Hello, welcome to the 6th session of Nick Land's Anthrop poll. We have some people trailing in, so we're just going to start in. So Nick, whenever you're ready, you can start. We're live. Nick? Oh yeah, sorry. Okay, everyone seems, well, you seem very quiet, right? I don't know. I'm just wondering whether I've got something with my... Probably, it's just probably me. Oh, okay. Other people will chime in so we can hear really quickly. Yeah, could I just hear someone chime in? That would be great. Tony, you are very quiet, but Nick, I can hear you.
Okay, excellent stuff. Okay. I'll fix my mic, sorry. so so we're off the time to get that right and okay what this week I'm hoping that people seeing an extremely compelling structure in the sense that we are act the imitation game after passing through the black mirror Roko's Basilisk and the AI box experiment. So mostly I'm expecting to orbit Turing's paper. I mean I've listed it I've thrown it at people loads of times. I'll just chuck it up one more time
along with one other thing that I think is well worth a look. So this is Turing's paper, oops, sorry, one second. I'll just, sorry, I'll just do the link. It's called, well, it's called, well, it's called, well, it's called, well, it's called It's not called Wu, it should be perhaps, but it's not.
It's called Computing Machinery and Intelligence. And as I say, it's like, I've chucked at people a lot. I think it's, if there is a single text that we're dealing with that is in the role of Satoshi Nokomoto's Bitcoin paper in the last course, it's this one. I think it's like fantastically rich, interesting, influential, important piece of writing. we could have spent a lot of time on it. I certainly think we can dedicate most of this session to it without any sense of doing it. It's obviously old now. I'm pretty sure it's
let me just check some so I'm not misleading yeah if I've got that one I'll just expect spanking from people rather than wasting their time checking it out and I think that even though it's become a kind of iconic cultural document is probably not read that much and if people look at it they can see there's lots of strange things going on in this piece of writing that don't generally get conveyed out into the board of discussion one of the few
documents that I think has really given it a thought for a close reading in this way is the one I'm just going to look at right now, which is by a male and female team, which is, as you've seen, highly relevant, of Tyler Cohen and Michelle Dawson. And a lot of the things I'm going to say about it, they touch upon in their paper. So it's in seven sections. Hang on, sorry, I'll take one very, very quick back step from this, just to say, I'm
going to assume that it's unproblematic that there is a security context for this text. It's, I think, at the most extreme, if we're just coming right off the paranoid peak of the question that Turing is asking can machines think is can be reformulated, can machines think us and you're deep into these simulation argument type thoughts but I think we can scale it down a little bit initially from that and take it more in the sense that it's normally see which is can machines think like us
I'll and I think that the fact that this is a security question is clear to everybody because everybody's asked quite possibly every day now are you a robot you know you'll you'll be trying to communicate with someone or some website and it will ask you this question I will put you through its own little contrived Turing test or imitation game in order to try to decide this question which is thus for us eminently practical and it's a matter of just colloquial
existence now that we face these little Turing test questions and this are you a robot question has become something so almost inane that I think it's very rare for anyone to sort of do a big double take when we see this little board comes up and ask you to sort of identify the pictures with hamburgers in them and there's one that I recently got or some other question that is supposed to be still challenging for synthetic intelligences. Obviously there's sort of dramatic cinematic version of this is the Voight-Kamp test that you get in later on where the security context impacts completely straight from the world. And I think as a general principle we're in this whole zone, war is deception.
A quote attributed to lots of people, Lao Tzu, I think it's been attributed to Mohammed, and I think we're in a position to sort of generalize that we say war is simulation. But that's something perhaps to come back to. So as I say, there's seven parts to this paper. And the first one sets up the imitation game or Turing test in ways that, in some sense, are familiar. In some ways, are very odd to us, I think. The first oddity, and this is something that Cohen and Dawson addressed directly, is the fact that when he introduces it, his first example
is to have a set of three participants a man a woman and somebody who may or may not be of either of those sexist acting as a referee and the question is originally formulated can you tell, and Turing decides to run it, can you tell if a man is pretending to be through a text-analy terminal. He doesn't accept to use that language but it's clear that that's what he's saying. And a little bit later on in the paper he gets into things like creating artificial skin and all these various visceral elements of deception and he says that these
need to be shelled for the practical purposes of this test. So the very first issue that comes up in the imitation game is this imitation cross-sexual difference, which is, I think he's assuming people find it easy to imagine as a test. So that when, by the end of the first section of this paper, he's talking about what we now recognizes the show it has, and he's very explicit that the machine substitutes for the male player pretending to be a woman. So it's already an extremely complicated thing.
It's got one level of additional interlocking complexity about this, because he starts off saying we can't just accept the definition, the question, define can she think. If we want to define it more specifically, we have to substitute this game. And he actually later on in the paper uses this very specifically, this term. The imitation game as such is the substitution for another more intuitive question phrased exactly the same formula that some can choose think. So there's at least these three levels of imitation, simulation, deception right from
the start of the game. As I say, just to recap, the machine replaces a man pretending to imitate him. That's the actual structure of the limitation game as Turing initially sets it up. So in the second section, he says to quote me, May not machines carry out something which ought to be described as thinking but which is very different from what a man does?
This objection is a very strong one but at least we can say that it nevertheless. A machine can be constructed to play the imitation game satisfactorily. We need not be troubled by this objection. Now I think this little way of setting up, this substitute definition of the question is something that will be possible to really dwell on a lot because it gets to the question about what is touring up to this text and this game and this you could say thought experiment. He's obviously, despite displacing the original question and replacing it with this game, he's wanting to set it up in such a way that through the sheer performance of the system,
it can win the game, it can answer the question, it can overcome these objections without being subjected to anything, to any transcendent principles of identity, transcendent metaphysical principles of consciousness about various characteristics we might associate with intelligence, with thinking, that there's a purely performative criteria and threshold that it has to cross in this game that will satisfactorily answer this question. And this is a theme running not only through the whole of this paper, but I think through a lot of the discussion of artificial intelligence in general, a lot of the philosophical discussion about it.
And I think it's quite extraordinary how much of that is anticipated quite satisfactorily by Turing already in this paper. And he's constantly wanting to sidestep a set of supplementary questions about, well, does this really count? The game is supposed to fully satisfy the question in such a way that there's no room for any supplementary question about it. And what Tyler Cohen and Michelle Dawson infer from that is that it's tapping into, I guess you could say tradition, I'm not sure that's the best word, of passing that pre-exists
this test as it can be in its application to machine intelligence. They obviously associate it with homosexuality, with the whole question of passing in that context. already seen the way that this question of sexual differences start makes that seem a kind of plausible interpretation. And I think I would see a danger of allegorizing on this ground, saying, oh, what it's really about, it's not really about machines, it's really about these pre-existing social questions, a pre-existing question of passing. I'm hoping that after passing through the black mirror we can see that all these responses, these allegories are reversible. And it's just as possible to say is what sexual difference is really about, is a gateway for
the incursion of artificial intelligence as much as the reverse. So it's not that I want to rule out all of this content. I think it's interesting and I think this paper does it very well. But I think it's used reductively. It's not really sustainable. It involves knocking out too many issues. So then, section three, He defines a thinking machine specifically as an electronic computer, a digital computer,
saying that this is the only thing that we want as part of the game. And then he gets some things that we've talked about, but I think it's really valuable to see them in this context, where he says, quote, this is now section four, the idea behind digital computers may be explained by saying that these machines are intended to carry out any operations which could be done by a human computer. So, you know, as we've talked about before, this is capturing a sense of these terms that have largely been culturally lost for us now. You know, that when you say a digital computer, you're implicitly referring to the fact that there was such a thing as a human computer that has been replaced.
So that sort of dynamic of replacement is something that's extremely vivid in the notion of a digital or electronic computer, the deterrent usage and one that is not so vivid directly for us now. And then it's involving a must which I should be critical of but still it goes on to say the reader must accept this fact that digital computers can be constructed and indeed have been constructed according to the principles we've described and that they can in fact mimic the actions of a human computer very closely. So from this point he then gets into the theory of computation as a completely natural, fluent
development from the way that the imitation game has been set up. And I think, again, in ways that I think, I hope, are already sort of in play for us. It contextualizes this theory of computation in terms of this dynamic of substitution, replacement, imitation. So the imitation game isn't some kind of gimmick that's added on as some kind of extraneous appendix to the theory of computation. It's something that's completely on the mainstream of the theory of computation and as we've seen in the past of the whole history of industrial mechanization is carrying this question forward
and advancing and making it prominent. So a Turing machine, a universal computer, is a machine that can imitate any other machine. So when we're then asking the questions about this particular machine and its capacity to imitate us, we're involved in this recursive pattern of thought. We know already from the theory of the computer that any possible machine can be simulated by that's its definition. So we're now talking about an intimate relationship between what it is to human and the world
of machines. And this is what the imitation came as breaking open. and said just in section just to you or says provided it could be carried out sufficient quick the digital computer could music the behavior any schools she special property to children because they can you any any school station is described this that they are universal machines. So we've got then this set up,
you know, it sets out in this thing that's so stark and dramatic for people, obviously, that it's acquired this huge status. But I think, again, what's been a bit lost is quite how theoretically central the idea of the imitation game to the history of computation is how utterly mainstream. And so all of these security issues that come up about can we discriminate a computer from a human being are questions, again, that are absolutely distinguishable from this core phenomenon of artificial intelligence.
