disagree with this thesis I think this thesis is pseudo problem from the Carnapian perspective why because to see the limit of something you need to have

gone beyond the limits. So, Wittgenstein, to go beyond the limits and see the limits of language as the limits of the world, he has only two recours. One, what Sellars calls the myth of the given, which of course I am sure Wittgenstein doesn't want to do that. Essentially, the sense data, foundationalism, in the sense that the sense gives us some sort of what you might call to be foundationalist knowledge about the world, about the state of affairs, which of course, Wittgenstein would have cringed in his grave, he would have heard that.

Second, it's a more modest one. And in fact, it is really interesting that the whole Vienna Circle, Carnap, Schellig, Reichenbach, so many of these Vienna Circle people, actually did start from a purely Wittgensteinian thesis. Essentially, Wittgenstein's slogan was like their slogan, their religious edict, okay? However, the thing that sets Carnap, late Carnap, not early Carnap, early Carnap is, you know, he's sophisticated and superb, but essentially he's just like a kind of what I would call a bad Wittgenstein.

okay so the thing is that how can so as I mentioned to you to see the limits of languages limits of the world then you should go beyond the limit in order to see the limits so what does it mean to go beyond the limit beyond the limit either in the sense of the sense that which absolutely out of the question for Wittgenstein or the question of the metalanguage. That every limit of language that we see already means that we are in a domain of a hierarchy of metalanguages or as Carnap would have said it, metalogic. Of course the reason

that he calls it metalogic is precisely because even to the end of his life Karna still have retained a little bit of that logicism of his earlier time in the sense that he thought that mathematics and all of this computation stuff can be simply thought in terms of pure formal logical means okay but nevertheless after the past four decades of advances in logic computer science mathematics and linguistics we know that there are fundamental correspondences undergirding all of these disciplines that they are that is what computational territorialism is simply

computational territorialism is like this that computation namely program as types mathematics a structure logic proof are all three manifestation of one divine essence, computational proto-foundation. And in that sense this is how we can think about the idea of metalanguage. Metalanguage wouldn't be just simply some sort of metalogic but it would be so you have

you are talking about in this domain or the territory or the purview of language. In order to see that limit, you have to go to ascend to a higher hierarchy where the computational, mathematical, and the logical inconsistencies and limitations of the previous language can be considered as limits. Otherwise, to talk about limit would be just nonsense. How can you remain a la Godel, remain in your purview of your current language and actually see its limit? It just doesn't make sense. Thank you. Yeah, it's something I'm still thinking about, but this is a good start.

Definitely, definitely, if you can, there is this blog post by Robert Harper. I know that his blog is no longer active, but I'm sure that some of his blogs have been replicated throughout the net. There is this magnificent, utterly, I highly recommend to all of you. It is like the best. It is a very short blog post called Holy Trinity. It's about this idea. How can we construct a hierarchy of meta-language by understanding the deep correspondences between math, the question of a structure, between logic, the question of proof, and

the question of computation, program, allotype. One of the things that is really important here, you might ask why am I talking about all this esoteric shit, you know, computation, logic, what has anything to do with all of the stuff that we have been talking about? Well, that's what intelligence is.

You see, what is intelligence? By intelligence I do not simply mean intelligent behaviors. A termite has an intelligent behavior. It is capable of creating a bunch of them in a swarm dynamic and make a massive ziggurat in their own terms or in our own terms rival to Ghazanbeel, the first ziggurat that humans ever made on this planet in the Elamite Empire but why are we talking about this well you see we are not simply making intelligent behaviors yes we are both

animals a lot of our in but also we have certain kinds of qualitative emergent aspects which can super ween on the very sentient namely animal biological traits that we have Like, let me give you, you know, coming back to the idea of this whole idea of new materialism, post-humanism, a lot of new materialism stuff.

let's say that okay a bee has a mind first of all i would say from if i was a carnappian which i am a fanatic carnappian in this sense i would say what does that even mean what kind of mind are you talking about mind in what sense you see this is why i say that when we are talking about the question of consciousness, mind intelligence, we have to explicate them. Otherwise, confusion ensues. And through confusions, we make so many wrong metaphysical, philosophical moves. We think that B is like a sapience. But do you actually, no, I mean,

really, even the most egalitarian of you, do you actually entitle a bee to the same entitlements, rights, the concept of justice and beauty that we hold ourselves to? I really want to know, how many of you do that? do you actually chomsky has this you know i'm not a chomskyian even though i have talked about him but i really think of him as a great man chomsky says something about this war on terror

that the war on terror do you know what it looks like it is not something like the time of Nazis, where they actually targeted innocent people and send them to concentration camps. He says that the war on terror is more like a human unknowingly walks on ants, not knowing that he is walking on living beings, precisely because they are not under his radar as living beings.

Questions? Now, this is, you know, we are diverging here, but I think that this is the whole point. That's why we are talking about AGI or AI in the age of intellectual scarcity because I think that this is the intellectual scarcity we try to be a egalitarian with regard to other animals but what is our egalitarianism where is it are we actually egalitarian to each other if we were truly egalitarian we should have a start from our own quarter rather than just saying that oh B has a mind. No, B has not a goddamn mind. It has intelligent behaviors and it has a great kinds of intelligent behaviors but it doesn't have a mind.

to have a mind is more than to be capable of solving a problem finding the best possible path to the nest and from there to the orchid. Mind comes with ethics and ethics is the very the other side of the coin which is called logic or conception.

Bring on the challenge here. Hello. Yeah, I think the bee is nothing without the colony of bees. And to talk about it, I don't know, maybe you mean the colony of bees when you say this. because one bee is part of a structure, you know? And maybe people are a bit more individualistic at times. Maybe these people. So I don't know.

