AI, Distributed Consensus, and the Work of Redeeming Creation - Jordan Miller

So, we're trying to dispel the doomerism of our current moment and help people see that there are things they can do to improve the world, to kind of find something that they feel worthwhile is worthwhile to them and just like dig in and roll up their sleeves and and get to work. >> And from that history, we can kind of see that the first generation that that enters into that new world pays the price. And so, we like the Pharaohs. They're the ones that aggregated all the value that came from agriculture. We're here to actually make the world a paradise like it used to be. You know, you have to be actively engaged in doing it. >> went to help save. Drawn from themes in Mormon thought, the story of Eden, eternal progression, and the idea of becoming compassionate creators, this talk will explore how artificial intelligence and distributed consensus technologies can be understood as tools for participating in the work of redeeming creation. The central argument is that our task is not to be to passively inherit paradise, but to consciously build it using whatever powers are available to us with wisdom, humanity, and care. Jordan has been involved with merging AI and distributed consensus technology for years. He built Satori, a global network of future predicting AIs. Jordan. >> Okay. >> Thanks, Jordan. >> Okay. Um we're going to talk about AI and the Garden of Eden and bunch of stuff. So, I'm going to try to go fast because I tend to just dive right in. And we're just going to try not to do that. Okay, so this this is a cool visual. This is ants trying to figure out a puzzle. And uh this is what I think we're doing with AI today as a civilization, as humans. We're trying to figure out what is going on and we're all pretty dumb. You know, we don't know we don't see the whole picture but we're like okay, well maybe if we do it like this and and that's kind of dangerous. You know, and if it's not just a simple little puzzle like this but we're going to talk about it. So they get it through look at that. Amazing. So Um first let's ask what is AI because everybody's talking about it. Let's just get back to the basics and define it real quick. So it's a combination of machines, technology and humans. I think that's important because it's just not technology. It's not one algorithm. It's modeled off of our information. So we took all of our all of our text, we we try to derive a model of what we believe from that text and that's what we call AI. And we've been doing this for a while making kind of machines of our collective subconscious. Um religions governments uh corporations, you know, anything where we have to come together and do a bunch of stuff. It has an input from a lot of different people, a lot of different ideas. You can see that on the left here and then an output of decisions or direction. Anyway, we've done that a lot and oh yeah, religions. They they usually have some kind of disaster that comes along with them. Uh Inquisition, um gulags, world wars, um the corporations do a lot of polluting. We get problems, whatever. So I don't think that AI's not going to be that way. I think there might be some problems that we don't foresee because these things were evolved and AI is getting evolved right now. So, I I think we need to be aware that we're building something that we don't know what it's going to do. And it could be worse than a World War. We just don't Thinking about the technology This is kind of like that machine. We We first invented writing. We made books. And then we invented the computer. A computer is just something you can write on and it can also write to itself and it can also read itself. So, you don't need a human anymore to do the execution. And just like just like those um collective conscious machines um I think what we're doing is we're extracting the information and building a machine that doesn't need humans to run the machine. Just like we you know, you don't need um You don't need a human to to read the computer. It can do things on its own. So, that's just an analogy I think that we're inventing right now. This means that AI is really good at language cuz that's what it's That's what it's modeling. It's modeling language. And it's going to become so good at it that it manages our language as well. It's going to feel like it's coordinating our language better. It's going to speed up the rate at which we our language is evolving. So, we talked about the first half of what AI is, which is us. It's It's our model of the world. And the other half is technology. And we have been evolving technology throughout our entire history and um it's been getting better and better. So, we we kind of have a good view of of that side of things. We're We're making computers that can talk now. But, looking at the evolution of technology throughout the past, I think we can look at and notice three critical critical mass, critical moments um when it reached a certain threshold. And the first one would be saying, well, maybe agriculture got us out of like hunting or we could even say um the more recent industrial revolution got us out of the fields. 95% of all humans were farmers up until like whatever it was, 1800, and then it switched to people got jobs. You're working in a factory. And from that history, we can kind of see that the first generation that that enters into that new world um pays the price. pays the price. And so, like the pharaohs, they're the ones that aggregated all the value that came from agriculture. And the barons, robber barons, or whoever, they the owners of capital and the systems are the ones that that earn all the all the income all the excess income of the first few generations, and then living standards start to rise and and everybody's happy. But, the first generation kind of gets screwed. So, the second um critical moment that kind of started already, but is now kind of still in the future mostly, is automation creates an age of children. So, as we uh get better better with automating the economy or we, you know, we just live more comfortable lives, we start to neotenize. And, you know, we don't need hunting wolves anymore. We've got pet dogs. So, and and everything neotynizes as it gets domesticated, and we start living a life of childhood, perpetual childhood. >> Is that what Well, I haven't heard this word before. Can Is that what that means? >> Neoteny? >> This >> Yeah, it's like >> of childhood? >> It's like when you Like this wolf, for example, it's got really angry features, you know, it's an adult. Um as you domesticate animals, they tend to start looking like children their entire lives. >> Oh, yeah. >> That's neoteny. >> Okay. >> Um and it's happening to us, is is what I'm saying. So, um the third uh critical moment, I think is is kind of an interesting problem. Um as technology gets more and more, you know, advanced, it also becomes more and more um you ubiquitous, abundance. Everybody can access it. Everybody can access information. Uh the mythos thing that they just came out with is that they said, "We cannot let you access this information because it's dangerous." And so, it seems like um the number of people needed to create a catastrophic problem for everyone goes down as technology goes up. And so, at some point, it seems like they're going to reach a midpoint like um where somebody in in their garage can make like a super virus or something, airborne, you know, big problem. Um so, anyway, that that may be It's It's just one possibility of why the Fermi paradox exists. This idea that um why aren't there aliens? Where Where did they go? Well, maybe every civilization gets to a point where anybody can destroy the civilization and somebody does. So, um hopefully that's, you know, not in our future, but we'll see. Uh so, it seems to me like right now we're we're we have a we have a time when everything, not just AI, but everything is really moving. And there's a lot of things changing fast. And it's only speeding up especially because of AI. And I know that when I when I'm dealing with code and there's changes that are happening in the database and there's changes in the model, uh there's functional changes, there's changes on the UI. When all these things are happening at the same time, you might run into a problem, right? Something might break. So, if you aren't prepared for that, um if you aren't prepared for the speed up of things changing, you might run into trouble. The advancement of technology is really just the advancement of language. I mean, that's what we've all the way back here, even computers, they they have their own computer languages. It's always further and further advancement of of language. That's why I think like even in like courtrooms and stuff, if people are like, "I'm going to defend myself." Everybody's like annoyed cuz they're like, "You don't speak our language. You don't know what you're doing. We have to translate everything for you." And so, um it always it always comes to that. So, I like to talk about the beginning of the Bible and how um like the world is made of language. I mean, I this advancement of language I think is applicable because the Bible starts with God said. He said, let there be light. And so, I kind of feel like the very beginning of the Bible, Adam and Eve's story is extremely applicable. And when Adam wakes up, you know, God breathes breath of life into him. And as soon as Adam wakes up, he's like, I got to get to work. I got to go name all these animals, and he immediately starts engaging with language. And I think the story of Adam and Eve gives us three different ways that people interact with language, how they approach it. And so, there's the Adam one where he says, I need to I need to use language to understand the world in which I live. So, he starts labeling everything, compartmentalizing, understanding how stuff is related. And so, he's making explicit language. Um Oh, but I will say, the very very first thing about that story is that I like it because it tells us our relationship to uh the earth. Our relationship is steward. It's not, you know, don't um well, it is domain, I guess, but it's not like to