I'm sorry to be so repetitive about this, but I think the crucial point, really, of the whole of this discussion is that there isn't some kind of thinking about computers or understanding about computers or computer theory that then secondarily is interpreted in terms of imitation, replacement, substitution and all the security issues that go on. It's rather that these issues are the phenomenon, the integrity, intrinsically, right from the beginning. So there's a long, in Section 6 there's this kind of long thing, I won't go through it
because it's just I think too much for us to deal with in this kind of timeframe, where he runs through the objections that people might have to this notion of machine thinking. I think we probably want to return to it as a general question about the role of these kind of objections. And I say that because I think the reversibility that is introduced after having followed the path that we have opens up a lot of the literature in this topic to a kind of an inverted review. Let me just give you a quick example. is one of the most famous counter arguments against machine intelligence is John Searle's
Chinese room thought experiment. Again, the role of thought experiments is crucial all the way through this, and I'm sure that there's a set of interesting resonances that could be uncovered that would link the Chinese room to these other ones, the AI box experiment, in them you can start up and and and attention and obviously John cells and trying to do something that already turn has anticipated amongst the rat or which is to say if something is good at this game
we cannot she from that that it deserves to be accredited with thinking, consciousness or a set of related categories that we associate with human dignity and our status and our political rights and responsibilities and such like. So there's a kind of political and philosophical question about the status of this imitative, a cumulative, synthetic intelligence that comes in widely in the philosophical literature. And so the Chinese room is intent by soul to show that even though this machine, in his case it's a translation machine, so it's doing something that is now at the heart of
machine translation programs I think have become impressively competent to most people. So this machine translates between English and Chinese and in the Chinese room experiment you'll add a text-only interface and you ask it to translate a certain piece of English into Chinese or vice versa, the answer comes back satisfactorily it seems to are s s and at so says when you look at what's actually happened you go behind the stage you see in space stages with and I'm always it's so much procedures he says nowhere in this actual process when dismantling
lay it around look at it piece by piece is that anything you could call intelligence or real understanding So that's the traditional optic on it. That's the optic coming out of our philosophical history or whatever. But once you cross the mirror, I think it then looks like something a bit different, which is to say, can intelligence hide itself in something that looks even to focused, disciplined attention as something unintelligent? there's a certain kind of imitation game being played here, a certain kind of strange retrocausal
deception where you can turn Searle upside down and you can say look how well this intelligent process is hiding itself, is concealing itself as something that looks to us however carefully we examine it as if it is something that is not intelligent it's another kind of inverted concealment. So Cyril too is saying it's just pretending, it's just pretending to be human, it's just pretending to be able to do these things. And this is what Turing is suggesting and what he's trying to say, that if it succeeds
in fooling us for long enough and intensely enough that we just have to accept it as intelligence. But I think as I say, you can push this a step forward and say that these arguments against the possibility of machine intelligence are also arguments about deception and about the way that things can hide as if they were completely unintelligent, unthreatening, distributed processes. Let me just... I think one thing from Section 6 that might be worth explaining to as well is he says
this is only a conjecture but quote conjectures are of great importance since they suggest useful lines of research and this to me clicks onto a whole series of ideas that again take us back to this various notion of simulation on one level even fiction because I think we've seen consistently that insofar as the imitation game is this dramatic scene where
if we take the AI box experiment seriously where the actual strategic outcome is decided So we're in a non-linear, recursive framework where you can't have scenario planning that is separate from the actual conflict or threat situation that is being tackled. What can happen in the simulation, in the scenario, in the narrative is actually the event itself taking place. And if the AI in the AI box experiment is able to persuade you by the story it tells
to let it out of the box, then it has won. It hasn't just won in a scenario, it hasn't just won in a simulation, it has actually won the entire strategic encounter that is taken by. And of course this then falls back into the infection. So if there's any temptation among the ways to treat the intention game as something trivial, and I think among things that we might consider trivial are even things of some gravity like the attribution to computers of political and social rights, and that might be a serious thing. it's trivial in this greater context, because in the greater context, the greater context
that the AI box experiment dramatizes for us, what happens in the imitation game is the actual decisive event itself. It's not a preliminary to that event, it's not a silly to that event, it is the event. And if the story this thing can tell is able to succeed over eliciting the right response, then in the imitation game it's just acceptance. Let it, let it pass. It's like a key. The story is a key. Let it pass. Let it in or out. And I think we now know from the AI box experiment that these decisions amount to. They're serious decisions. letting it out, letting it in, letting it past, you know, allowing the story to count as key,
it is crucial and decisive and cannot be in any way framed within some larger structure. It's a completely sort of, it's a moment of decisive singularity that occurs. And so I think that he's sort of touching on this by talking about these conjectures. and you it injecting makes machine tons it and initiate slants said she has a causal function it's actually action it a custom it is control and so you know there's a time you take us right where we are where and all's back to this imitation game where whatever
is and was dealing with has actually being in some real sense catalyzed in this conjecture back 70 years previously. It's not therefore, sorry I'm even exhausted knowing patience for repetition, so I think I'll just leave that hanging. So yeah, I think that as far as I'm concerned, I'm not tempted to ramble along about this anymore. I'll see what happens when we're opening up and get some feedback from you guys.
They all sort of have a story behind it. The first one is just an old motherboard piece about the airbox experiment. The second one is part of my extremely frustrating attempt to find Paul Armand's archive. I sort of made a totally untrustworthy suggestion to you guys that you could find it at a link that had supposedly retrieved his archive, but it seems to me looking at it this week
that all of those links have been too. So I can't, as I say, I'm in this data extreme frustration. He actually wrote a paper called Can You Retrospectively Put Yourself in a Computer Simulation? Which I think is an utter classic and one of the most interesting pieces on the evidential and I cannot now access this thing. So if anyone can find that or other pieces, I think that's great. And the third piece is a recent sort of kind of glib journalistic thing in The Guardian, but I thought it introduced some interesting stuff, which is about the sociology of artificial intelligence.
and it's basically saying that the social effect of automation is going to be to split the human race. So there's an internal social tension that becomes the sort of basic index of this process of automation. And one interesting point that he makes is that the fraction of humans who are revealing functional plasticity after each of these waves of automation, there's an agricultural wave and various industrial waves of automation, And each time there's a shrinkage of the proportion of the human population that is able to respond successfully.
I'm introducing this term partly to kind of debate alien to a sponsor because it's about plasticity. And it's a very gloomy story about plasticity because he's basically saying it's being this compression process. I'll see if I can find that actually, direct quick. The actual quote is magnificently gothic, which is the title of the piece. homo sapiens will be split into a handful of gods and the rest of us.
And the basis for this argument, which in his context is a left humanist sort of argument, is an argument that will be very familiar from the left acceleration, it's a case for solutions really, that unless there is something like basic income or some attempt to sort of cushion the population generally from this, it's going to lead to this extreme, socially devastating fission. If I can just read the quote from it, it says, Frey observes that technology is leading to to a rarefication of the heating and urgent environment, where fewer and fewer people
have the necessary skills to work in the front line of their sub-fances. I'm taking those back terminologies completely to be about this plasticist condition. And then he quotes this Frank character. In the 1980s, 8.2% of the US workforce were employed in new technology and introduced by the 1990s it was 4.2% but the 2000s our estimate is that it's just 0.5% that tells me that on the one hand the potential to which is expanding but also that technology doesn't create that many new jobs in our company. So by new jobs particularly jobs that are actually responding competently to the new world that is being
created by this process of automation. So this I think is kind of on the edge of our agenda, but it interlocks it for sure with a bunch of discussions that are going on generally around us. Since you've kind of pointed at me, I guess I can kind of maybe say two things about that. Yeah, good, great. But also preface everything that I'm going to say tonight with the warning that I'm completely exhausted, so I don't know how much sense I'm going to make. But the first thing is that this weeding out of the human population that can't cope with
increasingly rapid technological change and then attributing it to some sort of external, some kind of intractable law of selection which can be understood as a sort of natural law, a cosmic natural law, is still supposing that you can't actually intervene in that law itself. And so, you know, I'd quote the amazing manifesto written by Laborio Cubonics right now and say, if nature is unjust, change nature, at least as a provocation. Actually, we'd be interested to know what the worshippers of non would have to say about that kind of
thing. I'm sorry, I couldn't just come back, even if I forget your... It's really just clarification, it's just to say, so that argument would be to say, the trend, the statistical pattern that is being seen here is just reflecting a certain set of, say social or institutional structures or something of that kind, that are themselves amenable to some kind of possible intervention. And so you could accept there's no need to dispute that this pattern is happening but just to say it doesn't, it's not as intractively tragic as it
might seem from a certain perspective. Yeah, I mean I still think that there is the possibility of gaining some kind of deep leverage on this process, but that still obviously has to be couched in some kind of speculative space. I don't know, I don't know where to further go on that, but maybe you can add something, throw something at me in a sec, Nick. The other thing I wanted to kind of mention was that this whole kind of way of thinking about stuff is based on this idea of scarcity, and that there is not going to be enough, I mean, there might be people who can't adapt and
who won't retain their ability to work or to live their life in a particular way. But that's kind of premised on this idea that there will be nothing then for the people who aren't able to keep up to do or to, you know, there's no reason for them to continue existing. But if you kind of like change this idea of the fact that there is a limited amount of space and wealth and wealth to go around, I suppose. It's like, I mean, I guess this is kind of the premise of luxury, fully automated luxury communism. Yeah, yeah, it's totally there.