I don't know. When you say the bee has a mind, do you mean the nature of the colony or the individual worker bee or the queen bee or something like this? You see, the question of mind is quite vague. Again, it's exactly like life. When you say that, what is life? Like Edwin Schrödinger's famous book, and I really do suggest to read this book. It is a magnificent book. It is truly one of books that show you who we are and why we have to be humans with dignity without

actually trying to whitewash this idea that oh we are children of a sky daddy or we are some unique Aristotelian species at the ladder of us at the top of the latter of the species. No, we are just Darwinian evolved things, but we are Darwinian kinds of things, that we do have certain bright kinds of mechanism, by virtue of that we can do something else. To do something else means that we have certain kinds of faculty and capacities which can reformat or mere animality mere being apes or because

otherwise we are you know sure we are not apes in it in evolutionary sense it's just a bad evolutionary thing but we are the little is coming from the great aims. That's the whole point. What is exactly mind? To say that what is mind is already a bad question. It's a pseudo-problem. It's what you might call to be what Karl App calls the opiate of the educated. People who say what is consciousness, what is life, What is mind? What is justice? So on and so forth. If they haven't done the great job of explication, refinement of these concepts, they can do,

they can say all sorts of bullshit under such umbrella terms. Okay? So, what is mind once explicated? mind is the organ of a structuration simply in the philosophical term we call it an organ for arriving and intelligible intelligible intelligible is a qualitative is both quantitative and qualitative index. Let's say that an ant who uses dead

reckoning using fromance to get back to its nest is a great problem solver, right? It shows massive amount of intelligent behaviors but does it really have a mind in the sense that we talked about mind as a refined concept a refined mind as a refined concept means that the world doesn't give its secrets to us world does not reveal its structure to us simply by way of senses we use conceptualization we use scientific theories we use mathematical physics to discover sectors

of reality which are beyond our senses have any of you seen an electron a field dynamic? No, but nevertheless we talk about them, we speculate about them, we make accurate predictions about them. That is what mind is and mind requires conceptualization, theorization, mathematical physics, all of these in conjunction with one another. Mind is the organ to reveal the structure even if the world might not have an intrinsic structure.

And that's why I think coming to Alejandro and Hermann talking about this, this is why I said in the last session that all of these singularitarian technological cults actually do converge with this all this new materialist post-humanist stuff about you know a rock has an intelligence a rock has a mind forest think oh my smelly sock also thinks so on so forth but what does that even mean this is exactly you when if you really want to

be honest about this you see that the undergirding assumption of both these creeds is nothing but neoliberalism what is neoliberalism neoliberalism believes that individuals are the indexes of intelligence. Just like one ant makes a dead reckoning, or a human simply navigates, or a rat navigates a maze, so on and so forth. But that is not what intelligence is, intelligence as a consequential vector.

That's what intelligent behavior is. When we are talking about intelligence, we mean minded intelligence. In the sense that, neoliberalism is essentially an extremely smart Darwinian political Darwinian paradigm. So how it works is like this. There are only individuals, okay? There is no collectivity in liberalism. The collectivity is just not your collectivity, it's not the communist collectivity, it's not some sort of robust collectivity. So there are individuals. Each

individual has certain preferences, they have certain options. These options are not coming from rationality. Essentially the content of your belief doesn't mean sheds to neoliberalism. These are, you see, neoliberalism actually deals with the content of your belief, whether it is bad or good, as a noise. What is considered to be a signal, a useful piece or element or element for neoliberalism is simply that you can make preferences, you can make options. Doesn't have any, doesn't want to know about what the content of these beliefs or choices are.

Essentially, the competition of these choices and preferences, regardless of their content, which are basically deemed as noise by neoliberalist economy, are important. important because once these preferences put together in a multi-agent system in a very game theoretic framework then whatever preference whatever option that survives shall be said to be you know the best option the best preference. So this is the individual way. And then, so this is from the bottom. At the top, so neoliberalism is not just bottom-up or top-down political paradigm.

It actually has both a bottom and also has a top. So at the bottom you have these kinds of, you know, game theoretic formulation of individual preferences. On the top you have a model based on this very paradigm of individual competitions in terms of their contentless preferences or choices or desires, whatever. And that is called free market. Every market essentially allows you to enter in any sort of competition as long as everyone

can plug its preferences into this market. And then the interaction between this top and bottom is exactly what neoliberalism looks like. So neoliberalism is not just like some sort of stuff that usually Kitsch communists talk about, oh neoliberalism is bad at this. No neoliberalism is actually a very very smart game theoretic account of humanity, but a humanity divorced from content of their beliefs, simply reduced to a cluster of preferences in constant

competition according to the laws of Darwin or if you want to get even more radical about it as Nick Land would have said it according to the Hamiltonian laws of conservation of physics. So and here you get the vision of AI in land. Fitness criteria. Fitness for the competition. This is not really your typical social Darwinism. Social Darwinism usually has so much political baggage. I don't think that land version is actually social Darwinist.

Because social Darwinism, when you look into the history of social Darwinism, social Darwinism always tries to couple these fitness criteria or dynamics of competition with some sort up you know wishy-washy political agenda but no land doesn't actually bug down his version of pure game theoretical individualism with such petty political concern coldness be my god is he's actually quite right but but of course this is just a total myth a total myth it is based on a fundamentally pre scientific idea of how humans function

Questions? David? it sorry dial dial dial dial i'm sorry i like my eyesight is really bad dial and you were saying some stuff yeah open up open up can you hear me at all yes oh yeah um i don't know i uh I guess first of all I'm kind of I'll turn my face on I'm interested in the

pre-scientific idea of how humans function which I think you just said like land's coldness is based on I may have misheard you um no no in the sense that you see in the sense that the whole point is that okay number one do you believe in complexity or not I'm sure that land believes in complexity right okay but the thing is that land's notion of complexity is purely based on a stuff which are now called folklore of complexity sciences. I will give you a few

examples of these foreclores. The idea that variation equals complexity. Nonlinear dynamics equals complexity. Divergence from initial conditions means complexity. No, none of these actually you can in fact for each and single one of these examples, you can make a counter example of a very non-complex simple system that exhibits such features but cannot in fact be called a complex system. The canonical measure of complexity in today's complex science is something called a statistical stability. So what is a statistical stability? Let