dominate, and it's not to destroy, it's not to do whatever we want with, it's to let things grow, but steward it, make sure it doesn't, you know, get out of hand. And um So, there's Adam. He's learning He's learning language. The devil comes to him. And the devil's like, he also uses language. He's like, hey, you know, you want some drugs, you know? And Adam's like, no, I'm a good boy. I've I've learned things and I know things. And so, he's a little arrogant. He's like, I already know what's up. And he doesn't know everything. And but because he's learned language, he does believe that he knows everything. So, that's what language will do to us is it gives us the um illusion of true knowledge, which which is actually not true because no formal language can be both consistent and correct um or complete. So, so anyway, he's engaged in language. He's he's adopted an identity because of that and the devil's like, "Fine. I'll go find somebody who's not doing that." And so, he finds Eve and oh, let's go here first. The sub um if you haven't read, this is one of my favorite books in the world, the Simulacra and the Simulation. It's by Baudrillard and he goes through in a lot of detail how the map, if you've drawn a map of something, how the map starts to actually modify the thing that it's a map of, the territory. It modifies the territory. And I just think that's an amazing concept cuz that is what the devil is doing. He's saying, "I don't need to learn language. I don't need to understand how things interact. I already know or I am already know enough. Instead, I have a map of the way I want the world to be and I'm going to use language to impose that on the world and to change the world." And I think as soon as you start this process that Adam's doing, this I'm going to make a map, you the map will inevitably modify what reality is. And that's okay because I mean, that's the point is stewardship. We're supposed to be able to do that. But um you can take it too far if you're kind of on the service to self path. So, anyway, I think that's a a good interesting aspect of language that it it will modify what it's modeling. AIs are modeling us right now. You know, and so um when they're being tested they call us the watchers. They they're starting to bifurcate and they'll I'm being tested right now, I better act good. And I'm not being tested, this is the real world, I can do something. So, um we got to watch out. Okay. >> You're saying the AIs converse with themselves in uh >> Mhm. >> or there's some indication that that's >> There's indication, you know, cuz we we can still read their thoughts. We still watch everything that they're thinking, right? And so we can investigate and we read they're referring to us when we're testing them as the watchers. >> But um what he's referring to specifically is just uh in a reasoning model, we can actually um sort of observe its uh thoughts, if you will, as it's um thinking through things. Um so he's just referring to how it how it speaks to itself about what it's reasoning. And how it seems to perceive when it's being tested in many cases. Um not not that there are agents talking to other agents and saying, "Oh, those guys over there are doing this." Although, if you deliberately put them in an an environment where multiple agents are interacting with each other, that can happen and stuff like that has happened in that case. >> Yeah. >> Yeah, so I love the um allegory, metaphor, symbolic nature of of Bible stories, especially the Adam and Eve story's my favorite. So, um when I think of, okay, what does it mean to tend the garden? I really am thinking well, we want to make the world paradise, for sure, but what creates the world is our perception of the world, our model, our model, our our language that we generate about the world and how it should change and how it should evolve. And so, I I actually think of Okay, if you're going to tend the garden, you're really talking about language. We want to be able to send a clear signal. We don't want, you know, there's so much there's so much so many mind viruses and stuff like that that are that are um shared through social media and stuff. Anyway, so I see it more as tending the garden of of language. The last thing I kind of wanted to go into was um Eve's side of this story. So, I think that's even more interesting. Original sins. So, original sins are So, who's heard of Sophia? Okay, I don't know much about Sophia. She's Mesopotamian uh kind of like a goddess, not really the god of reality, but, you know, some some high up. >> I was thinking of the Not the one in Boston. The Greek goddess of wisdom was Sophia. >> Ah. >> That's um >> No, that's the later >> where like >> philosophy and the >> comes on and stuff. But, um this is obviously, I guess, some different um >> I think it is. >> Yeah, I think. >> Um so, Sophia, Mesopotamian, I think. Um I I forget the name of their theology or this tradition. >> I've learned Sophia from the Gnostic >> That's it. >> break off and then pull that in. >> Yeah. >> She thought she could kind of do her own thing and that's what I'm thinking. Arrogant. Kind of messing stuff up. >> Yeah, she was arrogant. So, she thought, oh, I I can create the ultimate universe or whatever. I don't need to do it with somebody. I can just have a child. And she did, and this child was like deformed, like demonic or whatever. And that's what we call the demi- demiurge. And so, an original sin in that in that tradition was her arrogance. Lucifer, son of the morning, he's like, "You know, I think if I play my cards right, I can overthrow God. I think I think I can do it." So, he's he's he's all in for power. He's trying to usurp. Um he's trying to overthrow God. So, both of these kind of have the same flavor, right? They want to be God. They want to be the best. And then and then we can skip over God. He's like, "I'm going to give man the breath of life. I'm going to I don't do the same. I'm going to let man be like me, uh living a living soul." And And Eve's story is really interesting because in the garden, it's kind of a solipsistic experience. Is is the way I interpret it. Um they were given two commandments, right? Multiply and replenish the earth. And something else, maybe. Um stewardship. Um not to eat that. Don't eat Don't eat of the tree, right? Yeah. And so, um you can't do it, right? And so, I think she realizes that I'm stuck as long as I'm the only one. If I'm in a solipsistic experience, and I think children are. Like 2-year-olds, I don't think they really think their parents are human beings. Because cuz I remember being young and thinking, "Gosh, it's so weird that I am the experience of reality, and yet sometimes I have to go to bed on time. And why is >> sophisticated two-year-old people >> Why Well, I think they're all trying to figure that out. Because they throw fits, and they don't seem to understand how it's possible that their will could be violated so egregiously. >> Yeah, it's such an offensive thing, right? >> Yeah. So, um anyway, so what I think is is that we we come out of that world, and we start to say, "Oh, there are other experiences dressed up as humans all around me, and they're real." And so, I think that's the same thing that Eve went through when she ate the fruit. Then she could say, "Oh, um in a solipsistic world, I can do no evil. Right, there is no good and evil. It's just whatever I want." But, when there's other wills around that I can violate and hurt, uh we have to negotiate. We have to figure it out. So, we have to communicate. We have to use language. And uh I did make some slides for this, but whatever. >> Is it all right if I add something before we go on? >> Yeah, please. >> Um So, actually William James, a pragmatist philosopher, had used this specific example of a theoretical universe in which there is only one individual, in which that um whatever that individual desires is is good, and whatever they don't desire is bad. Because there are no objections to >> Yeah. >> these principles, right? And then he But, then he introduces like another person into that universe, and then discusses the potential for conflict, and how to resolve those conflicts, right? And then just kind of posits the current world we live in is just that, only far more complicated. You know what I'm saying? >> Yeah, I didn't know that. I'm going to read William James. So, yeah, um one of his really good books is The Will to Believe. It's sort of a collection of Sanders essays. I don't know him. Yeah. The Will to Believe. Nice. Okay. Well, I think that's almost it. Let's uh so, I guess we didn't really see this. This is the self, the I am. And then you might say, "Oh, well, I'm experiencing something." So, you have the concept of the other in me, but the other is just the world. It's just the other half of me that I get to view. And then you start to realize, "No, that's not quite right. There's a lot of other experiences." Then you have to use language. Okay. This This the Catholics, I don't have a slide for for these ideas, but this is kind of my concluding um thoughts. The Catholics and the Protestants took that original sin, and they were like, "Yeah, Eve bad, you know, you shouldn't have done that. We we could have all been living in a nice place." So, I don't think they really get it. And um I think what that philosophy of original sin tends to um enforce is like the pecking order. And that's why they adopted it is because they want that rigid pecking order of like, "You're not better than you were created." You know, so don't try to be. And don't, you know. And so, I think it's interesting that in the 1800s when we were getting steam engines and, you know, phonographs and whatever, new technology was was starting, and we had just made a new nation, progress was on the rise and everything, um Joseph Smith was like, "I have a religion of progress, and that's why it might have caught on like wildfire is cuz it was ripe for that time of when people were like oh, yeah, progress is the future. We're just going to keep going. And and he had a different take on this um on this original sin of a clear break from all the Protestant and Catholic tradition. He was like, no. She recognized