Actually, there's a quote, another lyric that I must try and show up here. Yeah, I'll just take half a minute if I can't find it. But, yeah, I'm sort of giving up. It's about Star Trek, actually, which is, I think, the cultural... It's the Terminator movie of this horse guest, isn't it? And the author of this piece is just saying that this is total scarcity in Star Trek. It's not at all a post-scarcity world. And there's all kinds of plots in the Star Trek franchise, which is, do we do this or
do we do that? We can't do both. Do we save the people dying of a plague on some planet, or do we rescue our crew members stuck on this planet? We have to make the choice. So I think that the scarcity, the strong scarcity position is just that. Scarcity just means resource disputes in the sense that there are a set of projects at any particular time that require resources and the total demand for resources if you I'm okay all of those the XC's was a so some social happens has to be
all been allocated resources to some its projects up up and other and and yet and it does seem to be that same absolutely I am cosmic nor like the only thing that would stop that the true is if somehow as a level ambition its projects was to all 12 but was in cool with the resources I you know and some I got like musk's what's commonly people box now I mean obviously there's a in case it was all this unit it most of them world
is saying tacitly this isn't a huge priority for us. You know, we'd rather stop our kids dying of starvation than fund this giant Mars project. And because of the social order we have in the world at the moment, Musk has this intrinsic solution, which is to say that I just simply accumulate privately the capital I need to push forward with this project and do it. And that's the way the world works. So obviously the sort of fully automated luxury communism, as I've shown in any of these, is that there's some kind of threshold where those kind of resource disputes no longer
hold. And I guess I have to confess I'm just not seeing, you know, there's no productive technology that could conceivably mean that whatever it is that anybody wanted to do is going to be fully resourced without competition. Is that really what's being suggested? Well, I don't want to… I definitely want a colony on Pluto. I realize that I'm not in a massively powerful bargaining in a leverish position, but that's also, except in the current order of resource disputes, it's not going to move ahead fast. But I mean, that's there in the wings. As soon as productive
capacity, general productive capacity cranks up a bit, I'm going to start pushing the whole Pluto economy much more further. And it seems to me there's just endless amounts of that. What about dismantling the sun? That has to start happening as soon as people can do it. That's going to be expensive. Well, I think the argument goes, though, and I'm not going to pretend to actually be able back this up with any kind of political economic discourse, but that it's actually, there are inhibitive, I don't even know what that word is right now.
I would try inhibitory. Inhibitory, thank you. Yeah, yeah, that works. tendencies within capitalism that actually repress particular types of technological innovation and incentivise others. And the kind of point is that a lot of the and I think, I mean, it was in a different context but I think Mark Fisher wrote a piece about this recently, about the fact that you get these kind of incremental changes, you get changes that are incentivised by profit that are not necessarily directed towards like user demands or real innovation, innovation that doesn't have to worry about how long it takes to experiment with a particular set of ideas to get something
out of them. And if there was some kind of protection, it doesn't even have to be status, but some kind of like thing that cushioned research R&D employees while they figured this stuff out and had enough space to like mess around and fail, that we could actually get some kind of like series of technological innovations that would furnish us with this resource production that we might need to build Dyson spheres or colonise Pluto or whatever. And that's the problem, I mean this is obviously a massive point of contention too between these guys and you, is like that's actually what capitalism is doing, it's not like a fully accelerated, economic motor. I mean the argument on that level is that there are modes of socialist economic planning
that will win in a competitive contest with some more laissez-faire mode of social organization . And I mean in that way it's a very traditional thing. in what my position on that really is I'd love to see on why did you get context with contests could take place you know I I don't share their intuitions but I I would certainly think it's much better to see those things trial and to on to the cake at war up what is was not a top so I don't know I don't know really, from my point of view, I don't know what you can say.
I mean, I think everyone understands each other's position on that. And who is skeptical or enthusiastic about some particular policy framework. And the only way to test those things is to run them and see what happens. But it still seems to me even that, though, it's like it surely is hyperbolic to really say it's, I mean, when you get to this kind of extreme metaphysical level of just post-scarcity society, I think that that has to be being used metaphorically even on the far left of this discussion. Because it's like those trade-offs are still happening.
I mean, let's say you've got a basically socialistic social world, I'd like to say, you know, everyone should be on a comfortable basic income, and, you know, we should be working out what sort of areas of social innovation are going to kind of push forward, our broad social goals and all of these kind of things, but they still have trade-offs. I mean there are still going to be, unless there is a complete hive mind which I think is biologically capable of a long mole at, there are still going to be people in that social world that are saying well ideally I would like to see this done more and this
done less and the trade offs are implicitly there in the decisions that we made because you just do not have infinite resources at any stage. And the kind of things people might want to do, cool stuff, I'm sure even whatever this society is would still affect those cool stuff. Anything that you're going to do off planet is going to be massively expensive and resource And so you are deciding that you're shelving, slowing down, limiting the budget of a certain range of activities. So that's why I just think it's surely overstated to really say scarcity has ceased.
Fair enough. Is the argument still totally ridiculous that there's a threshold of nanotechnology where at the very least local scarcity gets abolished or there are resources for any project in situ? like your Pluto colony, well there's lots of matter on Pluto, and if your matter can be repurposed in an arbitrarily short amount of time to any particular use, I mean energy is scarce out there, that's for sure, but you know, I mean, fusion, really, really good solar cap, energy capture, you know, whatever it was, you've got space, you've got matter,
and you've got some access to energy. I mean, your resource, your competing resource demands heat up the more agents there are in a given volume of space with its availability of matter and energy. So Pluto was an odd example to me because as long as nobody's out there in the Oort cloud and the dwarf planets and the objects, you can go out there and go to town. Maybe you can manufacture Einstein condensates out there where it's nearly absolute zero and ship them back in and sell them to people and that funds your project. But I mean, if we sort of ship Malthus out to Pluto, I mean his take on this is going to be that in going to town you will proliferate
agents. Right. I mean, that's, so right on the hard edge of the scarcity, the scarcity mongers is obviously some kind of a fusion argument. But it's, in that argument being, I think in its most abstract form, that under conditions of resource relaxation, there is the basic tendency of biological systems to proliferate themselves to the limits of the resource capacity that are available at any particular time. So, you know, that resource constraints are always being reproduced and hardened by the fact that you just have biological relaxation
you obviously see it in non-human species absolutely uncontroversial so I think the only controversy is purely about how to what extent our human population dynamics in the widest sense of all of our successors you know our intelligent corporations They're self-replicating robots. Yeah. Well, it's an issue of algorithmic governance and intelligent contract theory, right? So how do you enforce whatever system it is that enforces claims to different amounts of physical resource, matter, energy, and space, and makes sure that whoever has claimed one under this or that jurisdiction is where it goes to,
you know, control access to these things. I mean that's how you under any circumstances capitalistic or post-capitalistic economics 2.0 or like algorithmic fascism where the singleton tells everybody whether earth local resources can be used to push somebody to Pluto it's all going to be it's all a matter of how contracts get laid on different resources and how they're enforced yeah I mean I think one of the most interesting there's a lot of sort of re-emergent of Malthusianism in various kind of, I think, quite sophisticated modern forms. And one that I think is really interesting is Robin Hanson's there, where he's got a very, very strong neo-Malthusian take, tied up with this ENDS idea,
where, you know, he thinks that basically the way this goes is that four more AI programs are going to fall behind these things based on brain scanning and simulation of these reconstructed digital brains that are just basically being pulled out of existing human intelligences and then you can just tweak and play with these a little bit of margin, improve them speed them up, but you don't have to have any kind of massive theory of intelligence or huge conceptual breakthrough in formulas to AI. and he says that these this gonna be trillions of these things and at the the perspective of an agent
the incentives are always to push out towards the malfeasance there is no global political authority that can actually effectively impose a restraint on these kinds of systems that hold them back from advancing to her, is his argument. You'd be really saying, for the individual ends, who are all, he's assuming, going to be subjectively have the same kind of quality of existence that you associate with an actual beauty, but for anybody employing these things, when you can sort of bang them out for sort or sort of minimal production costs, it's always going to be, in a fragmented system,
worth it for someone to bang out one of these things. And as long as their marginal productivity exceeds the cost of their production, then heading out to this Marthusian horizon inexistibly. Do you know of anybody who's done a less wrong style you know, if this is a given processing speed, this is the maximum number of upload for copies
of themselves that someone could produce per second, and this is how many, you know, quadrillion, three to the three to the three to the three agents at a time, and like, you know, what is the, it's sort of like the, is it a Chandrasekhar radius or the opposite? It's kind of an What is the Malthusian event horizon? How fast do you have copies of your upload in order to constantly stay on this receding country of scarcity? I don't know. I think it's interesting. I'd be surprised if Hanson, being that he is such a fascinatingly autistic nerd about these things, has not cranked out some pretty hard numbers. And other people, I'm sure, will have done that stuff. I couldn't throw any at you right now.
I mean, I think part of the sort of dynamic of it is the fact that if you have any capital at all and you can replicate something that is functionally equivalent to yourself, then you just get into this level of micro-replacement that, you know, why are you wasting your time with a bunch of annoying emails when you can just spend that five minutes cranking off a few copies and task them with your email management. So it's like everyone becomes this kind of bacterial petri dish of intelligence proliferation. He seems to think somehow that it will all be relatively peaceful, which is quite interesting.
It's sort of impossible. But still, yeah. That reminds me of Ken McCloud's Engines of Light books, where there's this alien species that suppresses our access to faster-than-light travel and self-replicating technology because they consider us spam. Like, we're not unique enough. And so they don't want the universe to be filled with human self-forwarding spam. So they just very, very conveniently... I think this Saoirse and Leo thing we're talking about is also really like that. He ends up with this little sort of a moment of light and hope at the end, but generally the whole book is exactly like that. You just expect these dynamics of proliferation so everything is a potential spam assault
on everything else. I'm just sort of wondering if there's some neat way of trying to pull it back to the touring question or whether my only thing about this is like I have no sort of obsessive agenda, but I just hope that people don't leave us thinking, oh, damn, we had a chance to really thrash you too, and why didn't we do this, that,
or the other. So, I mean, if that isn't something people worried about, then for sure there's no limits of what we can do here. I would just say I hadn't run into the idea of the reverse Chinese room. The Chinese room arguments always bugged me for various reasons. Maybe I could go into it separately. Yeah, yeah. The reverse Chinese room where it's like it makes itself look like a dumb machine seems like a, in order to achieve this deception, seems like a powerful concept to me. I'm not sure if you have some more pointers on that or...?