Let me just tell you, it's the idea that complexification requires a hierarchical stabilization of statistical decisions at each and every stage of the functioning or the structural evolution of a system. in the sense that, for example, a very, very rudimentary, perhaps even misleading example. So don't put this on Twitter that Reza said it because it's just bullshit. But I'm just giving you a little bit of this. So when you are making a wall, you have to put your bricks in certain and certain kind

configuration you should make sure that they are leveled you know you put a level on them you put the adequate cement between them so the first row of your bricks is what you might call to be degree zero of a statistical stability in your system then after that you make the second row The second row should absolutely conform with the first row. Essentially, as you build the wall, each row of the bricks, the cement, and elevation, and the levelness of these bricks, put constraints with regard to everything that you put on top of it.

This is what complexity is. The same thing about humans, sure, we are all animals. To say that we are not animals, it would be just like some sort of theological bullshit. No, we are Darwinian, but the thing is that these Darwinian evolutions, the certain kinds of mechanisms that has afforded us, have allowed us to create more and more constraints. to the certain point that these constraints were no longer naturally viable, naturally viable, okay? You know, there are massive amounts of literature on this point that why do you think

neoteny, which is one of the greatest, you know, features of the great apes, including us, a certain proportion. Essentially why is that our brain... Neotony is the idea that, you know, basically the complexification of the brain or the central nervous system at some certain point when it wants to become complex, it requires a massive amount of time for development but also a great size of a structure. Okay, let's imagine a counterfactual cartoonish scenario. Like imagine if this evolutionary thing had been progressed further and further,

you would be 70 years old and you were still behaving like a goddamn child. That's what neoteny is. But also at 80 years old, your head would be the size of probably your house. I'm just making a cartoonish Walt Disney example for you. The thing is that natural evolution is fundamental, fundamentally under constraints already entrenched in your evolution. And there are certain limits in the evolution because if these limits, the so-called thresholds of complexification are transgressed, you go extinct.

this is why when you see among the more stabilized forms of life on earth they don't suddenly after like 100 years make a god damn arm out of their head like imagine you know you created 300 nose on your face that just doesn't work. You will die. You see, evolution is a complexification process. Addition of complexification is always disproportional to variation. So as the complexification increases in

organisms, variations turns down. You cannot have so many appendages or massive amount of variations in your structure so the reason from an evolutionary perspective that humans became such big deals in their own shitty worlds precisely because they managed given their own complexity the platform the evolutionary complexity they managed to create certain different levels of complexity which were which were not entirely natural or even though they were naturally evolved

the way that they used it was not exactly natural reason rationality inference concept use is naturally evolved but doesn't mean that is unnatural to say that is unnatural doesn't mean that it is supernatural Shall we have a break? Maybe we can have the responses and presentation

after the break. Yes, that would be super, super, super, super. superb so then let's have a five minute break or longer or shorter as you like excellent love you guys see you 30 seconds from now Thank you.

Thank you. Thank you.

Thank you. Thank you.

Thank you. Thank you.

Thank you. Thank you.

Thank you. Thank you.

Maybe Sam should say something. Sam somehow inadvertently revealed himself that he has been reading intelligence and spirit. I hope that he wouldn't have done that, but now that he has done that, maybe we should ask him something about some sort of comments. Yes, so what would you like me to talk about? Whatever you want to talk about. Well, yes, I did read your book.

I liked it very much. so I remember last time you were going over some themes that you also discussed in intelligence and spirit you talked about the role of global Lyapunov exponents in the description of nonlinear complex systems and how that was tied into Rodin's account of the post-human future along with Humphrey's notion of diachronic emergence. Yes. So I grasp all that, but I was wondering if you thought this was the only model

that we have currently available that would be appropriate for Rodin's disconnection thesis or whether there was some other model for the divergence of, for the radical divergence of physical systems. Yes. Well, yes, that I actually want to talk about is, but you remember, you know, after the exegesis on the global Leopold of Exponent and also the judgment under uncertainty, I made like one or two paragraphs on this idea of the computational asymmetry.

Sam, can you hear me? Yes, I can hear you. Okay, okay. Sorry. So, you see, I'm going to talk about this. But the thing is that here a problem arises. So you see, the way that I think of transcendental philosophy, I mean, with regard to Kant, sure, I mean, Kant is just like, you know, his logic is Aristotelian, you know, he makes so many dodgy moves, so on and so forth. But who doesn't as a philosopher? I have done my share of dodgy moves, you know, quite often.

But here, in the spirit of the criticism, we can think about today as, in fact, transcendental philosophy, or what you might call to be transcendental epistemology in the vein of Kant, actually tries to refer or address to certain kinds of questions which are now becoming the most important questions of theoretical computer science. In the sense, you know, for example, the question of alien. How can we ever see an alien? By alien I don't mean people like us, like in Arrival, that, you know, they talk, they

make some sort of, you know, alphabets that is alien to us, but nevertheless we can put them together and communicate with them. I'm talking about Balsamanian aliens, in the sense that even the forms of intuitions of such aliens is fundamentally different than ours. the very principles through which the three synthesis, which are, you know, central to basically sensation, perception, conception and imagination and schematization would be

fundamentally different. How can we ever, ever come into contact with such aliens, right? Like Fermi's paradox, where are they? Where are they? The thing is that here a problem arises. The problem how I think about it is that, so you talk about, what do you call it just now, what was it, the last word that you used?