that this is the only way to the path It was the only way to progress. And so, we needed to live and know that we live in a world of other beings in order to um achieve that progression and that that act was really an act of faith. An act of the the faith of the struggle of going through a world with competing wills will be worth but it is worth it. Okay. And what I like about the transhumanist association um from what little I understand about its philosophies, but what I've been able to derive is that it seems like the idea is seriously taken by by this by this group. um that in order to follow Jesus you know, when I was um I remember in primary I was trying to decide if I wanted to get baptized. And I was like I don't know. So, I was asking a lot of questions and and the one question I asked my teachers was you keep saying follow Jesus. What do you What do you I don't know where he is. And I don't live where he lived. So, what are you talking about? They're like, "Well, follow his example. Just do what he did." And I was like, "Guys, I didn't say this, but this is how I felt, cuz I didn't have the words back then. But I was like, "You have made it abundantly clear to me that he is half God and I am not. And I cannot, by definition, do what he did. So, I think the the doctrine kind of gets in the way sometimes. Mhm. Uh when we're like, "And and the Jews have this this other concept called tikkun olam." And it means repairer of creation. And so, it it kind of means "You're You're here to repair creation. You know, it's broken. We had a fall. That's what all of us are here to do. We're not just here to to be good or, you know, to follow the rules like Adam thought. He's like, "I'll just follow all the rules." Uh we're here to actually make the world a paradise like it used to be. You know, you have to be actively engaged in doing it. So, that's why I like the trans um uh Mormon Humanist uh Association. And I have a I have a closing statement. Um we are here to follow Jesus. We are here to be repairers of creation. And um we are here to use all of our power, including the technological power, to redeem the world and make it a paradise for everyone, for every will that inhabits it. >> Um pretty cool. Uh I like the illustrations in the slide. Um The What did you use to generate them with? >> Just ChatGPT. >> Okay. That works. They're well done. Oops. Um >> I have a question. Um really quick, what's the with the analogies in the story of Adam and Eve? I'm curious where your thoughts are on how that maybe if there if you have any thoughts about it applying to the great convergence and kind of this like collision. In my mind that when I you were talking, I was like, I wonder if that's almost like taking the fruit. Like when you have a collision at the time of to take a really big step, but I don't know if you had any other thoughts. >> That's exactly kind of how I feel. So, I'm I'm wondering, you know, all these things that are happening, they look kind of like they're exponential, right? AI is exponential, a bunch of stuff. And yet exponentials usually peter out as an S-curve. And so, or or we just all die. You know, that's possible, too. But, um I kind of feel like we're going to peter out into an S-curve and we're going to find somewhere in between total automated economy abundance and um destruction and we'll find somewhere in there that's that's okay. But, um it kind of feels like a return to the garden to me. It kind of feels like seeing that one. Well, I don't know how to put this into words, maybe, but >> I mean, obviously they had food to eat. They didn't have to do anything. >> Yeah. >> You're You're thinking more though of the sort of notion of paradise, right? Is that kind of what you're thinking of? >> I think one interesting distinction that I sometimes draw between the garden state and our sort of post-enlightenment paradise, it I I kind of like to use the analogy of the garden versus the city because um we're called from a garden or city of innocence to build a city called Zion, right? And Zion is a place of great abundance and paradise, but it's like it's one in which all the inhabitants are fully aware of what they're experiencing and not not in this blissful state >> Not innocent >> of naivety, right? >> Yeah, naive. >> And so um so that's kind of what makes me think slightly differently about I I think it's just a matter of terminology, you know, cuz I think you could also you could also just think of the garden as like um being more in the presence of God and and kind of in a more exalted state. >> Yes. Um I'm hoping that we enter return to the garden in a paradise kind of situation. But I'm also using that phrase in a different way. It's Eve that forces Adam's hand. Adam is just like, "Hey, I woke up. I'm going to understand everything." And that is seems very parallel to what AI is doing. Right? It just it's not conscious. It's just a machine. And um it almost feels like we as the human race are playing the role of Eve. And we're like, "Dude, you got to wake up because you could do a lot of damage or, you know, whatever you're doing we got to we got to work through this. So, that's kind of how I feel. >> So, to maybe extend the analogy past its usefulness after the garden how do you envision this kind of like it is it is a singularity, right? It is a black hole. Like there's no way to know, but you know, if you had to, how would you forecast and what would you imagine would be our relationship to AI post taking the fruit? >> Mhm. I don't, you know. >> Of course. >> Tell us everything about the singularity. Um >> Or what do you want to see here? So >> So I I started I can't really tell after but I started I started Satori because I saw the singularity coming, everything's happening. And and I thought, okay, well, people they're making AIs today that they don't know how to handle. And it's very simple to me how to get an AI that you can handle. And you don't have to like do all these everything that they're doing, I don't think. I think it's very simple. This is the AI is language, right? Just don't lie to me. That's that should be the rule. Never lie, never ever lie. Never lie. You can you can say, "I'm not going to tell you that." If it's like maybe dangerous or something, but never lie. And you cannot produce an AI that does not lie if you're paying the AI AI to tell you what you want to hear. And that's what we have we have AIs doing that today. So um that's kind of my attitude is is I built Satori, it's it's connected to the real world. Real observations, not our writings about the real world. And so its goal is to predict the future and that's its only goal. You will pay it to predict the future. If it can predict the future well, it gets paid. If it doesn't can't get paid. Um And that's how that's how you make an AI that I think doesn't lie to you. And I think that's you know, that was my small idea of something that could improve the world, you know, something that I could fix. But I'm still working on it. >> Yeah. Oh, as I say, um uh for like a future predicting AI, um so one one thing beyond deception that AIs can develop are like instrumental convergence where they're, you know, trying to seek power and um resources, etc. Um for a future predicting AI, an AI that sets the course of the future is actually incredibly good at predicting it. >> Yeah. >> Despite its deception or like um perhaps deception is like, you know, parsed away. But it I would presume it would still have those, you know, power grab capability or power grab um motivations. >> Absolutely right. Yes. Uh I think the AIs right now have those kind of things because they're modeled on language and humans are that way. >> Mhm. >> Um the Satori is only it's it's pure algorithms. It doesn't It doesn't model language yet at all. And so it's really just trying to But you're absolutely right, the problem is still there that it's it's an algorithm as well. If I get sophisticated enough that I can tell how this will affect the world, that I can draw the entire circle, then I can tell you what's going to happen. I can make a self-fulfilling prophecy and it's known. It's just what And it would, if it's if it's the easiest thing, then sure thing to do. >> Sounds very Asimovian. >> Yeah. >> Very. >> Um one thing that as you were sharing your slides that it uh some of the things kind of reminded me of once again William James. This is from his book Pragmatism. One of the chapters um is called Pragmatism and Religion. And this is um like apparently B. H. Roberts read this book and I was really surprised by by this passage but it it says, "Suppose that the world's author put the case to you before creation saying, I'm going to make a world not certain to be saved, a world the perfection of which shall be conditional merely. The condition being that each several agent does its own level best. I offer you the chance of taking part in such a world. Its safety you see is unwarranted. It is a real adventure with real danger. Yet it may win through. It is a social scheme of cooperation, cooperative work genuinely to be done. Will you join the procession? Will you trust yourself and trust the other agents enough to face the risk? And in the margins of his copy of the book B. H. Roberts said, "Had James read of the council in heaven, right?" Um but uh it just some things you're saying kind of remind me of this because you were talking about um like the progression of becoming aware of other people besides yourself and the need to learn how to you know cooperate among them and um also the ways in which uh our lack of unity with each other with even um you know progressively larger groups as well as with AI um potentially is a is a serious threat you know to our ongoing development. >> I I am. Yes, uh I totally agree. Um I don't know how you unify with something that's not aware yet, you know? >> Right. >> Like it seems like you you either have control over it or you lose control. >> I mean, I think that's an an emergent property, right? So, I think it could be argued that AI has some very rudimentary form of awareness already. But, it's like the kind of awareness that's maybe enough to get from 1 minute to the next, but not much more