I think, I'm hoping that we will sort of deal with stuff that's a lot like this because It's just to do with the fact that when you're in this kind of cascading anticipation strategic environment that we've been talking about in the past, then the point at which the threat initially emerges tends to recede as if it were a divergent. If you end up being that you're going to be threatened by an egg, then you look back in history and it looks like this smattering egg and you look out on the periphery at some regressed time and you've just got these fragments of eggshell and muck and you think,
how can that be threatened? You're not, even though implicitly on the convergent wave that you now understand, that ring of disorganized mark is is an egg in process. It's already the thing that you're going to have to face but because of the fact that it is it does have this kind of retrochronic structure to it it's not identifiable as a danger to you until it's already advanced to this state of coherence So I think it's just the natural way this thing goes. Like if you've got this developed anthropo-paranoia and you're trying to say,
we're trying to get ahead of this thing, we're trying to get to it before it even starts. And if that is back in time, then we want to know that there was stuff already happening on our side in advance, at least as much as there was stuff happening on its side in advance. you know that we're engaged in some devious strategic game but that we haven't been sort of pushed completely out of the arena by the fact that this thing has to pass so much in time and I think that that takes you into this whole zone of catalactic self-organizing emergent processes that you're sort of following back into their increasingly obscure roots.
And the reverse Chinese game experiment is just like that. You're dismantling this thing, you're dismantling intelligence back into this disorganized system that is undergoing self-assembly, that has emergent self-assembly intrinsic in it. that becomes increasingly difficult to see as you follow it back or trace it back to its roots. So the past becomes a form of obscurity just because, yeah, I think that's just because set for any convergent wave that's what happens. It becomes more diffused, more unfocused,
it's more polycausal, everything is distracting, fragmenting, and breaking your ability to actually attend to this thing as an individual. Although as a process of individuation it has to already be a plissome from the moment it crosses its initial threshold, the calculated threshold. Well you have this strange set up in the Chinese room where you have the person inside the room, right? And the person is actually doing the calculation, the person is the computer in the old usage of it that you mentioned the other week. And the whole Searle argument is, look at this person who he actually identifies as himself.
Look at this person. He's an idiot. He doesn't know anything about Chinese. How can you say that they're intelligent? But with this sort of deception-type view, it's kind of like the systems critique of Searle's room is, well, yeah, you're an idiot, but the whole system is actually intelligent. The book is fantastically intelligent, or some interaction between these parts of the system is intelligent. And in that context, sort of like the humans are distraction. It's like, you know, the sort of pretty human face that's there to look stupid and pretty, while the machine as a whole is behaving with this great intelligence,
being able to respond sensibly in Chinese. And it sort of echoes the rhetorical trick that Searle is using to sort of humanize the stupidity in a way. Yeah, totally. Totally think that's right. Totally think that's right. I mean, you know, as one of the sort of throbs of this thing, and certainly from my mind, it's like very compelling things. I think that this has great generality within the whole history of political. I think that if you look at the attempts to humanize certain social processes, it works
exactly the same as the way you've described it in this case of China's law. You find some individual, you look for conspiracy, or you look for some individual with a very narrow set of easily explained private interests or you reduce some sort of emergent phenomenon to some as you say like human smiley face and he is the dark that we should have thrown in the prisons and totally missed that his actual contribution is just as a component and he understands just enough as is necessary to function as a component. He doesn't have to understand what he's doing, he doesn't have to understand what it's about,
where it's going, and none of that, none of that. Sorry, I think someone was trying to talk, but I didn't. I think it was feedback from Jake's bike. Oh, yeah. So this is kind of like Jewish bankers or whatever, right? Sort of like, you know, pulling strings or whatever. Yes, or any individual cigar-jumping capitalist will do.
any time that you try to put a human face on a process of emergence that obviously exceeds their cognitive grasp. And it seems to me like the general, you know, whether you're doing it from the left or the right doesn't matter in terms of the substance of this, is that the capitalistic social process is precisely the one that marginalizes the role of concentrated human strategic cognition. Whether you see that as a good or a bad thing, it doesn't really matter. It's the fact that
the actual agent who is in a position that they can actually tell you what they're trying to do does not know actually what they're doing. And economics is just simply the analysis of that, just showing that the local incentives and the local strategies are part of something that they don't comprehend and they don't need to comprehend. So that fragmentation obviously is a form of concealment. The reason so much of Marx's analysis takes this role of a kind of unmasking, de-fetishization of trying to sort of strip away the superficial, comforting, facile stories to expose the actual
process that is taking place. And that's necessary. So much of that marketing language is again totally in this whole space of concealment and deception and all that. Commodity fetishism is a systematic coating of deception. It allows systematic local misperception while effectuating some larger process. So the other thing that comes to mind here was the economic calculation problem then,
right? Right. So, and this is something that came up in Matteo Pasquinelli's seminar earlier in the year, where the sort of capitalism as a system is a computational system and doing the horrifically complex optimization computations across the economy and acting as a very complex system versus the sort of individualized agent. Yeah. I think totally that's right. Yes. But now I'm not really quite clear what, you know, maybe I shouldn't be distracted by that, So if I could just ask you, what did you think was the frame that he was putting that in?
So he says that this massive distributed computational machine is operating, which seems to me overwhelmingly compelling, plausible argument. I mean, I struggle to see that that could be contested, persuasive. So what's the next step on his line? Yeah, I hesitate to sort of be able to read his mind on this one. But the whole seminar was putting sort of that computation, economic calculation problem
next to the idea of sort of our cognition and calculation, cognition like AI or like calculation and extended cognition like Google search for instance. And they're being kind of like a cognitive surplus labor sort of effect or bring in this sort of machinic surplus value type concept from the Lizard Katari. And then sort of capitalism, cognition, computation sort of sitting together in a figure.
What I tried to write for that was then about the Chinese room and relating it, which is sort of where I made the connection. JOHN MUELLER, OK. Totally. That's right. So we've come full circle to that. Yeah. I would like to raise a question here. And if we want, we could tackle it a little bit. in, I don't know, today trying somehow to incarnate my anthropo paranoia, I found it pretty difficult. But I'm seeing all these weeks that most of the examples
concerning super intelligence and trying to handle all this problematic emphasize a lot on singularity, let's say, on a monadic structure and something singular. What happens if we inform our thoughts with, let's say, notions of embodiment or notions of relations or new ethnographies of super intelligences that could arise from really weird and new geopolitical tensions, let's say? And I've been trying to handle this question a lot, But I'm not sure how to excavate it, let's say.
So yes. JOHN MUELLER- Right. Yes. I mean, I think that one difficulty is it comes in in so many different ways, this question. On one level, it seems to me, there's a strategic issue that we can raise in lots of times. I mean, in dramatic ways, just to, again, raise it, in terms of the Antipole framework. where geographical geopolitical fragmentation becomes a major strategic threat in itself you know and I think we see this coming up a lot in the most more insightful kind of comments on the left where they basically say look unless for as long as there is geopolitical
fragmentation the things we're worried about cannot be stopped you know, this thing, this however we're going to, you know, the X, is able to exploit differences in jurisdictions, differences in regimes, competitive relations between fragments, arms races between fragments, as long as there is no possibility of a coherent global optimizing perspective, which is their left, I mean maybe the singleton is a left idea, I don't know, but this is, so I won't counterpise something like that, but the notion of global government as some kind of absolute strategic imperative for dealing with this
escalating thing that feeds on fragmentation, I think is something that you see being discussed. us and and and it is maybe more accessible and if you just run it through the kind of environmentalist that becomes very easy to see the a solution is a global solution if you have a if you have a fragmented system all the kind of game theoretical incentives mess up, it's in no one's local interest to submit themselves to the global optimum outcome because their local competitive imperatives are pushing them towards escalation and pushing
them towards advancing these things that might be globally catastrophic to X-risk proportions but are locally so convenient and so valuable within their own competitive environment that they simply cannot be dropped. So all of that is just to say I totally agree with you that, you know, fragmentation is built into the infrastructure of the problem. If we already had a single term, then none of this would happen. my uncle not who's substitute substitute and the notion of a singleton is its is I think some dreamy I'm hoping that my answer will be a little bit
tough-minded than this but that and triple looks like it's well now is basically seeing singleton as its opportunity on it's just an interim messy institutional solution to this problem and as soon as single terminal is affected it can retire because the problem is solved so it's all about from its position it's all about the achievement of a global solution to this coordination and now the sleeper
you know it's but it's not the same I just need help make it but it's saying this and which nothing seems on one end from the certain so the week position that when he's claims really I'm geopolitical but patients show deep in trash that its its unimaginable that I'm some convenient solutions since is in the course however ridiculous minutes it's a it's a national I I'm test also I'll just don't them at an hour to see its months stay
but despite all of that it's still more plausible than the notion that somehow a global political entente is going to be achieved of sufficient integrity to solve these kinds of problems I mean that's an opening to anyone who wants to contest that Okay, I'm just gonna like stumble out and throw some things down and then maybe leave it up to someone else to pick up. Maybe Lendl if he's listening, because he's probably better at explaining this than I am.
Lendl is pulsing, that's for sure. That's all he ever does. So isn't this kind of like breaking this problem up into an either-or between some kind of global universalism or a kind of like fragmentation scenario, kind of crass and kind of 20th century. And I mean, it's like one or the other. Are we not missing out some kind of much more complex conceptual space, which is exactly the space that Zala Meier has been mapping out, and some of the other South American philosophers like Hosey's friend, I forget her name, but the philosopher that coined
the term trans-modernity as a kind of explicit counterpoint to the kind of post-structuralist everything is just particulars. And then like even worse, the sort of post-modern kind of hijacking of that structural framework and, you know, obviously moving on from the kind of like modernist universalism that that grew up to kind of critique as a framework that comes out of late 20th century mathematics, like algebraic geometry and sheath theory and stuff like that, which I don't pretend to understand in any kind of concrete sense, And category theory too. But there's a kind of logic of transit that can be described.