Something that I said? Sorry? something that I said yes yes yes you you said you said a specific term but it was basically this idea that you know so okay how you know is there is there is there a way that we can actually talk about the aliens without actually you know overemphasizing this you know model of dynamic divergence aka maximal Lyoponov exponent yes I would say that

Kant already had thought about it and if as I mentioned to you if we read Kant in contemporary from a contemporary perspective from the contemporary lens we see that the Kant's transcendental philosophy is essentially what you might call to be a problem of computational complexity from the perspective of of theoretical computer science. So we have such and such levels of computational hardness, you know, we can parse this kind of things, we can parse that kind of things, we require this amount of time, we require this amount of memory, so on and so forth. Okay. But what

What if we actually come across an alien whose level of activities does no longer correlate with the hierarchy of our computational complexity? it is non-computable for us. Well, that is another problem. You know, how can we ever talk about this? Sure, there can be such a stuff in the universe. I mean, we cannot just say that, oh wait, they don't.

Well, what is it good? What is actually beneficial to such a speculation? Everything that we should talk about the non-computable indexes of intelligence should come through the labor of expanding of our computational resources, rather than just simply speculating about that which is not inside the purview of our computational complexity hierarchy. Because wouldn't it be the case that if we are trying to speculate about such stuff

without actually expanding our computational resources, that kind of discourse would amount to exactly the same thing that the scholastics were talking about, Neoplatonists, you know, abstract negation, apophetic theology. So this is why I say that I think that Rodin's thesis is fundamentally inconsequential. It's just that the way that he goes on and talks about it, and by that I mean that he tries first to get rid of the conceptual resources that we have in order to posit a disconnection

a disconnected uh intelligence that really to me is as bad as saying that oh we just resurrected the no platonist god no if there is a way to talk about this stuff coherently rather than making armchair speculation, we have to start to alienate ourselves from the kind of humans that we are here and now to different kinds of human with further computational capacities. Otherwise we could talk about not just post-human intelligence, we could talk about angels, devils, we could talk about ethereal

beings, so on and so forth. Should we start with the presentation of the student? Sure. whose first whose head is on the guillotine first I remember oh my god Giancarlo is not here oh yes that's the best way the best way the best way is

is that we have this kind of mythology in Iran that many martyrs were telling to the Arab Caliphs that, oh, you can take my life. I will live, I will give my life to my beliefs. When they ask them to come and basically being hanged, they don't show up. this is John here at this point so shall we have to respond to respond to Sam Sam Sam is going to be the respondent yes

So I'll switch to Sam. Sam, are you here? Yes, I'm here. Sam, may I, if you are not in some sort of compromising situation, would you be able to turn on your camera? Sure, just a moment. Okay. You have to podium we are all ears Okay, and this was the response to the rodent tasks from last time

Right, okay. So this is sort of a continuation of the question I just posed. so with regard to the rodent text the model that was given for the disconnection thesis was that of nonlinear dynamic systems described via globally Apanov exponents

and these systems cannot be adequately used to model artificial general intelligence as we're talking about it insofar as the kind of disconnection that Rodin is requiring for his thesis is a mathematical idealization of the global Lyapunov exponents occurring only over an infinite time period. So that was one aspect of the criticism from last time. The other aspect that I recall was

that of a base rate fallacy. insofar as the use of, say, self-autonomizing technological systems or self-replicating technological systems is a subjectively selected subset of possible priors, which could be used to predict the trajectory of post-human intelligence and one that's selected without any real logical rigor. So in that sense,

it's inadequate because it's without any real method. So does that adequately get at what you were saying from last? Yes, yes, yes, yes. Please go on. Please go on. Okay. So in light of that, then, and this goes back to the reference to Jabo's blog post on Pascal's wagers, which you used to describe this sort of epistemic failure.

This is as much a question, again, as it is a legitimate challenge. the question I suppose is what what methods or what sort of process could be used to properly identify the trajectory of the development of intelligence based on current activities or current features of artificial intelligence research or human intelligence. So I suppose that was really the question or challenge I wanted to pose.

Superb. Any more? Any more? Anything? That was about it, I think. Okay, superb. Okay. Thank you so much, Sam. Thank you. Thank you very much. Really appreciate it. Sure. Do you still need my camera on or? No, if you want to do something compromising, turn your camera off. Thank you for considering me. So, yes, I think that I talked about the model for, you know, disconnection thesis

as GLE. As I said, that GLE global or maximal Lyopoulx exponent is just simply why I would say that is what you might call to be an unespoken implicit assumption behind the narrative of the disconnection thesis. It's not by any means a model that I would say that Rodin would actually say that oh this is based this is where i could get this from no actually i would say wouldn't would vehemently say that no this is absolutely not the case but nevertheless

if he says no to that then my answer to that would be then where did you get this from in what scenario in what you see you see this is the whole point I told you this all of you that that I didn't try to attack Rodin's position from perspective of no rationalist boring Brandomian stuff. I don't want to play that kind of role because I know that he is going to have more massive challenges to me. All I'm going to do is actually using his own

assumptions against his own argument. And then I would say that, okay, if you say that the divergence from the initial conditions in the sense that the disconnection on valid post-temen would vastly diverge from its human substrate, where did you get this from? You know, obviously you don't believe in norm as intentional bloatwares. You think that norms should be actually naturally grounded. I mean, so far as they are naturally grounded, then you cannot just selectively say that, well, I actually choose to naturally ground

these in Darwinian or some other kind of theoretical pattern. No. When you are in the framework of naturalism, unfortunately that is the whole point, and that's a very dangerous thing. When you are in the domain of naturalism, there is no button to it. The ultimate button is physics itself. So you have to talk about these kinds of assumptions that you have made by way of physics and physical models. Because then it really doesn't mean anything if you say that, oh, I think actually norms

are naturally grounded. Naturally grounded in what? In your selective, you know, basically semi-normative, semi-physical ground? No. If it is not, cannot be explained by contemporary physics and complexity sciences is not naturalist. It's just another fucking pseudoscience. So that is my challenge to Rodin. Let's take this seriously. So you are using this theme. You are using a theme. You are also saying the norms are going to be naturally grounded.