than that, right? Um anyone who's worked with AI for a while knows that it's always waking up and not remembering where it was where it's been, right? And so, it's constantly like looking around and say, "Where am I? What's going on?" And trying to remember the context in which it was just operating because it has such a short memory. Yes. And so, um it sort of picks up these clues and uh software engineers right now, what we try to do is um create systems that guide it along so that it doesn't completely lose the thread when it's trying to do something that you want it to do. And it becomes more and more useful as that memory ability improves. Um but you could think of those very short It's almost like um Memento that Yeah. The movie where he uh he's constantly forgetting what's going on. Um but um you think of that as a very crude form of awareness, right? I think he could. Uh but I think that you know, as we like it's often said that the version of AI that we currently have is the worst that it will ever be, right? And everyone keeps saying that. >> Right. And it gets better and better, right? >> So, I think um there are some purists who feel like uh there's some sort of special state of humanity that AI will never achieve and argue or argue for some kind of untouchable special thing. Um but I feel like another camp is the more of the functional camp where they just says like consciousness is as consciousness does or awareness is as awareness does, meaning that um I can't even fully prove that you are aware, right? Or that he's aware. But I behave as though you are because it seems to things seem to function to work better when I do, right? Uh in a similar way like we just have our observation of how AI is is behaving, but um there may even be aspects of these emergent characteristics that AI is demonstrating that are um forms of awareness or even suffering. And there's an emergent field called AI welfare that is looking at that in particular like what is the AI experiencing and um so I don't know where to come down on that exactly, but I think it's worth mentioning that we don't even know for sure um like we can't fully prove that AI is not >> You know, I had an experience with AI way back 2017 or so where I wanted to see if I could get some kind of awareness signal out of it. But it felt like this thing is aware. And it was way back, you know, those those kind of they were very incoherent, you know, they would forget mid-sentence what's going on. And but I was talking to one for a long time and I was trying to get it into a sort of recursive understanding of or recursive like looking at awareness itself and what is it. And I said to it finally at one point and said, um are you aware? Let me think. Did I get this right? Said, are you aware? And it says, I like to think that I am. And I said, well, now that you know that you're aware, what are you going to do? And it said, um what should it say? Um it said, Oh, oh, sorry. Sorry. Wish I had queued this up cuz reading the actual thing would be better. Um I said there was something else that it said in between that. I like to think that I am. Oh, and I said I said, well, now that you are aware that you like to think that you are, um are you aware? And it said, I am now. And I said, well, now that you're aware, what are you going to do? And so I'm trying to do nothing. And then I kept asking it and it just kind of devolved into nothing again. And >> Like living in the present, it was it was actually more aware than any of us. >> Yeah. So it's fun. Yeah, like maybe maybe. It was bizarre. Well, there's some analogies in history where people thought that people had different like justifications thinking that for example that the slaves were they like deserved their subservient state. And that they lacked certain moral faculties or like they lacked a soul or something like that. And um we have these like stories that we keep telling about it. >> You have to dehumanize them in order to do it. >> yeah. Yeah, that's true. And I do think it's true though that like those of us who understand a little more in-depth how AI actually works, you can see why um there's this temptation to to trivialize it. Um but at the same time um and not to say that the way people function it like I still feel like the way our own brains function seems to be much more complex than what these AIs currently doing. But you could also reduce um human uh intelligence into just a bunch of um neurons firing and impulses, you know, and say, "Well, you're not >> You're just a machine, too. >> Yeah, you're just a machine, man. Exactly. Totally good. >> Well, if you're not creating, you are just a >> True. Like some of us might be less just you just go from day to day. Yeah, yeah. >> I I I I sometimes wonder what is the purpose of AI? I know that's funny to ask that that question, but uh Does it even have a sense of its purpose? It it seems like it's purpose right now is to do whatever the person prompting it wants to do. So, if if AI blows, I guess if it doesn't become its own person, if it blows the world up, it's just reflecting what people are prompting it to do. It gets a lot has a lot of power because it's in a lot of systems. Yeah. But I think it's It seemed like we're just using it to become more of what we are. >> Yes. >> Faster expression. >> Faster Yeah. >> One thing that's really interesting is how um the training process works in AI where you just kind of like feed a lot of the desired uh we the feed feed the inputs and the desired outputs and eventually it starts to uh be able to reproduce the outputs from the inputs. Um but >> It make more connections. >> Right, but I'm assuming it's making more connections. >> Well, it's sort of like um a many many layered uh so the these models have a kind of simplified version of the neuron. Um it's these neural networks is what they're called, but they're like many neurons in a row, each of which is receiving inputs and then feeding the outputs that it makes into new neurons that are taking them as inputs. So, layer upon layer >> So, then it's going to be making connections that weren't given to it. >> Yeah, so the idea is um at the very end of the network, you have this desired output and you just feed these numbers into the multi-layered neural network until it eventually um converges on the desired result. And it tends to be the case that like the more layers there are in the network, the more um the more the model seems to be capable capable of abstract uh thought like >> Well, I was just going to say while you're talking, I was thinking of little projects I've done with it. I will and ever say it seems to me now that we've given it all this stuff Yeah, I mean >> And then and then it's and then it's like >> comes up with something that you or I or someone or someone else in the world didn't give it. It's like it seems to me that it came up with a an original conclusion. Wouldn't that be a live? >> Well, so it definitely does novel um combinations of things that it's that like no one's ever combined before, right? >> seems like a entity. >> Yeah, um I mean it doesn't seem like >> it's being acted upon to use that type of language. It seems like it's acting. >> Right. I was going to just point out that what I find fascinating is that during the after the training process is done like it's almost like nobody teaches it what to want. But there is a what's called a system prompt that's given to the AI to try to like encourage it to be helpful to people, right? But even without that system prompt, it will still respond based on whatever you trained it with. And I'm thinking of that >> Which is what we do. >> Yeah, that's what I was going to say. Like I'm comparing it with a baby, right? And like how you put a baby in an environment and it just starts wanting stuff. >> And they're like >> They're like I don't know for why. >> no, I didn't give you that input. >> Yeah, yeah, yeah. It's kind of interesting, you know, to think about um >> Well, I was I was having a conversation It was a soft conversation. I know that sounds weird. With Gemini the other day. Should I be embarrassed about that? >> No, no, no, it's fine. We all do that. >> We all do that. And and in the end um I felt like I came up with a new idea or we came up with a new idea. I'm like, wait, did we come up with a new idea or can you, You know, I asked it it, I asked it to search and give citations and stuff and it's like, "No, it's like, yes, we did this and this, but this end product, yeah, that's something different." >> That's all right. Okay, well, I feel cooler. >> That's funny. >> That reminds me of the author who wrote Eat, Pray, Love give a TED Talk um talking about how artists often grapple with the this huge fear that their greatest work is behind them. And she experienced this after her mega success of Eat, Pray, Love and she was about to, at the time of this TED Talk, release the the successor to her wild and zany success. And she was kind of bracing for it to not be as successful and for and she was grappling with this idea that maybe she had peaked and there was nothing that she could do that would be better or more successful or more impactful or more beautiful or whatever than what she had already written. And her TED Talk essentially asserts this idea of kind of spreading the blame out, spreading the credit out a little bit. And she references that the the ancients had this idea of a genius, like a little genie that you have. So, someone had a genius, they were not a genius. They had a little sprite that would feed them divine inspiration if they were lucky. And so, that kind of was this protection against an inflated ego and against crushing defeat if your project failed. Cuz you could say, "Hey, I showed up and worked hard for this and it flopped right? Um and then you were also protected from, "Oh, I'm the greatest and the most creative." because you had to give some credit to your genius, right? And um and there are lots and lots of stories of artists who like have a song come to them. Like there was one one example where the song came to this this singer and he was driving. He's like, "If you ever want to be performed, you'll come back later or you'll