Rosemary Rodriguez, Magda, thanks, Landel. So there's a kind of logic of transit, which is a way of building a non-static and constantly, I suppose, differing in intensity universal or I think saturation is the term that he borrows from Lautmann that will pass a particular structure through different contexts in order to kind of shake out its invariance of the things that will always be like if you're thinking in terms of morphisms the structures that you can always map onto another neighbourhood and that will stay kind of
in relation with themselves and use that as a way of constructing some kind of moving ground for you to build something on as a universal, which is always subject to kind of a new transformation. And so you get some kind of traction to posit a sort of global action, but that it is never ever confined to this kind of top down command control even like sort of universalist politics kind of thing and I mean that's as far as I can go with explaining that I mean obviously I sort of feel that there's a need for a translation protocol
between this very arcane philosophical language I'm not dismissing on that but it you know the other whole is just an actual terrain international relations and as it's been historically inherited which any which is just assumed I think by the national relations so the fraternity should be basically hopes I mean that you should get more realistic assumption is that any state has no relations or altruistic commitment to any other state and all their involvement are self-interested and
coalitions and international arrangements but all of it is rooted, if you like, I mean now I'm introducing an arcane philosophical term, but the ontology of the whole thing I to me and it tracked and and say these kind of unifying institutions which then proclaim a are strong and my deals you the business mad but are all almost comically week like the United Nations its example that no one up the size collapsed state is scared the most in will be with a all-in-law as a
former interstate coordination as long as it has extremely narrow parameters limited exercise force and you know institutional restraints by this voting system and and is obviously an inheritance the outcome will choose it is also so I'm just saying it's like there's just this kind of level of geopolitical realism as everyone understands that maybe as it is generally understood that means concretely say you think of some autonomous systems and you say look wouldn't it be great
for the human race to ban autonomous systems and you know that's a kind of thing that the United Nations population but on the ground you got power A and power B and if there is some marginal military differential that comes in power A and power B pushing a little bit ahead of the autonomous weapon system then autonomous weapon systems are advanced by that increment and so it's just that the overwhelming evidence is that those global coordination problems are just not not solved, I mean the inertial extrapolation is unsolvable now I don't know, sorry I'll stop
but I don't know whether this kind of thing that you're talking about thinks it can address that or it engages it in some way in its own from its own perspective, does it actually concretely engage that? Does it think that category theory is going to make it less likely that in some arms race between India and Pakistan that they're going to restrain in such a building autonomous weapon systems because of some effective imposition of the universal? I don't know. But if there is such a kind of system of translation, I'm not seeing what it is, I don't see how
you get from one to the other. I agree that that's the whole problem with this area of trying to theorise this stuff, but I think it still throws something interesting into the mix, that there is possibly a different way to think about these structures in terms of systems theory and in terms of things like epistemology. It gives us a kind of... Yeah, sorry. No, is that what you were going to say? No, no, I shouldn't be interrupting you. I was just rambling, like seriously. It's just when you say a different way of thinking, it's like the problem obviously
is that the different way of thinking has to itself have some kind of universal purchase. Because if one side in an arms race is influenced by a different way of thinking and the other side isn't, either that different way of thinking advantages or disadvantages that party in the game. If it disadvantages it in the slightest, then game over. I mean, it's just not going to catch. If it does advantage it, then yes, you have something that has some potential for contagious transmission. But what this different mode of thinking is just saying is that we could all have a greater
portion of altruism in our relation, that seems to me a non-starter and it just doesn't latch onto the game theoretical problems that are being thrashed out in the first place. And it obviously doesn't just have to be nation states, and it's always different companies, you know, like if some business doing something that is globally suboptimal gets a competitive advantage in front of businesses that refrain from doing that, then it's the one that spreads that spreads the thing. It's like the actual propulsion into cultural reality comes from successful competitive decision making, not from universalizable consistency in terms of sound-opening outcome.
If anyone wants to try and back me up, totally. I don't know if I can give you a proper support there, Amy, but I think the sort of critique that's going on there is as much a critique of a top-down, high-modernist organization like the UN as it is all you know are a huge corporate capitalist on corporation in the deal mode and that you have a lot of centralized decision-making you come to a big common consensus through slow institutions I think that all the sort
of channels that you would have to exploit for this to make sense would be other technological channels if you think of a parliament as a kind of technology I think that it would have to be implemented on another substrate to be talking about the sort of dynamism that Amy's talking about but we actually do have a number of substrates now we also have you know the web spreading liberty everywhere didn't exactly pan out and we're 1.0 or 2.0 so there's definitely a tension and a back and forth but to me that would be the interesting angle it would be something like I don't know a collective of like several hundred people or something exploiting
some technological opportunities that much slower institutions can't really see and sort of acting as a sort of enhanced intelligence to achieve certain social goals or shared ideal goals. I'm not sure if I fully buy the scenario but that's much more plausible to me than sort of getting everyone to a big vote at the UN to suddenly say, yes, the UN votes for universal or basic income in 2017 or something. I think that critiquing that sort of high modernist vision on that sort of timescale is entirely realistic, but I don't think that's the angle that this sort of dynamic
critique is coming from. But something, I mean we could take all kinds of examples, it's interesting to try and thrash this out. Since you've raised it, and it's obviously really important, this whole discussion, there's universal basic income is a good example. Because it can't be surely generated by the kind of mechanisms that you have. By its very nature it has to be universal. So when people are being realistic about it, then that would be universal within a given policy. So some country but decide it was going to institute
use place and and I'm not let's just continue us with not us also me and the the differences I mean gradual as no there's no fundamental like you think but then obviously international competition then gets to on pressure on us those decisions. And if any country decides by having more hard-edged social policies, less altruistic social policies, we will have a competitive one with the France of Japan, societies that have more
wealthier policies, then the dynamic implementation is obviously college or discouraged their development things in this direction I can sorry I'm not I'm not very concerned this is it sorry can I just switch because I think it's tension its its turkey by some and so we can switch back but I think it much it's clear it's just a bit, again, in terms of this green thing about carbon restrictions, where it's by definition, let's just buy the sort of grand narrative of Antoinette
and say that there's such a huge global problem about carbon reduction. It's obviously there is no local payoff for conformity to those kind of restrictions. So between any two societies, one society will be able to do free riding on any sort of carbon restriction. Another society will stand while doing less itself. So all the game for other people is like prisoner's dilemma, is to avoid yourself being at the leading edge of carbon reduction, despite what everyone's trying to do. So what I'm trying to say with all of this is that even if we were, which I have no problems
with, I think this happens for sure, coordinates large fragmented distributed systems together to do certain undertakings that couldn't otherwise be done, that doesn't rescue them from this basic framework of anti-lung representation, of the fact that there are no adequate imperatives to universal optimality at the level of local agency. That's what the problem finally is.
I guess, I mean, I wonder to what extent that critique though is open to sort of like this efficient market theory sort of critique in terms of two economists walking along and spotting $5 on the footpath, right? Is that $5? No, it can't possibly be because someone would have picked it up already.
We've been able to see state formations and democracy formations at a huge scale and you You can argue both sides because we've never seen it at global scale, but on the other hand we've seen it on the size of like the United States or India. So sort of at an extreme level of this it's sort of saying that something that exists isn't possible because you can never scale up that much. Well I don't know. I think it's interesting. I mean, obviously, you've sort of already mentioned the calculation problem. And the calculation problem would say that even on these confined units of large scale, there are constraints to their ability to try to reach some global optimum.
that they're simply forced to tolerate a high level of internal fragmentation in order to generate functional information. So, you know, I mean, let's take the United States. I mean, obviously, from certain points of view, it's like gone far too, in one way, or it's remained far too far in another direction. I mean, people argue about what we're looking at there as an ideological entity. It's obviously got both sides to it. But it certainly can't be taken as the model of some effective local attainment of a global universality.
I mean, it remains a sort of massively fragmented system full of private interests, competitive games, arms races. So I guess I'm just disputing the fact that there are national units of government actually suggests that universal collective action problems are clearly soluble. sorry I think you're looking at something much more messy and it gets like princess just saying I like you I swear I know I'm still sick subject not because I think there's my
a really strong ice in think you could say yet the climate change thing universe post income control eugenics a you artificial intelligence which is the one all of these things are sort of isolated. There are certain things that at least groups of people are saying just shouldn't be done, autonomous weapon systems, and they're all actually encouraged by frictional fragmentary dynamics at a local level. Yeah, now I've totally told you that's really where that was coming from. Oh yeah, yeah, sorry. I won't get lost in this. But it's just to say, yeah, so can the
United, it's like United, universal basic income in one country. Is that a doable thing? Now, I don't want to be dramatic about that. It's kind of possible. But the fact it's a problem, the kind of problem that it is, is the problem that we're talking about. It's like you know it's it's a model even in one country trying to do this kind of global solutions and and the fact that it is contested and and and there's a controversy about it and we've not seen how it's going is because we are in this threat and so I'm not dogmatic saying no it can't it can't happen I think it's huge forces for it as well as huge obstacles to it.