take this to its ultimate conclusions. Let's go back to not like quantum physics. Let's just go to a little bit more, you know, more friendly physical vista, which is called complexity sciences, where we can actually talk about different qualitative trajectories, phases, spaces, systems, so on and so forth. Would you be able to tell me how actually your theme can map onto this model of complexity science? number one

number two math idealization yes I mentioned that maximal or the so-called global Lyapunov exponent It's just simply a math idealization. Essentially, it is a model which is fundamentally idealized and can only be applied to certain kinds of simple models. Within mathematical physics, inside the domain of complexity science, the application of global Lyapunov exponent in the sense

that there would be an explosive growth of divergences from initial conditions holds no ground. Because such an assumption only holds ground inside the mathematical idealization of time when we actually apply this mathematical idealization to a physical target system we see that the remember I remember I I can't remember who asked me. I told them that, you know, think of this as this diagram.

An input goes in, it's called the perturbation of the system. Then there is a system and from the system there are these squiggly arrows coming out. Each of these squiggly lines represent possible trajectory for the evolution of a dynamic system. And each of these are plotted within a state space of the system now the thing is that in reality namely outside of the mathematical idealization local divergences within the trajectory of the systems do actually converge over sufficient time

rather than diverge. They converge upon initial conditions rather than diverging from them. Hence the folklore that the dynamic system given sufficient time its evolution would be fundamentally different from its initial condition. That's just a folklore, a folklore that is true within idealized mathematical systems, but not in physical systems. And if Rodin actually thinks that human is a physical system, which I think he does, then really such model holds no ground whatsoever.

So, base rate fallacy, selective views, or probabilistic priors. Yes, I talked about this and I'm going to talk about this a little bit more today, so I'm not going to go over it right now. Then the last one, what method or sorts of method can be used to predict the current trajectory in a genuine manner if I'm not mistaken that was Sam's last question the thing is that there is no such a thing as a such a method you see

Yes, we can do predict the trajectory of a dynamic system, but when you add the adjective genuinely or in a genuine manner, then we are no longer in the business of real physics or a physical target system. The word genuinely is already extremely biased from a probabilistic sense. Yes, we can predict it, we can predict it. But a stochastic fluctuations

can go all sorts of ways. all we can do predict is that given such initial boundary conditions such dynamics that is unfolding through the system given the sets of the trajectories or tendencies of the systems after a perturbation this might be the prediction of the system this prediction is not a genuine prediction what what do you mean by genuine prediction as if it is going to be certain no it's not going to be certain that's why it's called a prediction a prediction that is genuine

early in so far that it actually abides by the resources of a statistical and probabilistic analysis we have in our arsenal right now here and now that's it nothing more nothing less to say a genuine prediction we are lapsing back in the Laplacian universe such that you know if the motion of the celestial body is perturbed in such and such way because of these causes, it's definitely, genuinely

going to be in such and such position with such momentum and speed in such orbit. But that's a Laplacian universe. That universe has been abolished long time ago. We are in the domain of a statistical analysis at this point the domain of a statistical analysis is quite very different from that Laplacian determinism it is essentially something like that that I hit my head my hand on this table under such and such conditions in time t1 space one such force such momentum such salinity of the material

all of these conditions according to the statistical probabilitical formulation we can expect we can expect that when I do this when I hit the table with my fist there would be a range a range of statistical probabilities under which the effect would be this or that you see when we are talking about a statistical universe we are not talking about certainty certainty just get rid of that that's just a pre-critical concept we are talking about the statistics

Doing this means that inside this threshold, certain events might occur, certain events might not occur. But nevertheless, this is the range of possible behaviors of cause and effect. This is what usually is called in philosophy of science, nomological expectancy. nomological in the sense of law and expectancy expecting that such and such things will result within the statistical range if such and such conditions are met that's it

questions Valentin, I see that you are opening your mouth and you are saying something. Just say it, man. Go on. Let's spill the beans. I think we already discussed the first time and I can't wait to switch to Malik. But do you have any sort of criticism here?

Anything that can help us or other students to get this idea a little bit better? I never really believed in what humans to do. I never really believed in this kind of prediction. The fact that we can... Prediction in the sense of Laplacian universe? Yeah, I mean, think about the sciences which try to work with what they call complex systems. Think about any kind of attempt to create such a science.

It's always, for example, huh? You see, my apologies to cut you short. Just simply, I'm going to give you a little bit of an interim and then please do correct me. You see, when we are talking about prediction, as I mentioned to you, prediction is also a concept which is clutter, wake. We have to explicate it, okay, in the robust probabilistic statistical sense. To many people, prediction means that you are predicting that when I actually hit my wall, I will have a fucking, you know, I will faint.

Okay, something like that. That is just bullshit. That is old notion. predictability in a statistical sense is a range of a statistical numeracies within which you can actually see a pattern with the threshold of possible. This event might happen or this event might happen with certain variations or this event might actually do happen. It's a range. It's a statistical range. Prediction is not something like a prophet making some sort of predicting about the future.

It's not like Nostradamus' idea. a statistical prediction, it's just simply a range, a range in which a pattern will be derived and we can of course interpret this pattern in very very different ways according to our conceptual theoretical tools. If we talk about prediction, I just thought about it because when you read Malik or rather Esposito, this account of financial markets structuring our contemporary world, it means

that prediction means something very specific in this very particular realm. And it might be so that the concept of prediction we are thinking of right now is slightly different from some other possible concept of prediction, which would be used in a world not Right, right, right. But that's why I haven't got to the Sohaeus. But don't you think that that very idea prediction is pretty critical? Yeah, so for me, I'm always very skeptical about this kind of stuff, because I saw a lot of people trying to do something about what they call a complex system.