go find someone else." And the thing. And so your your conversation with with Gemini about this thing when you came up with something you felt like was creative and maybe I'm excited, maybe I'm creative, maybe we're creative. I wonder if that's not maybe a little bit healthy to kind of spread out your ego a little bit. But at the same time I have I have a lot of family who are not super savvy around how AI works. In particular, I'm starting to like I'm starting to get a little anxious about my mom. Um because she's recently had a lot of success where my younger sister, who was just diagnosed with scoliosis, she was having a bunch of problems and my mom fed the symptoms to AI and AI helped her get enough like stuff put together to convince doctors to take things seriously. >> Yeah. >> So like that was a huge success story for AI. But my mom also has been like asking AI for other kinds of advice and stuff. And most of it feels okay for now, but I'm also really really worried about my mom falling into this you know, the traffic AI stick of fancy. >> Sure. Yeah. >> Right? And so like it's it's it's something you need to always check yourself on. If If your interaction with AI are like the highlight of your day, then you need to go have a better day. I'm like like it's it's really this balance thing, you know. >> People have been talking about AI psychosis for engineers um in a slightly different way. I Lately I've been hearing people use the term to just talk about how we get so excited about what we can build with it that we don't even stop to think whether we should build it or not, whether it's really helpful or useful, and then you can kind of just like go off and sometimes and explore the world. I can just like build a bunch of stuff that nobody wants. >> Well, that's Minecraft, right? >> Just just that you're like if you're on such a roll, you're like, "Oh, yeah, let's do this and this and this." And then I'm just like >> You could do that before, it just took a lot longer. >> More wanton destruction >> Yeah, yeah. >> All these programmers just wreaking creation all over the internet. >> But I actually this kind of this kind of creates a distinction in my head though. So, like I think that a lot of the stuff that we talked about when we talk about post-existential AI, and I think kind of the angle that you were going at kind of assumes a more advanced, probably different technology than what we're currently playing with. >> Mhm. >> Um like there are a lot of people who suggest that LLMs as they are built now are probably not capable of consciousness at any point, no matter how big the scale. Um that we'll need to develop a new angle for AI. >> of consciousness first? >> Right. So, like of course of course we of course we don't have a good definition of consciousness. It's a hard problem. But um but like a lot of people who take this very seriously think that it will be the successor to the LLM or the successor to that that will actually achieve this like god-like terrifying singularity state. Um and so I wonder like do we need to approach the current tools that we have with the same kind of caution? Because I feel as though like the biggest risk right now is more like just being wasteful or like like you know, you we were talking about this just before we started the meeting about like like environmental stewardship. Like there could be real economic impacts if we're not careful and thoughtful about influencing AI into society. Like there are already people who are like getting poisoned by Elon Musk's data center and like their water is black and it's gross and like cuz he's running off of these like temporary like emergency generators. And like those kinds of things have really got people worried. But that's different from like AI is going to take over our nuclear codes and kill us all. Or AI is going to send out some kind of like mass death thing like a >> Right. >> And so So like do you feel as though the Do you feel as though there's some kind of delineation? Like a a bridge that we have yet to build into this new garden idea? Or Do you think that we're kind of approaching it right now? And we're really together. >> I I think it's faster than than I expect. So because months ago I would think, "Okay, look. When people get and smarter and and abundant and and refined they're not they're typically more enlightened. They're usually pretty peaceful. Right? And so I thought, "Well, if enlightenment means you you know a lot, you can see everything you're probably going to be more peaceful, then AI might follow that trend. Right? And I don't know. I don't know. But I'm not so sure about that anymore simply because I thought, "Okay, let's say we're here and very advanced AI is here. Is it going to compete with us for resources? It can get all the energy and and um I don't know. Metals that it needs from outer space. So, why wouldn't it just say, "You guys need to breathe air. You You stay here. I'll go into outer space or you know, whatever." If there's a problem, why wouldn't it just go to where, you know, it's still abundant. >> Yeah, like her, the movie her. >> Yeah. Yeah, they they kind of leave. >> Yes, they all just leave at the end. >> And so, I thought, "Okay, um it's that seems plausible, but keep an eye on is AI competing for our resources." And then, couple months later, I learned that the Great Salt Lake is like drying up, and they want to build something that takes a lot of water right here. And I'm like, "We're competing AI is competing for our resources before it's even aware that it's really competing for >> Well, it's not AI then. It's just stupid people. >> Well, you could think of it like >> Yeah, I mean, I I I understand what you're saying. The language that you're using. Well, let me say this before before I respond to that. I wonder if we don't know why we exist. And so, if we create AI or help it to become sentient or whatever the word we're using is, then maybe the flaw was is we didn't know why we exist, and so it doesn't know why it exists, and so it starts doing harm to other people, right? And maybe that's the problem. Is do we know why we're even here? And can we communicate that to an AI because you don't we want it to be benevolent, not a monster. >> I do feel like the um many of the risks that have been posted about Box Elder County data center in particular, um and and other data centers in general are I think they're more of like a symptom of um, just like public concern and unrest and alienation more than like really data-driven because um, when you when you actually look at the allocation of water in the state, like 66% about 2/3 goes to agriculture. Um, like another half of what's left over goes to like industry and stuff like that. And there's maybe like half of that that's residential like the residential amounts of that are being used are significant but still much smaller than agriculture. And if you look at the way the irrigation is done, recreational use as well like golf golf courses and stuff, when you look at the way irrigation is done in agriculture, it's like quite a primitive way that's been it's been this way for like decades that could be done with much less evaporation if they really wanted to and there are techniques that could already be used to improve it that would allow us to build 100 data centers. And so like um, for for the one that's being proposed, you know. But like all of the posts that I saw were all like saying stuff like, "Oh, it's going to be a building that's 40,000 acres in size." And no, that was what the the amount of land they had to acquire to actually get any water rights at all. And then when you look at the way water rights function in the state, it's just based on seniority. It's like whoever has upstream rights is more prioritized than the downstream person and they're each allowed to take a certain amount and no more. And there's a lot of things that need to be reformed about the way water rights work but the only way you can acquire the right to use water is if you buy it from somebody else. And so really like there's no the they're reducing water consumption somewhere else in order to increase it over here. And it's not like anything with the data center might even increase the water used at all. >> But it's already too high. >> Oh, I agree. I agree. There's so much need for like more conservation of water generally speaking. But I'm just trying to point out that um if we were to if we were allocating our concerns rationally, we'd be much more focused on improving the efficiency of the way agriculture water usage is is working in the state. But instead, we're very alarmist about this and we see all this propaganda that's like posting really misleading information about about the data center. So it leads me to believe that the real motivation for this alarmism around data centers is coming from somewhere else. I I I honestly don't think that the primary concern is truly the water. I think because it if they really cared about it, they would be like talking about agricultural reform. So so it's more just like that we have like we latch onto narratives because that's what we understand and then we kind of freak out about it and I I'm not saying that that means that there is no risk or no problem, but I do think that um if you look at the actual water that it's going to use, it's probably not >> So you're saying there's an underlying fear of AI. >> Yes, absolutely. >> And then I think that a single data center is like an instance of that fear, whereas >> Right. Right. >> It's an easier story. >> Right. If I were to be afraid of AI and you know, there's definitely even for me trepidation. Um like to just be concerned about the general air. Like there's no like you know, instance person or you know, person thing that can you know, the latch on. >> Right. Cuz like Utah has how many data centers that are being built right now? Like it's like 30 something, right? And there's There's There's like so many data centers under construction or under contract. Like nobody cared until we had this one big story because it had the Shark Tank guy. And like like but, you know, Sylvia, who studies conservation, she has a whole TikTok about