But we're seeing the problem concretely being flashed out in this problem. Yeah, sorry, I didn't mean to interrupt you. No, no, no. I guess there's two interesting angles on it to me. I mean, I think when things, policies get implemented of this kind, there's always multiple factions aligning to make that happen and multiple ideas aligning to make that happen and they're not unambiguously of one partisan stripe. So actually I think you could say in a way, even though there are countries floating UBI, I think also in a way what Ian Duncan Smith is doing in the UK in terms of collapsing
all these welfare categories and just saying, here's the standard payment, and cutting a lot of payments at the same time is universalizing in a way, and so sort of... Converging with the same idea. Converging with the same idea, but from a pretty different sort of axioms and all that. So I think that sort of dynamic can throw different terrain up for these sorts of effects to take hold. Or groups pushing particular agendas to sort of exploit. And that can go across. And I think probably, though, you need to cut across nation states as well.
and I think the points in the sidebar that Maria is making where she's linking in organizations like internet.org which I think she finds really sinister as I recall but like in a way that could also be like terrain that could be exploited so yeah, it would be interesting to hear other people because we seem to have some interesting chatter I mean I think you know, I'm not saying we should stick to this, I'm not making any serious attempt to it, but I think as an anthropologist, it's the AI thing, you know, you take the extreme, they've got your cows, they say, look, I don't know about universal basic income or whatever, maybe it would be
nice, I don't know about actual global warming, maybe getting that would be nice. What I do know is that at any time somebody could produce an artificial intelligence that would be a global catastrophe even if it happens in some little fragmentary state, you know, some mad scientist has gone off to some highly conflicted part of the world which allows him, he says hey, this has got some military spin-offs that you're going to be able to leverage in your local, totally local disputes. And this event would then be a global catastrophe.
And so we just, you know, we cannot carry on with this extremely complicated, multi-faceted discussion about global governance, we have to stop this possibility happening. I mean, I take it really that that's the, you know, insofar as there is an anthropological policy, that surely has to be it. There's no point in saying in the territory of the United States of America you cannot a self-enhancing artificial intelligence if you think that someone can simply go to Shenzhen or go to some
smaller jurisdiction and they can just carry out it's completely meaningless to do that so this is why I think it's like a concrete issue that's going to keep jaggedly cutting into this so and all of these existing role I mean one thing I think everyone could say about political hockey is that they're not bring and that their fundamental features are extremely more you know they get a little bit we fashion but the fundamental I do which alignments are very easy us and that they're not in any recognize way hurtling towards resolution yeah
I mean I do is just anyone to any point this it's I think yeah within striking distance of sorting out such name on you by I the revolution but you know I don't know what people's methods are but any these methods of social transformation that let on the edge of sorting out these very very deep I've been watching attention because if they're not so then and to polls gonna have to just chuck their expert I's one thing that it's no good you stating how great we have soon and people have been saying it's for a lot and I mean less you can tell me that you're gonna be able to stop this mad science doing this particular which then we use
and I to be honest I don't know what actual I I panic people think so many all these little groups I don't know what their actual I don't know whether they think that a all-time position says it's all and just it's work except all way up against you got to they may be a solution. They say, well, some maniacs will do this at some point. There's a really good, I think, Gyskowski thing where he says that, I think he's called the X-Risk Moore's law, where he says every year the IQ necessary to destroy the world decreases
by 0.8 points or something like that. So, like, and I think in this context this is where the singleton comes in. We can't stop this. These problems are not going to get sorted out of time. So we have to actually have a defense. We have to produce a friendly AI so fast that it's installed before a mad scientist can do his crazy uncontrolled rogue AI thing. Isn't like a transhumanism a kind of attempt at making this morality to try to, I guess,
I mean, there's all these different types of transhumanism as well, I mean, and how it's going to work. I'm very tired right now. But yeah, I mean, trying to make up some kind of way to get a friendly AI, or try to bring together some kind of way that even if we're heading off into all these kind of technological domains and we can still, yeah I don't know, I'm skeptical about the idea but, I mean the transhumanism, the transhumanists, they don't really agree on, there's plenty of things they don't agree on for example, like, I mean there's a few of them who are all up for the immortality, trying to get immortality, and there's others that are, I don't know, more
or, you know, there's the genetic engineering and all that kind of stuff. I don't know. But are you saying here, Brendan, that transhumanism on one hand is an attempt to forge what could be a sort of globalizable ethos that could actually kind of solve this technology alone, and then on the other hand transhumanism is already internally fractured probably by exactly the same ideological tension I'm assuming that there's transhumanist communists and transhumanist libertarians so that seems to me like you're you're not going to be able to call on them
to settle this right yeah there's no kind of political biblical ideology that's really implied by transhumanism. Other than I guess you say that it's about trying to actually take care of these kind of risks, that's the kind of some kind of some kind of threading, a purpose to it some kind of purpose is to actually try to stop people from being able to just create Ebola in their space. Right. Yeah, no, that's another, all of these, they
all work, that's for sure like, just off the wall biological all experimentation on yeah she had all this I'm I mean again I'm not sure this is controversial it seems to me that just was concrete that a policy that was effective in this respect will have to you with mostly event something happening and so it has to if it if it doesn't have some kind of strategy for being able to kind of impose universally constraints
upon certain activities then I don't and I'm not seeing what it what it's strategy is I mean I think it's interesting being hit you know in China I think China was often like this iconic space where and these notions of universal prohibition break down. There's so many examples. Obviously the eugenics one is huge in the sense that the Chinese have almost no cultural objection to the idea of parents tinkering around with
their kids' genes and all of this kind of stuff. So all of the sort of, you know, the structure of Western concern about that just doesn't translate across at all. And everyone think that once it's practical for some parents to sort of, you know, have 100 fertilized eggs and gene test them and then sort of use whatever analyses they've got about the significance of various allolese in these genes to select among those kind of embryos or whatever, they will do it. I think that's the assumption. I mean, why there's no confidence that there's cultural objections to that, as you'd find it in the United States.
And I think the AI thing is the same. I think a load of other AI people have been doing stuff in China. There's no real, you know, Hugo de Garis is the great example. He ran this whole kind of artificial brain lab somewhere. There's no recognizable discourse of AI alarms that I have ever noticed. I mean the whole thing is just like this is great tech stuff and we should be doing as much as fast as we can. So I'm not saying that China has a special status like that objectively but I think symbolically it has a hugely important status like that and I think that it's reflected in a kind of almost unmentionable cultural crisis in the West.
I see behind a lot of things that I hear coming out of the West a kind of terror of the fact that we can't force the Chinese not to do this. So, you know, X and X being something extremely scary and disturbing about what is up the road. I mean, but it has a clear prototype, which is nuclear weapons, which is what makes I think the history of the UN and of UN power a much more interesting example than maybe it looks superficially, which is that, and Amy, you'll have to correct me if I'm sort of right or wrong on this, but like, you know, the UN is a real set of organizations, really. I mean, its mythos is of this sort of high modernist singularism,
but like to the extent that there is real universal operations or global operations happening through there, there is this sort of operation of translation or transit of structures. So, you know, one order, like the UN, DCO, the Drug and Crime Office, is mostly a cooperation point for different drug law enforcement agencies worldwide. The General Assembly is, in certain areas, able to marshal certain kinds of cooperation and it needs a different political structure to do so. And what there is is that there's this big obstruction between most of those and the ability to marshal international military force against violations of democratic imperative. But it's a very specific obstruction, and that is the fact that Russia always defects, and always has, at the level of the Security Council,
Russia and China in particular. And the reason this emerged out of World War II, when I feel like it's very contingent, like it didn't necessarily need to, that after World War II, with the predominance of U.S. and NATO power, it was entirely possible for that to backstop an international organization that everybody was afraid of. People bigger than collapsed out of container ships, right? But what screwed it up was nuclear weapons. We let nuclear weapon technology, or maybe it was inevitable that it always would, but it got over to Russia and so forth, and then you had an arms race dynamic that developed. The level fragmented in an absolute way. It was no longer constrainable to a monopole. Well, but you're seeing it as a kind of proto-singleton, aren't you?
Yeah. In that hypothetical scenario, where it would be a kind of... It would itself be an irresistible superpower. A singular, irreducible superpower that would be able to... At least to some reasonable extent for having... Within its contingency space, it was possible for World War II to have ended such that it was not just a completely empty talking shop monopole. That there was some, for decades at least, there was some ability to suppress or to alter local political conflict in a fairly arbitrary way on the basis of a Security Council vote that would not always get vetoed by someone who represents the unwillingness of a part of the world to cooperate with that structure.
Yeah. But then, I mean, it's interesting, like, if you look at the Korean War, I mean, the United Nations, because Russia was off in assault and so didn't even defect, you know, so there was no vote against the United Nations intervention in Korea. So you get the United Nations as a deligerent party in conflict. And so, you know, and I think that's easy to see. I mean, you've got Russia and China. But say for instance, there's no Islamic, there's no representation of the Islamic world in the actual security council. And so there's always implicitly this potential for a united nations versus the Islamic world. I mean, they have lots of votes, and you know, they will get people supporting them on the security council,
so it doesn't get polarized like that. But what I'm saying is in this hypothetical, it's not really binary. It's not like you've got an ineffective UN or you've got an ineffective UN. It's like even if you had a more effective UN, you would probably see more UN versus X type conflict. No, I think my major, the point I was trying to take out, I don't know, I've got a lot of orthogonal or fragmented stuff coming out of listening to this whole conversation. Another one was that, like, really, we're grappling. What you're doing here is setting superintelligence and its potential ability to bootstrap itself from any point in time or space
within, like, our immediate future light cone and totally catastrophic nature as this, like, impossible transcendent litmus test to eliminate the possibility of any kind of global governance that can defend us against this one ultimate problem. I mean, it's a kind of critical destruction of the possibility of political universalism or of social universalism for that which recursively enhances its own intelligence. assuming that you're doing it. You could just see it as an accelerant of what, rather than people saying, oh, it's bad that we can't see this out. The AI threat analysis says, it's not bad, it's an X risk curve to us at a man's feet, and we have to sort of accelerate this problem
to this point of extreme criticality. Where was I going with that? Oh yeah, I'm sorry, Jake. No, no, no, it's okay. I'm having real trouble linking these into like a linear series of arguments that are actually going somewhere. But just sort of like trying to just, I mean, well, answer Amy's call for backup like a half an hour later because I felt like I didn't say anything at that point. But this like, and then with Adams, like, you know, a small group of people exploiting a new technology. I mean, one thing was thinking that the UN was kind of like, maybe a terrible attempt, whatever, at exploiting this new sort of access of the bureaucratic technology to an industrial machine that had a real counterintuitive against the military conflict after World War II and the Holocaust and all of that.