It can be cybernetics, it can be analytics, it can be something like that. And you always look at it, and you see they publish basically two kinds of articles. One of those, they look at a very simple thing, just very simple, like I'm kind of . And they say, this is negative feedback, this is amazing. And then there is other kind of articles where they, in an incredibly speculative way, generalize it to such an incredible speculation about morality and future of humanity that it's a... Yes, yes, yes, yes. No, no, I think that we are both on the same page. Essentially, this is the whole point that I was, I'm going to build on.

and this actually becomes extremely, extremely weighty for people like land. And essentially that they have this precritical notion of probability and predictability and from which they make a new ethics of posthuman. An ethics of posthuman which is nothing but a relapse to the most infantile intuitive notions of humanity anthropocentrism 101 Well, in case of land, I guess we will come to it.

I know that you are a little bit close to land. Don't worry. Your secret is fine with me. I'm a friend with land. It's all fine. No, to be honest with you, land is not, to be honest, you know, OK. Land's politics to many of us, including me as a communist, looks atrocious. But I'm not going to simply say that land philosophy is utter crap just because of his politics crap. No, if you wanted to say something like that, you should have gone for the Quine.

Quine was a racist. He was a supremacist. He actually was one of the most conservative people on the planet back then. But he was also a great, great philosopher. You know, I do not believe that philosophy and politics can be separated from one another. They always do converge. But we should not also be so blatant to conflate every philosophical thesis as if they were political theses. Let's talk about land.

I think there is a big problem with any kind of science which uses as an object, what they call a complex system. It's just like from this very beginning when you say, like, physics is not enough for me, sociology is not enough for me, Marxist analysis is not enough for me, I want to study a complex system because it's such an interesting thing, you just lose the idea of object as some kind of a scientific thing. You say that the complex system is exactly the thing that transcends object, but you still want to study it using scientific method, and now you basically keep picking your axiomatic at more or less random.

Yes, I agree. But, Valentine, you see, the thing is that I mentioned this earlier on, you see that this view of the complex system study, complex sciences that you are telling me has been, of course, the very paradigm from, I would say, Bert Lampy. when he wrote what is a complex system to Norbert Weiner to at least 80s but no longer complex science believed in such things complexity is not about an object

Complexity is about what a system does. And what a system does is by virtue of a certain kinds of structures or configurations of structures. Have you read James Ladyman, What is a Complex System? Definitely. I absolutely recommend this to all of you. It is one of the greatest essays written. It's highly cited. It is by James Ladyman. It's called What is a Complex System? You can find the PDF online easily. A TED talk apparently. It's probably discredited the post.

Or if you can't find that, read this one. one I'm pretty sure on this I have dementia which I know that I have there is another essay on this topic by James Crossfield James Crossfield used to be the head of you know Santa Fe complex science department the name of this essay is the calculi of emergence to calculate our emergence

okay anything before I deliver the last below to my greatest friend David Roda no okay Rezo sorry I made a question on the sidebar but it's too noisy here so I decided to write my apologies how do I actually get access to the sidebar I don't have the sidebar here there is up there

in the top left there's a little box oh yeah I side okay okay a little message box yeah yeah okay okay my apologies sorry I'm bringing my face I I'm legally blind so I have to Herman, would you be able, okay, I got the gist of your question, but would you be able to elaborate this a little bit for me?

I guess my understanding is that when you try to predict some phenomena, you have to make a certain decision, then it's functions. I guess, like I was saying, how do you go about choosing this? It sounds like arbitrary and you were talking about two distinct types of induction. I'm not familiar with the logical induction, I'm familiar with the frequency conduction. Yes, yes, yes. Well, Herman, this is not really a question that a philosopher like me can actually give you the best answer.

Do you really think that if I could answer you this question, I would have been teaching you guys? No, I would have been getting like the big bucks from some sort of institute. But let me just give you an example about this. You see, this is actually a very, very serious problem. This question actually has been proposed by Kant, even though Kant really didn't understand the weight of this question, even though he was in converse, he was basically criticizing Hume problem of induction.

So you see, okay, let me tell you this. So, you see emeralds, okay? Every time that you see an emerald, the emerald is green. Emerald is green in time t1, emerald is green time t2, t3, and so forth. and then you say that after the hypothetical time, TN, all emeralds are green. You see, all emeralds are green. You are making an inductive judgment. An inductive judgment is all emeralds are green.

But you should understand, when we say all emeralds, we are not making a logical statement per se. It is not like that all bachelors are unmarried. That's a deductive. That's a deductive judgment. Here, we are making an inductive judgment. the designation all is not about deduction it's about induction it's about the frequency of the instances of the green emeralds that we have seen so far okay okay so this is the first example now Now let's think about a different scenario, a scenario that has been proposed by Nelson

Goodman in Fact, Fiction and Reality. Superb book, really, really highly suggested. So imagine that I see time T1, emeralds green. Time T2, emeralds green. T3, emeralds green. Time Tn, up to time Tn, all emeralds are green. Good, fantastic, so great. But after 20m, for some reason, I see blue emeralds. Now, the conclusion of this inductive judgment

wouldn't be the same judgment that I just made, namely, all emeralds are green. The true judgment would be something like that all emeralds are grew. Green before time, Tien. Blue afterwards, hence grew. Or like all crows are black before time, Tien. But then I saw, you know, white crows. And I said that all crows are blight. black before time Tien

and after time Tien white now from the problem of induction both of these judgments are actually justified this is called the rule of induction simply by way of observation or the frequency of observation you cannot make judgment that all such emeralds are green or such crows are all black you are also justified to say

all emeralds are grew or all crows are blight okay now the thing is that then how can we go about this disparity between such judgments between black and blight green and grew well the observation based induction absolutely doesn't give you any solution to this problem this is why it's called the riddle or the problem of induction it is one of the most nasty problems of philosophy the thing is that they are both justified the reason

that you are not using grew instead of green or blight instead of black is precisely because you have been brought up in a linguistic framework in which black and green are your entrenched predicates or adjectives. If you were brought up in an alien universe where grue and blight were your adjectives, you could as well say crows are blight and emeralds are grue, and you were good to go.