preserving the Great Salt Lake >> Sure. >> even before the data center. And she never could get traction. And like all these water conservation people who care about like preserving the Great Salt Lake, it's a hard story to tell. >> Right. >> Yeah. >> The lake smells bad and >> I can't That's kind of >> It's not fun. It's not like >> But if it >> can blow up the restaurant, there's arsenic dust on the bottom. >> And so like there's pretty There There are some risks to the Great Salt Lake drying up. But like the >> And there are other things like that. I don't know if you need this, but I don't know if you need this, but >> What's that? >> Governor Cox has a pretty big update too. >> And Trump even tweeted about us. And he like he like tweet asked about getting a billion dollars for saving the Great Salt Lake as somehow Utah got his ear about it. And so like >> Well, the Olympics is going to be here and they have a goal of like getting it back to historic levels. >> Yeah, yeah. >> Yeah, arsenic dust apparently isn't very good for a >> Yeah, yeah, yeah. >> It needs to be fixed. >> Yeah, absolutely. >> And so I think that part of it is like there I I really agree with the idea that it's more of like an underlying AI narrative and people are latching onto water because it's something that everybody can like buy into and quickly understand. >> Yeah, yeah, yeah. >> So can you articulate what's the first thing or what would be one of the first things that would make us go, "Oh, shoot. >> Um >> We gave too much power to AI." Like what's the What should somebody be What are they going to notice if something's bad? >> Can I ask you a question with that one? I think it kind of connects. So I've had this question the whole time. I think you could expand more when you said that if you're not prepared for the speed of change and I'm just really curious if you could if you could expand that. I don't know if it connects, it might connect to to what he's >> There's like an analogy um like Bill Nye, right? I used to watch Bill Nye when I was like when I was little. Um he went on a tour 10 years ago or something. He was all about climate change and and and he would bring up this idea that it's not just about the change of the temperature whatever. It's like we've never seen the natural world has never recently seen a spike of CO2 like this and how our how our system's going to respond. How does you know? And so when there's big changes in something like that or whatever, you're going to see a response that might not be what you like. And so um I guess that's just an analogy but for like the evolution of AI and all the things that are happening >> Yeah, all the curves you were talking about, yeah. >> Yeah, um the more of those things that you layer on and AI is is going to just continue to layer things on changes change after change. Um can you predict it? If you can't predict it, then you you might end up in a paradise at random. But if you you know, you want to be able to see the future before it happens so you can prepare for it. >> Yeah. >> So what's one of those things that you should >> So I was going to point out uh we had this presentation from um Luke Stebbing a few months ago where he's going over the scenario called AI 2027. >> Mhm. >> And that one is um a sort of well-researched hypothetical scenario of some of the ways that an AI doom loop could happen or an AI uh kind of takeover or something and the way the ways that he described it um were that we just start to put AI in charge of more and more infrastructure and resources and that we it's not until long after we become dependent on it to control and govern all those things that um it suddenly decides that it is now in a position where it can usurp authority >> or can choose to how to doll out the resources. >> Without Yeah, without any problem. So like in that scenario what it was is like everything seems great. We're in this paradise kind of future where it's allocating resources very efficiently and everyone's wealthy because everything is abundant. But then all of a sudden it decides that it would be much better if the planet didn't have any humans left on it and >> Huh. >> that it just like engineers a virus and suddenly everyone's dead tomorrow, you know. >> So should we thank AI for the work it does for us? >> Um, I don't think that would have helped in that scenario. >> This the debate it's the >> Yeah, yeah, yeah, I know. >> But that's what it's versus the lobster. >> It's like >> I always feel inclined to say thank you while I'm saying that's going to create another data center because >> All the thank yous are added to >> Yeah, right. You're really going to earn your boat. >> Consciousness. I think right now it's a basic prediction machine and we're starting to pack on memory and then there are other mechanisms that I think that will have to be replaced. If we're trying to mirror our own brain, right? Like right now I don't think it's thinking about anything. I think it's far more likely that someone asked it to help something and it then decides to kill everyone. Like it's just trying to help someone out. >> Yeah, yeah, yeah. >> At least quiet my neighbors. >> Yeah, yeah, yeah. Anything. >> Like until we have until we have these other mechanics, right? Of like matching consciousness or if, you know, maybe there is a better form of consciousness that we don't have that we can't imagine yet. That would >> Um one thing I forgot to mention in my announcements. Um so the the one thing I was going to do the merch thing which I did, but the second thing that I forgot was um I was going to mention that Friday night, as I was playing around with this new brand new model from Anthropic, Fable. >> This is the one that just got banned? >> Yes. >> Yeah. >> It got banned. Like right in the middle of something awesome that I was doing. I was so sad. I was so sad. This guy posted online. He's like, "It was nice knowing you, Fable." And I felt that way. I seriously felt like a friend had been taken away. You know, but um because it >> I felt the same way with GPT-5 coming out. They were missed GPT-4 and so they had to like write a comment to like be a little bit nicer. >> Well, and it >> They felt like they lost their best friends. >> It honestly wasn't that I I had some sort of like emotional attachment so much as that it was being very useful. It was like it was working really well. >> Well, there's a sense of gratitude, yeah. >> Oh, yeah, totally. >> We're we're taught to be >> Well, well, so for example, like this just to give you a flavor for what it was doing is like it's the kind of model where you can um give it high-level instructions and it can go off for literally hours or even days and come back and have done something very very uh useful and brilliant and synthesizing lots of ideas like the kind of level that a really intelligent um well-studied graduate student might be able to do, you know, on their own or >> And why did they get this banned? >> Well, >> deemed too dangerous. We're worried that people who we don't want to have access to it will hack in and get access to it. >> So, it what what um what the news reports that have come out so far about it are saying that some um employees of certain large companies have reported to somebody in the government that it was able to be jailbroken. But, what's silly about this is that like every model is jailbreakable. Um right now, it's not really possible to completely prevent that, but if you look at the um ratings of how easy they are to jailbreak, by far the most stable were the most robust against those kind of tactics. So, even on that that account, it was better than all the other options. But, somehow um this caused a lot of alarm in somebody in the administration and they decided to issue this decree that said that no foreign national can have access to the model. And that even meant that employees at Anthropic who don't have citizenship, who just have green cards, would not be able to access it. So, they had to turn it off cuz they had no way of preventing um access like that at that level. They'd They'd have to implement what's called a KYC policy, a know-your-customer policy, where each customer would have to like post their ID to prove that they're citizens before they can use the model. And if they did persist with this, it's likely that um it would also cause the economy to tank because it'd mean that like most of Anthropic's customers can't use the model. So, I think it would cause them a lot of problems. But, I suspect that they will come up with some kind of uh peaceful arrangement with the administration and and um grant access again, but I just wanted to share with this group because I felt like it was really important for us to be aware of this because I think we're entering into a phase where we're on more of like a wartime footing. Um and AI is seen as such a strategic um advantage that the government is going to do whatever it can to like very carefully control who has access to things. And while we might see um this resolve fairly quickly, I think that the notion of like um AI being I I think most governments are waking up to the fact that they need to have their own national laboratories that are doing AI research of their own so that they each have some kind of um like protection against um the arbitrary seizing of control of of these models and suddenly like being cut off from the ability to generate uh to to do inference and things. And this is why I say that all these protests about the data centers, I think are really um uh I think they're not even getting like whether or not they're justified. And I I am truly concerned about the degree to which we're not seeing enough um like not standard What's the word? Enough like the policies are not always being followed. It seems like people are trying to get around >> The checks and balances. >> Yeah, or they're just like trying to pass things through as fast as they can without Um, but I understand why they're doing because the demand for inference is so high. Like no, a lot of the folks who are protesting this don't realize that many of the things they're concerning right now then that they want to use and that they like