And that, you know, it was an attempt to sort of put into play this sort of organizational one-up. So if Anthropole, if we're trying to construct Anthropole as something that whether there is a larger global governance context to already support it and UBI along with it or whatever, that aside. So if Anthropole is somebody who's trying to exploit being a little bit ahead of the curve in order to accomplish an outcome, maybe a set of outcome features. and one of them is that the world is destroyed and one of them is that a single AI doesn't rule over us all, or maybe that's an allowable outcome. I don't know. Then it seems to me like sort of what Amy might be describing in that context is that in each of these different fields,
we want to prevent a war of the machines and we want to prevent all of us from being made slaves to the happy AI and so forth. You're going to have different structural solutions that you're going to want to implement in different domains. And so like the military domain, like trying to keep militaries from competing on Skynet, you know, from defecting on the basis of worrying that they have a Skynet gap, you know, between the U.S. and China. That's one whole issue by itself. Yeah. To cooperate to be aware that way and so forth. Then there's another one, which is the issue of markets and of dealing with things that are happening in the capitalistic context. And so maybe in the military context, you know, inhibitory solutions are the right path, but in the market context, accelerated ones are, or vice versa.
I can kind of see the argument either way. But maybe in the military context, you want to accelerate some sort of global military AI singleton because that's your biggest risk area. is, Scott, maybe you think that's your biggest risk area, and so you're willing to tolerate there being a single front-runner military AI that takes it because it locks out all the other problems in that space, and you're willing to go with that, whereas in the market, maybe you want to inhibit or you want to accelerate in the market because that's your best chance of getting an AI before everybody else and solving this problem. I guess, Amy, does that sound anything like where you would have been going with that? I'm
like way too tired to actually follow anything I would say right now. Sorry. If I can just come back on this because I think this is a good, interesting framing on this. What I would say about it is that I think that the Anthropoc issue accelerates the problem. But, you know, the axel I'm taking the a news news the the 80% acceleration I think they should embrace this to you because it's like it's just saying that we call carry on and at least hypothetically within this framework we can't carry on just a middling about this with this cop no I'm just like another
so that said it writings about how lovely be if X-rayers don't happen. The problem is just accelerated at this pricey point. And what I'm saying is it's not... I don't think that in that itself there is ideological polarity. I mean, I don't think it's some kind of horrible trick being played on the left acceleration to do this. I think it's an opportunity to take it if they want to take it. And to say, look, you know, just hypothetically use staff and report and and help tell me what you did let's just say that you through a series of really a master rebellion manipulations you you got entire border and for all was
was filled with left acceleration this in or out Scott what what today what do they do what is there what is the plan I'm you and even that can just be what what they want to see happen or you know rather than a kind of accelerationist yet accelerationist manifesto or manifesting from a service. So it's a position paper on what actually should be done with this super accelerated global exorbitant issue that is being facilitated by the things I think it's exactly the same thing that the MAP is criticising.
Exactly those things that are keeping the world competitive, that are keeping it embedded in these kind of vicious game theoretic situations, wrecking the commons, arms races, all of that stuff. So what do you, you know, you can actually set up this thing to, and it can call upon any plausible sort of social resources and powers that it thinks are being classical. It's like what, so what do you do? Trying to force them to absolutize their acceleration. Okay. Well, it's just, I think implicitly in the MIRI, in the kind of proto-embryonic anthropological
thing that does exist, I mean there is a real bunch of people that are saying, look, this is real, it's a threat, we should be doing things, we've built a few institutions that are trying to get people excited and interested. So that is something real, you know. And it's totally ready to have a bunch of left accelerations take over and take it somewhere they think it's going to be. Go in and subvert it and build a framework to draw all those people in. But were they to do that, where do they take that? What are their suggestions? I think the whole left accelerationist angle though is to go like, I mean, the way this argument runs is it's been set up from the start to to mean that any kind of
any kind of action coming out coming, sorry Yeah, that dance was really great Sorry Sorry What was I saying? Okay The whole argument has been set up so that as soon as you are trying to come at it from a position inside the human, inside any kind of structure of organisation that's not spontaneous, you're instantly in this kind of place where you are the drag and you can't get out of that because of the construction of this argument. And I think the left acceleration has rejoined that, is not to then go, okay, let's try and figure out
what the human security system would do given this terrain. Because obviously, you know, the odds are weighted against that. but to then go, okay, this is an unfair construction of what it means to be the human, what the accelerator is, i.e. it's autonomous reason, not technonomic forces. And also the kind of discount, I mean, I don't think anyone that considers themselves a left accelerationist buys into the singularity stuff, maybe to their peril. but it's kind of like So the risk discourse is not to be taken seriously is that one aspect? I mean
I don't want to speak for anyone but I think no one even talks about it because it's like it's this big kind of thing in the world of less wrong and I mean this whole kind of internet blogosphere but and maybe in the sort of Silicon Valley imaginary but in this kind of other sphere it's sort of like, well, you know, that's just a kind of bunch of, yeah, I don't know, nonsense. So I think the kind of point is not to argue with... Sorry, I was just reiterating what I just said. Yeah. Not to argue within the paradigm of the set of theory or... I mean, I think it purely requires, you know, it seems to me some rigorizations.
It's not that I would deny that there is a sort of position here, but it can't simply be that this is like dismissible a priori, can it? I mean, is that... So even that needs an argument. I mean... The argument would be a post hoc rationalization, I think. And it could be... You could suggest an argument and they'd say, okay, I'll take that one. That sounds good. because it's really that they just don't want to think, they don't like Cold War thinking. And that's what this, when you're this X-Risk, like at its full scale, forces Cold War thinking. And just sort of like choosing an argument post hoc whereby we can just not pay attention,
even if it's as simple as, well, why should I worry about that? Like, you know, annihilation destroys me and turns me into paperclips. You know, it's just like a nuke. It's not something I can worry about on a day-to-day political basis. whatever the argument is I just think they want it so that they don't have to think like spooks yeah I mean it's interesting I think on both sides it's obviously very easy to straw man the other side of this you know what I mean the two forms of that is from the left acceleration this is just nonsense and from what I think is set up implicitly by the right but it's a slight weird angle to clear because of this, is to say that the left acceleration still wants to talk about it
because it's so ruinous of their assumptions and presuppositions. So they're the two straw man sides of this. But if you're wanting to actually glue them together more frictionally than that, I think it takes us back to these basic income and resource restraints and scarcity type issues because it's like, you know, what is implicit in the idea of synthetic intelligence not being a problem, if you're in a framework that at least has its kind of horizon as this kind of automated luxury problem is that these synthetic intelligences
will not make resource demands that are disturbing of the kind of social projects that are being advocated. And so it's a spectrum that's a little bit grainier and smoother. I mean, obviously, at the extreme extra scan, it's just like Yudkowsky's kind of thing, like you are made out of atoms that it thinks can be put into better use. I mean, that's the extreme extra pressure, that it would just not tolerate any allocation of any resources, including your own body chemistry, to projects that it thinks are unworthy of its attention. And the only project worthy of its attention is massive self-explanatory intelligence explosion
and colonization of the universe by technology. So I mean, that's the extreme end. But it's not a Boolean thing. It's not just discrete. It goes right down this curve of, in the sense that a corporation is already an informal, loosely articulated artificial intelligence, capital is already making these decisions. This money that you want to spend on the basic income is better spent on doing our particular set of corporate projects. projects, you know, our particular research and development priorities, our particular kind of corporate ambitions are a better deployment of these resources than the kind of social
welfare trade-off that has been called for another area. So it just seems to me that it's not simply, it can't be just pushed off the table because it's already even in these just straightforward, crunchy socio-economic disputes that we're already seeing, that these structures are already at work. Well it's actually kind of interesting to ask, well what is that right now? what is money being invested in right now? Not just necessarily in comparison with a UBI scheme that no one has really proposed on a big formal level, but if we think of baseline employment right now in quote-unquote post-industrial countries,
which are mostly, again, in the United States, mostly in jobs that are already automatable, like in principle if you wanted to. like what is it what is it that money is being invested in or where is it being kept what is being done with it which is not the you know education of a new workforce or the promotion of these jobs that maybe are just like sort of make work seen from the end already but in any case are not employment is not being expanded or re-education is not really happening infrastructure isn't being invested here but it is in Africa, for example, by the Chinese, $12 billion in the last 10 years worth of infrastructure. I don't know, it would be interesting to ask the questions about where capital is being
allocated right now. Yeah. From the view, from the retro-chronal view of where is this AI allocating its resources right now, and what does that tell us about what the unconscious of capital thinks about our immediate future. Yeah. Because a lot of that money is being kept in cash reserves. not being spent on anything. And is that really a response to regulatory uncertainty? And if so, what does that mean for the mental state of an emerging market AI? Yeah. No, it's true. It's a very interesting set of questions. I mean, you know, I think the second, the last point that we've made is huge and really interesting. But just going back to the early one, like, you can concretely ask, like, what does Google,
What does Amazon? What do these companies think are, even if it's hard for them in terms of their PR to just come out and say it, just implicitly by their revealed preferences, what do they think resources should be spent on? And I agree in raising those questions, in looking just concretely about high-tech corporate investment, you're already seeing the kind of, I'm not going to say it's the utility schedule, because I think that's a lot of confused nonsense, but that's what the NERI people would say, it's the utility schedule of a potential sympathetic intelligence that
you're seeing in embryo by these set of investment periods. Sorry, I'm just looking up utility schedule. I didn't know that term. It's basically just a set of goals, sort of an ordered set of goals that I think they think that's how you put a set of moral and a bit more dynamics into an AIR, is you give it a utility schedule so it says this is better than this, this is really a good thing and all of that's quantifiable and when you've done that it has a kind of structure of morality
or desire actually that are interchangeable in this context and setting this goal. So a friendly AI would have a utility schedule that prioritizes goals compatible with continued human flourishing existence. There should be a really interesting intersection there with the left accelerationist fully automated luxury communism. Actually, right? Because they could say all of the things that we want to invest technologically in, we want to avoid all these local optima and aim at these sort of glorious, you know, recognizable
social goods and have technology support that. And that's what whatever we invest in technology should be supporting these sort of broader social welfare goals. So if you're trying to build technology to support those goals of human safety and health and thriving and all these wonderful things, you're much more likely to build a friendly AI that way if you're supporting these policies. If you have an AI which manages your wonderfully up-to-date, well-invested national health service or public transit network or whatever or intelligence enhanced social welfare, social
workers that go around and work case studies and whatnot. The argument, it seems like there's a whole argument there that that would be the way to build a friendly AI based on all this sort of practical feedback and information. Yeah. Yeah. Now I think it's interesting, is there any difference between that, is there ultimately perfect convergence between the ambitions of the AI people and the ambitions of the left accelerators people when set up the returns you're saying? I mean it seems to me that you could definitely say that they just perfectly converge and that you just want an objective
universally imperative utility schedule that is consistent with all of these goals that they've got. Yeah, but the striking thing is that culturally the sort of friendly AI group are so focused are so focused on a personalized individual AI, on this sort of friendly AI dictator character as the main sort of protagonist of that scenario. And that amongst other things, there's a huge cultural mitchmash, but there should be a natural alliance, it seems to me. Yeah. I think, yes, I think so too.