So this is the whole point. This is one of the greatest mysteries of philosophy. Essentially, this is the point that we should understand that whatever kind of judgment, perceptual judgment we make about our inductions are not just simply neutral, as if we were just simply neutral observers saying that, oh, this is green, time t1, time t2, green, green, green, green, green, and then something like that. No. The reason that we function like that,

we make judgments like that, we say emeralds are green, it's precisely because our conceptual, perceptual variables are entrenched in long years of our linguistic entrenchment, if we had a different language, in which it could admit perceptual, conceptual poles such as green and blue, grew or blue and green bleed or black and white blight, then we could make such judgments.

That's, and you see, this is not just simply to show that we are habituated to our languages. a greater lesson that can be learned from this thing is that the so-called alien if you actually really think about an alien it can have a different language in which grew is a natural adjective unlike in our language where grew is an unnatural predicate They would see emeralds as grew rather than green. Question, question.

Thanks, Reza. You mentioned Goodman. So would that be a good book to start going in depth? Yes, yes. Okay, if you want, okay. Fact, fiction, and reality is more about the problem of induction. It's a really great book. But if you really want to look into the very problem that actually gave rise to this group paradox bullying products fly products this kind of unnatural predicates and adjectives of observation i would say read goodman's total gem it's called the world it's called the ways of world world making

excellent thanks anything anyone okay let me let me just finish this and then we go our way. So, my apologies.

Sorry, I lost the page. So we talked last time about the Pascal wager, which Nick Jabo, with regard to these kinds

of singularitarian accounts of HGI as disconnected calls a Pascal scam. So to continue, this disproportionate wager, as a reference to Pascal's wager or Pascal's scam, This disproportionate wager on the magnitude of uncertainties is a speculative trend that aligns the disconnection thesis with other singularity-driven scenarios, where bets on the rise of a malevolent or a benevolent superintelligence are being touted even less discreetly, as Pascal's scam.

Example, Skynet, Paperclip Maximizer, Rökels Basilisk. However, what connects extreme scenarios associated with judgment under uncertainty is not simply the biased over prediction or under prediction common to them, but also the central role played by intuitive impressions and adumbrations in rendering them extreme. Their radicality is fabricated by those exaggeration-prone cognitive habits

that belong to an image of the human whose diagnostic prognostic abilities are still bound to its evolutionary infancy, as yet unfettered by critical rationality or science. Rodin does signal caution about the exorbitantly speculative dimension of unbound posthumans and instead favors in quote, graduation from speculative metaphysics to a viable cultural research program, end quote. But the radicality of the disconnection thesis and the strangeness of unbounded posthumans

are precisely founded on this unwarranted speculative metaphysics. Once this unverifiable speculative dimension is removed in the graduation to a viable cultural program, the disconnection thesis loses its radicality. and the unboundedness of posthumans appears more casually and normatively constrained than Rodin's claim. Sorry, causally and normatively constrained than Rodin's, that Rodin claims.

Although Rodin resists the characterization of unbounded posthumans as alien or inherently uninterpretable, there is no indication as to what the interpretation of unbounded posthumans would entail, particularly when the accounts of intelligence and agency provided have rendered them devoid of any conceptual content, semantic content. Even if we approach the interpretation of unbounded posthumans from a computational standpoint, and no longer from the perspective of predictive accuracy or theoretical conceptual fidelity,

interpretation of such a phenomenon would be so costly that it would become completely unfeasible. Computational cost grows on average for an observer as it climbs the complexity computational hierarchy. In other words, the size of the observer's internal model grows as it attempts to model or make predictions about phenomena at higher levels of complexity. This increase costs the observer. For example, if the observer is a biological organism, it costs physical and metabolic resources.

The problem of computational cost has also interesting implications for modeling. Those models, in this case, models of intelligence, in the case of Rodin's disconnection thesis, whose measures are set on a higher levels of complexity are not optimal or even feasible models. They are, in fact, cannot be called models because every fucking model requires a certain balance between order and chaos, noise and signals. If you increase the size of the model, you get noise.

You get less amount of signal message. If you increase the size, if you decrease the size of the model, then the noise increases. So the entire idea of the model making in science is about this balance between noise and signal, order and chaos. Otherwise, literally, you don't have any sorts of computational effectivity to interpret the model. Model becomes as alien as angels dancing in the heavens. Those models, in this case, models of intelligence, whose measures are set on higher levels of complexity are not optimal or even feasible models.

For instance, Charles Bennett's logical depth is a measure of complexity of a string S in terms of the time needed for a general purpose computer, a universal Turing machine, to turn the shortest program that generates S, a string S. The problem with Bennett's and other similar models is that they attempt to interpret or measure the complexity of an object, example, general intelligence as a Bennettian deep object, from the uppermost level of the complexity computational hierarchy.

In the case of Ben's measure, the most powerful and resource-consuming class of formal languages and the respective automata, namely the universal tiering machine, computable class, which is the most cost-consuming machine in the hierarchy of computational complexity. In a starting from the upper level of the hierarchy, they run into the problem of poor effective computability. So you see, the more the cost of computation increases, the least effective computations become.

Effective computation is equal to the level of computational interpretation in a rudimentary sense. So, for example, you get some sort of text that is utterly made of some sort of really, really strange crypts and stuff. And it requires a massive amount of computational cost to decode it. You see, this amount of work that you put to decipher it is proportionate, okay, to the decrease in the amount of effective computational interpretation of the text that you are trying to decode.

Just like a future intelligence. So if the future intelligence is such a big deal, okay, it is not really has any sort of palpable computational connection to our resources, then how are we going to interpret it? Remember, the whole point was that Rodin tried to tell us that this disconnected future intelligence is not uninterpretable in principle, in quotes, in principle.