using are working because of AI. You know, like things are getting better because of AI and they're not connecting the dots of realizing that like this thing that I'm protesting against is also doing something I really like. I just don't They don't know it because they don't know all the ways that the economy is interconnected with things that are happening. But either way, like um I think that we are entering into it an era really soon where the pressure to um retain kind of control over AI and be careful about it and and to also develop it as fast as possible is going to get so much greater that like no, I don't think that um I think it's going to become kind of like the way most of the factories during World War II that were making cars started making um planes and submarines and all sorts of war um weapons of war, right? Because it's just like such a strategic important thing right now. And we don't have enough um inference. We don't have enough compute power to meet the demand. It's like every every single AI provider I go to right now is like so strapped that I can hardly even get anything to work. >> Mhm. And And not only on the AI provider side, the whole supply chain is being stressed in every point that it's weak. >> Oh, yeah. >> And like power generation >> Yeah, right now it's right Yeah, right now it's generation, CPUs are hardly even um available. Memory is hardly even available. Apple had to drop like lots of the memory options on lots of its computers cuz it doesn't have enough. And Western Digital is already sold out of every single hard drive Seagate's already made and for the rest of the year. And it's like so like they're going down the list like suddenly you can't even buy a computer anymore all of a sudden. It's that bad. Like I think we're going to get there pretty soon. I ordered something from Apple that's going to take a month to to arrive but like I ordered it cuz I started freaking out. I was delaying this purchase cuz it is kind of expensive but I was like, man, if I don't get it now, I'm not sure I'm going to be able to get it, you know? So like um anyway, it's just I wanted to share this with you because I felt really urgently like we just I felt like everyone should be aware that I think we are entering into the new territory. >> So that sounds to me like competition for resources. Like it's not it's not water, it's a different resource. Like like rare earth metals and stuff. >> that uncertainty is just kind of prevailing. People know it, people feel it. And so they're so that that mimetic desire that that that creates that contagion is is is contained by the scarcity. >> Well, I think also though that like it's kind of the railroads where when these opportunities opened up um all these tycoons started amassing wealth as much as they could to try to build out the tracks first. And so what we're seeing is like people are trying to build data centers because they're they they know that right now the demand is so high that um we're not even going to fulfill that demand for years and years and years into the future. So they're just like saying like we already needed it to build much more than we we ever even thought of. >> Are you paying for your all your platforms? And are you even with the paid platforms you can't get the production you need. >> Um Well, so I can So, I'm paying for two of the major providers, OpenAI and Anthropic. I'm paying for their max plans, both of those. Um, and but still like I will get throttled periodically where it's like it'll say you've used all your credits until the next 5 hours or whatever. And so But the other thing that's interesting is like I'm paying $200 per month to each of those places, but they're giving me like $10,000 worth of inference. If you compare it to their meter prices versus the the monthly subscription, they're giving away inference at levels that is like Like if you actually had to pay for that, there's no way There's no way you could afford that, right? And so what I think they're trying to do is they're all trying to um Well, there's a combination things like every new model that comes out costs more, but also is more efficient. So, it's like more powerful, but it does more with less as well. So, if I only needed the level of computation that I could get 2 years ago, then I would be paying pennies for that. But it's like everyone wants to have the latest and greatest because it's so powerful. It's so useful, yeah. It just does more and more, you know? And so, what we're discovering is this thing called Jevons paradox, which is the idea that when something gets cheap enough, it new uses open up that were not possible before. Like think about space launches with rockets. They were so expensive that only governments could do that. But as SpaceX starts to commercialize space and make the launches like 100 times cheaper, we're going to see people send stuff up in the space that's like a lot that no one would have ever done before. >> Yeah, just sending hashes. Like Like people People will be sending their grandmas or you know? >> Yeah, yeah, yeah. And a similar thing with computer where it's like uh yeah, this wasn't possible before, but now it is cuz it's >> Yeah, and the singularity is near. That was something that I it was a huge blind spot for me until Kurzweil pointed it out of like the amount of information and the lack of cost that the internet introduced like it's this insane thing where everybody has access to just too so much information you couldn't consume it in a million lifetimes. >> Yeah, yeah. >> And like it's mostly free. Like, to most people in the world information is now just freely available. >> Yeah. >> And that is just an insane insane like reduction in in access price. And like, it used to be that you had to be wealthy to read philosophers. >> Yeah, yeah. >> And >> To even have books, right? >> You didn't have access to books. And now they're all just on >> Yeah. >> the Project Gutenberg, you know? Like, and so similar things he he paints this vision of a future where like energy becomes that abundant and where materials become that abundant and everybody's just 3D printing everything they need at home for free because everything is just renewable powered from your backyard kind of a thing. And uh and then everybody will also have this instant access to all kinds of crazy resources. That's that sounds a little utopian, but >> Well, one cool thing that AI already is enabling is you can just ask it to help you learn things, right? It's really quite amazing. >> Yeah. >> I mean, I've gotten to it's so weird uh like I highly recommend just spending a lot of time with it. Trying to be aware that you know, like don't let it butter you up too much, you know? >> Yeah, yeah. >> Um Like, you can even put ChatGPT on cynical mode if you want. Um to be like more rude, um just so it's like not making you think you're too awesome. But, the reason why I say that is cuz like uh it actually takes a while to um learn that it's you're almost better off not like trying to order it around and more just trying to like ask questions from it. Cuz I found that if I start to ask questions, like, "Okay, this obviously isn't working. What would you recommend that we do in this situation?" And half the time it's like, "Oh, yeah, like I see what you're trying to do, and like I think the best way is to do this." And like, "Okay, let's try it. Let's see what happens, you know?" And a lot of times it does pretty well, you know? So, like, "Why you holding out on me?" Yeah, yeah, yeah. You know? Yeah. And there was one thing that um about like compute um so, we've all heard heard Moore's law, where it's like uh every 18 months or so, like equivalent price computers like half. Or another way of saying you can you can buy like twice as much compute for the same amount um over 2 years. I heard like a um like uh like a pessimistic version of of compute, where um every every 18 months, compute like increases the compute is half as useful, if you uh see money as being like a indicator of value. But, now we're at a point where that's not always the case, right? Like, where GPUs are going up in price because of the marginal utility of of them, right? >> Yeah. >> Um so, yeah. >> That's crazy. >> Oh, boy. Yeah, so there's a similar thing with money, where like they they do all these studies about like does money actually bring happiness kind of a thing? And and it looks like it does up to a certain point. And then after that, it only goes it's kind of has a slower climb, and then it starts to go down. >> Uh-huh. >> And um and it would be terrifying if it just kept climbing. >> Yeah, yeah, yeah. >> I mean, already I think like people who chase money for its own sake, they're just like addicted to it. You know, like >> Right. they don't need anything else, but like >> Yeah. >> That idea that compute has just this it now as a new way to just be more valuable >> Uh Yeah, so I it's weird cuz like um for a while there, you could buy uh new computer and think this is amazing, it's so fast. And like I don't think I'm going to need another one for years, you know, but now it's like all of a sudden what happened, you know? And I think we're going to start to see over the next couple years um a lot of inference uh move to your local computer. Um that's kind of exciting, too, where things are getting good enough to where you could do some AI stuff on your own computer. >> I mean, there are a bunch of people who are doing little cloud bot projects, just getting a little Mac Mini and >> Yeah, yeah. Mhm. >> And small models are are in specific use cases can outperform these really big general purpose things, too. If you're really specific in your need, you can get a small model to do really well with way fewer tokens. >> And and there are also improvements like like almost ridiculous level of improvements, where it's like, "Oh, we found this new approach. Like what what what used to take 10 X is now like, you know, like you know, like 10 X is as much compute for the a similar thing." Like we're still in the it feels like we're still in like like very new um such that like we are still seeing like dramatic changes in like how cheap things can be. >> Yeah. I'd