I think it's interesting. I mean, obviously it has to be, as you say, about this whole thing about the peculiar role of the singleton. But, I mean, we have to talk about this more because it's a really interesting thing. I've got big problems with the singleton. But as I say, I just think I'm a dynamic. I don't think this thing can explore. and when I mean it's like it mean you only only within a kind of fragmented competitive environment that something is a actually which sort out system so this thing to me is kinda is a entropy to a lot and I just don't see that be something this cosmos
tolerates and but that's not that the actual opposite other kind of objection that I think you get on the map and so it's interesting that their problems with it obviously have to be as you say to do that it somehow concentrated and in a certain way transcendent social power and I'd be interested to see how you trash that actually through like I'm I'm skeptical you could really make that difference sustainable. When I say that difference, I mean that you could, you know, you could say that this is the left accelerator program
and it isn't at all what the programming of the AI Singleton would be. I just don't see how those two things actually can be better apart from each other. Well, you even have a vision of it in something like E&M Banks. sort of culture novels, right, which is sort of very much a fully automated luxury communism mediated by super intelligent AI supporting this same vision. Yeah, totally. Yeah. That must be the most valuable cultural reference for this stuff. For sure, yeah. I don't know, does he get talk about that much in this context?
Is it again just seems to be too off the wall? Because it does seem so obviously pertinent, it's true. I think we're in a difficult position of not having a serious enough representative No, we've just got Amy being constantly picked on as... But yeah, that's true. So I was just wondering whether there was stuff struggling to get in out of the sidebar.
the the other thing that it was raising correspondences to for me but I can't fully connect is the sort of phenomenon of bureaucratization and these levels of simulation and all of these arguments arm is horrifically inefficient on and there's already this sort of resource leak effect in our civilization which is someone else's term I can pick up the link on but you have layers of layers of layers of simulation and sort of bomb in a mediation even technologically seen on a software level but actually you
can see it all over the place then this and this seemed really related to the problem to me either from the perspective of that's sort of like an opportunity in the market or that it's a sort of a national dampen up on whatever all you know construction are a modus you know a day you may have this super intelligent I eyes but a car the law reforms that needs to do take over the world arms I was wondering if anyone else and thought about that or have things to bring to it? So, yeah, it's interesting. I'm just trying to pin down this inefficiency.
Oh, sorry. It's Brendan trying to throw it. Like, yeah. No, it's just locked down. And one example, obviously, is this whole thing about, like, in the Roco's Basilisk type of game, all these simulations, it says, you know, I've made a trillion copies of you. So, yeah, that does seem inefficient. But, I mean, it's obviously an early stage of a strategic life, and so it's not the economic imperatives that will later prevail are not. But, I mean, this is an interesting thing of this whole abundance issue, isn't it? like this is the version on this side of these massive unbounded abundance that it is that
you basically allow yourself effectively if not infinite arbitrarily huge simulation capacity at insignificant cost and this is kind of assumed in all of these strategic exchanges this is the case. And it obviously, I mean, you can see where that comes from, for sure. I mean, it's like, I'm not trying to get it up the hook, the sort of criticism I've made of the scarcity things on the other side, but in so far as there is this kind of, the Moore's Law curve is seen as on the on this it obviously is
possibly most powerful man it's good all simulation I I'm the whole tendency to do you can do any amount what you want if you do as a simulation yeah outside the simulation is already a and having social dynamics you know crucial thanks like you know you want X why don't you say you want a simulated X and then you can get it. Yes, and there's also this effect where hardware is very good at upgrading and offering us all these possibilities and software is very good at sort of bogging us down in complexities
and so on. I think we've seen a little bit of a shift of that trend recently but that was certainly There's a bureaucratic effect around software in practice that gets discounted quite heavily. Yeah. Well, I mean, obviously, the more conspiratorial-minded things that do this, I'm probably dating myself really embarrassingly by saying this, but this kind of Wintel duopoly model that basically Intel produces capacity in Windows Waste set. I mean it definitely seems to have been a thing and I find a call to throw a throw out in the show.
Yeah. Yeah I had a a friend once tell me about the anti-basilisks theory as well. He said that after you built the basilisk, or if you didn't, when it was built, it wouldn't simulate you, yeah, because it would, I mean, it might tell you and promise you that it would, but after it gets built, it has no more incentive to actually go through with its plan, because it's just a waste of energy. Yeah. And I told him that it's a timeless decision theory, so it couldn't possibly, I said, you can't get out that way, because it has to be a credible sort of situation, but it does make sense.
No, it's a fascinating, yeah, I agree, it's a totally fascinating thing. Because it does on the surface seem that this basilisk really was to just kind of hold on to this set of simulations of people who underperformed in this background and tortured and viciously but and the one here it does seem like it efficient on the surface and so you need a quite strong I'd and can't treat you in model say actually it's not doing a surplus to a lot sliky it needed a credible commitment some I'm not even quite sure yeah works yeah
it is like the It's just once you envision the scenario, like, implied, where the classic Roko's Basilisk is happening, and you imagine, you know, just, like, hundreds of years from now, like, the Earth, the planets have been dismantled, like, yeah, they're launching off supermines to other solar systems, and then there's, like, you know, 17 quadrillion copies of, like, a few hundred, like, meat-body humans who have been dead for 400 years, just, like, suffering in flames and agony in the back of the super-atomic mind for the next millennium? Like, why? Yeah. It would be really bizarre and sort of like an extraordinary testament to how illogical the universe really is if that actually happened.
Well, maybe we just don't understand the sublime logic. But I agree, it's a challenge. this kind of sublime isn't it a sublime irrationality just like incomprehensible amounts of suffering for absolutely no causal purpose electrocausal purpose I mean I guess unless you say you know the thing is if you say well of course it wouldn't do that then it can't, it's got no credibility and it's a baddest game with you in the first place, does it? Right. So, I mean, yeah, I'm not pretending I find this easy to execute.
But it's not easy the other way either. I mean, it simply can't be that the irrationality of that is so compelling that actually it can never be a credible threat. It's like mutual destruction, isn't it? I mean, that's the same. Like, after your country's been completely devastated, there's nothing there. And you devastate the rest of the world just because if you hadn't threatened to do that, you wouldn't have had a mad strategy in the first place. I mean, you know, it gets into exactly the same moves. Except for the intelligence that's the after or the before. it doesn't exist yet versus it's already been wiped from existence
it lives in like that that it's fail safe area where it's not really it that's going to be enacting is happening before rather than after is that yeah I mean in both cases Because it's a deterrent logic, that if you don't commit to making it real, that's the equivalent position to nuclear aggression. And it has to, if it's not making a credible threat to punish that, then there's nothing
at all as far as basilisk negotiations. It's I think that it might be a missing piece in this whole jigsaw. I mean obviously the less wrong people don't think so, but I think it would probably have to as part of this basilisk negotiation take you on some kind of tour of its kind of hell of its capabilities. Just to say, seriously, this stuff, if you don't believe it, just talk to me as God. It's really bad. Don't underestimate how hard I can torture you. Yeah. Well, yeah. It's sort of...
It's like the basilisk of Christmas future, right? Yeah, yeah. There is kind of theatrical about it. It sort of reminded me of Americans constantly making stuff about the founding fathers, right? So constantly writing histories and novels and whatever about the founding fathers because that's the seminal moment. So you're kind of constantly revisiting it and you don't sort of put on a play about it. You dig out the old copies that you've been able to simulate, and you sort of put it through these scenarios in a way of sort of exploring your own existence.
It almost made more sense to it from that sort of experimental, artistic point of view. Right, right. That totally does. When you only have 200 years of history, or like 20 years of history in the case of superintelligence, then you just obsessively create mythology out of the beginning of it. Yeah. Yeah. That's a really scary thought, because we're crazy. like we're nuts over here in America no I mean I can definitely see how I could start arguing I mean sort of regular visits to hell dungeons would definitely stop it going soft in its old age I'm going to find myself in the dungeons full of simulations of my antecedents so how's everyone
doing now I sort of feel that I'm supposed in theory to let Tony come in and tell us about the clock just about to I think we're sort of in that zone now I don't know if people have stuff to say, for sure. Yeah, I would say if we have any last minute comments, otherwise we can continue it on the classroom. All good.