But now, if the said unbounded post-human is beyond the hierarchy of our current resources, then even, let's not talk about rationalist, conceptual, intentional interpretation of such things. Let's talk about a computational interpretation of this intelligence. How are we going to computationally recognize it? How are we going to calculate it? How are we going to, in fact, recognize it as a pattern that can count as an intelligence? Because if this very pattern requires so much computational cost beyond our purview right here and now,

we are not going to be able to interpret it. Increase in the computational cost equals to reduction in effective computation. Reduction in the effective computation means that you are not going to read this computationally as a pattern. Hence it becomes arbitrary. You could say about this post-future intelligence, as I told you, that, oh, this is a skydive, this is God, this is an angel dancing on the tip of a needle. but so he genuinely tries

to create some sort of minimal connection between this unbounded posthumum and the current resources that we have but okay let's get rid of all the intentional stuff because he doesn't want to talk about that let's talk about the actual computation from a computational standpoint this doesn't work either You cannot interpret such pattern as a pattern. Because the increase in the computational cost equals to impoverishment of your capacity to interpret a computational string as a pattern.

if you can't see it as a pattern then it can't be any other kind of pattern so why are we calling it a posthuman intelligence you could call number P or an angel or my partner or my cat as the posthuman intelligence so why are we so picky about this so So, starting from the upper level of the hierarchy, they run into the problem of poor effective computability.

in the sense that we can never be sure whether or not we have found the most efficient coding for what looks likely to be a random pattern. Already shorn of the constraining continuity with those deontic normative attributes of agency that make an intentional somatic interpretation possible, which, of course, Rodin doesn't want to talk about them. So be it. And that's the whole point. I mentioned last time that I'm not going to attack his position from my no-rationalist position. I'm actually trying to attack his position

from the very premises that he has built, his own account. Already shorn of the constraining continuity with those deontic normative attributes of agency that make an intentional semantic interpretation possible and now suffering an arbitrarization of computational interpretation, going to lack of effective computability, the unbounded posthuman can match any random pattern or description of any system. Absurd questions such as whether we can regard a galaxy, the number pi, or an angel as post-human intelligence become genuine topics of debate.

Even though Rodin attempts to leave some space for the interpretability of the unbounded posthuman as not uninterpretable in principle, this minimal space is avoided by the very criteria that are proposed to unbind posthuman intelligence from its human substrate. The room reserved for a level of interpretability in order to prevent the collapse of unbound post-humans into unintelligible alien-ness shrinks to nothing.

And just as the line within the estrangedness of unbounded post-humans and unintelligible radical alienness blurs, so does the distinction between the disconnection thesis and the speculative, apophatic or negative theology. Thank you very much. Questions, questions, questions.

I know that we are a little bit over time, but don't worry. No questions? Okay, if you don't have questions or you are too shy to ask questions, I can give you one thing. Now that the Google Classroom apparently is on, put your questions on the Google Classroom

and we will go over them. those of you who have immediate questions about words, terms, whatever, please do ask. Maybe now is the time to ask who will do presentation and response next time. Patrick, are you up to this idea that we should toss a coin? We are no longer asking for volunteers. Sadly, the coin has not, I don't have a coin with 15 sides on it.

Otherwise, I will do that. No, let me tell you that, you know, I'm infamous for making inappropriate jokes. Let me give you a joke about French Revolution. So there are a bunch of people who are going under the guillotine. You know, obviously no one is going to volunteer to go first. Well, we are the Jacobins guards. Let's toss a coin and whoever gets the coin is going to go under guillotine next session. How about that? Why not? Why not? Yes, why not? But don't we already have Giancarlo to talk about Malik?

Well, Giancarlo didn't show up. But anyway, what about Malik? Giancarlo is a smart guy. you know he's like he's like you know yesterday I mean last session he was part of French Revolution now he's part of the common have you read this book called the dice man yes very minor work of philosophy Yeah, that's great. Good arm. Great that you've read that. There is a story called the Lottery of Babylon,

which I think is just amazing. To be honest with you, one of the things, you know, maybe, Patrick, would you be able to go offline? I mean, offline for a second and then come back. I don't think that's possible. What do you mean? Oh, okay, okay, okay. I wanted to go off the record. You know, so when I was a child, one of the first books that I read, it was a book that truly, truly scared me, traumatized me for the years to come. Have you read the Black Tulip by Alexander Doman?

No, hardly not. Okay, okay. Read this. In the first 20 pages you get the actual version of how French Revolution looks like. like and if you don't be yourself you will get the real version of the French revolution in my class okay who is going to say something next week please someone i can i can do it okay super thank you so much another person another person

herman how about you sure why not okay super fantastic you see we are not we are not living in the reign of terror we are still have a semblance of idealitarianism after a few threads everyone said that why not okay that's super fantastic thank you so much my friends okay so we're um talking about malik right is there any chance we can get uh like uh not meaningful, but the key parts of this essay.

Yes. So basically, I really do hope that the Google Classroom is up. I will, tonight or tomorrow, I will upload. There is this essay by Suhail Malek. It's not part of his new book. It's just an essay that he wrote a few years ago. is called Contra Contemporary. And I have highlighted a few sections of it, which I think that we can read, precisely because I think that they are relevant to the stuff that we have been talking about with regard to land, road, and so on and so forth. And then read that, and then we'll talk about it in the post-stream. And of course, I will tell you

what chapter of Boe Strong you should read. OK, great. So we're not reading ontology of finance then? No, no, no. I actually thought that the finance piece is good, but it is not really that much relevant to the discussions that we have so far. So yeah, no. I actually find this essay, which is superb, super, really good. It's really fantastic essay. Sounds good. Top notch. OK, friends. Shall we say goodbye at this point?

OK. Love you all. Thank you. okay thanks reza see you guys bye bye see you see you see you next week see you next week