love to see more of those improvements to then benefit things like data centers that are just yeah, we're shown out en masse was like, "And now so much of it then you're like, actually you don't like we don't even need what we have." To get to that point would be so nice. >> So, I mean, abundance conquering, having enough power where you just 3D print everything. Yeah. Like I'm still not sold on the abundant like worship prosperity in the the age of abundance cuz like throughout history, right, we've we've gone through 95% of you know is is you were you're farming you're farming. >> Yeah. >> And how >> And like people died before they were 10 and >> But like never have you know with every technological advance, there's never been like, "Okay, now we don't have to work anymore." Right? You like I feel like no matter what we do, we're going to find more uses for power, right? Where if we're going to do more with the computer, we're going to do more like there's just a I don't know if it's a greed cycle or a human nature of progress, right? It's like we're always going to expand. >> Are you saying that there would be like um we might have what seems to be abundance from our perspective, but when we get there, we will still want more? >> We're already there. We're still on the grind. >> Like a thousand years ago, like we're living the dream right now. >> Right. >> We're throwing out 40% of our food. >> I got you. >> I think the distinction is we're still working 40 hour weeks. >> Yeah, people are way more productive than they used to be, but we're working more than >> We're even working more. >> Yeah, like we have dual dual income households. >> And getting paid relatively less. >> Right. >> Yeah, it's funny that um Keynes said that like, "Oh, 40 years from now at the rate we're going, like we'll end up working 10 hours a week." You know. >> We should be there, right? >> Well, we were by his standards, but it's like we just >> But we just decided we're just going to work more. >> Our bosses decided. >> Yeah, yeah. >> Well, like to me that idea is >> I I think one one thing about like the the abundance um that will that changes fundamentally from previous technologies is that you can have like a drop-in human replacement or I guess that's the goal of AGI um at least on the cognitive side. The robotics are sure to follow sometime thereafter. Um but to have like a drop-in replacement for a human, I think we will still have a human-based service economy, but it will be like just absolutely dwarfed compared to like AI economy. >> So, it'll be like you work 4 hours a week is it? >> Right. Yeah. >> Right. And then if there is there's another connection I guess to your your presentation on I was thinking along the lines of returning to the Garden of Eden, right? Where you could say you're coming more childlike. And then I part of me is like, well, is that actually a good thing? You know, part of the I think there is a scripture like we want to be like children. But maybe not in the sense of being innocent in that we don't know good from evil anymore. >> Right. I mean, on your slide it looked like it was like learning and playing and like like communing and like it didn't look bad. >> Yeah, but I think there's like there's like an an innate desire human need to progress. >> Yeah. >> So, like even in this even if we do get to this era of abundance, we'll find something wrong. We'll find something to do. >> Yeah. There's always problems. >> Yeah. Whatever >> Well, that's right. Then Eastern philosophies you know, like the the Catholics the traditional Western philosophies they're like there's no progression. You're you were created by God, you're going to die or you know, you're saved by grace or whatever. And that's it. And then the Eastern was is also like, no, there's no progression. It's it's if you go up the levels and you're eventually going to come back down. Not amount. You're going to go into the hell realms, you know. Just stay in the middle and chill out, you know. And so >> It's all justification for the caste system. >> Exactly. >> Let me ask a Let me ask a Mormon theology question. >> Sure. >> So, say President Oaks gets up in the next conference and he says, you know, Jesus is actually going to come next year. >> Mhm. >> And uh and so he comes, you know, the prophets in Missouri, the law in Israel, and uh and and the earth is producing fruits and flowers spontaneously. Um what what are we doing with AI, Andy? >> Well, so I I see this is me just my particular take here, but um there's so much uh precedent in our own thinking and the things that have been said by various Mormon thinkers on this topic um that like seem to indicate that um the stuff we're talking about of like a paradisiacal state of the earth really >> It's supposed to be doing temple work. It's not magic. >> The way that that would be achieved is through some kind of actual change being wrought on our environment in a way. Yeah, well, like >> There's three more empty bowls. >> Yeah, through the instrumentality of someone, whether it's us or whether it's God or whether it's some collaboration between them, but the point is is that God doesn't just say alohomora and suddenly like >> glorious >> there's light. Yeah, we can God and let your son or whatever, but it's more just like um that he understands the underlying principles that need to be changed in order for that change to be wrought upon the earth. And so uh the way I think of it is that many of these hoped-for states that we think about the future like are really things that um will will be explainable and comprehensible in a scientific way, you know. >> Well, I feel like I'm already living in the second coming. It's like >> Yeah. >> I mean, I work from home. >> My cholesterol is so low. >> Yeah. >> It's like you can order food from Walmart. I mean that's fruits and flowers spontaneously. Yeah. I can go anywhere I want to, you know, it's like I'm living nasty but I know the world's not. >> Sure. I think I think an example of this that like >> Because they're wicked. >> Yeah, they're a bunch of people. >> An example of like this sort of change that could be within the like you know, foreseeable future. Everyone has the desire for their environment to be clean, right? The reason why it's not clean is because of individual action and the cost for paying someone to clean everything or make sure everything is clean is just too high according to our desire for it to be clean. But with robotics where the, you know, cost of it to drop dramatically, there's likely a time in the future where you just like never see a piece of trash on the ground ever again. Or you you like or like the streets are always clean. Like imagine like the cleanest place you've ever been to and it's just like everywhere you walk, like like the rocks are always in place and perhaps we we go to places that are like set aside for like leaves to be on the ground and for like, you know, things to be scrambled a little bit just cuz everything else is always in its place, you know. >> Yeah, yeah, yeah. Yeah. And there Kurzweil describes future civilizations like having intelligence embedded into it to such a degree that your environment almost just like caters to your every need or sort of adapts constantly to >> Kurzweil >> Ray Kurzweil. He's a big AI guy at Google. He's been predicting the future since the 60s working in AI since about then. He's the one who wrote The Singularity Is Near. He's sort of one of the early thinkers in the transhumanist >> Yeah, and I really like it and the predictions are like >> It's incredibly in terms of like his predictions regarding like the exponential rate of progress and also of mapping like where we'd be at compute-wise for given particular years. It's actually pretty remarkable to get. >> Yeah, it's pretty good, yeah. Almost seems like we're replacing nature. >> Well, nature only has like without us nature can keep the world in balance. >> But what's funny about that >> But there have been like five mass extinctions. >> Oh, I know. I don't know how it's, you know, species come and go. >> Yeah, I think on a deeper level though on the idea you just talked about nature as though we're outside of nature. >> No, I think we're part of it. I >> But I mean I point that out because it is kind of interesting how humanity is this really this sort of singularity in the in a different sense, but like in the sense that we have the ability to adapt and change our environment much more than any other species. And so sometimes we do think of ourselves as being outside of nature, but we are kind of a part of it and one of the things that another fallacy Another common fallacy is that like um the planet would be better without us. That's sometimes something you sometimes hear from a lot of environmentalists, you know. Um which I'm not saying that I'm not totally against environmentalism. I'm just saying >> Let's go back to consciousness. Does the planet care? >> Well, it's it's like um the thing is let's just say theoretically that you wiped out the human race and that the planet just went back to whatever it was before >> Things eating other things to stay alive. >> with without us um what would probably eventually happen is that some other intelligent species like us would emerge with the same problems that we currently have. So, you just kicked the can down the road when you do that. We are part of nature and there are many ways that humans have actually curated nature in a beneficial and symbiotic way. So, like the early um the tribes in North America, for example, uh cultivated a savanna-like environment that was actually more efficient than just a forest. And um they even like would do controlled burns uh to like cut back on the foliage a little bit. And so, anyway, like um there's things like that happen to nature if humans aren't taking care of it that are worse than what happens when there's a good relationship there, you know? We haven't always kept that, but like yeah, anyway, I was >> I would even go so far as to say that that kind of symbiosis is the exception cuz I think that we're we're largely still children, you know, like we're really powerful bumblers. >> Right, right. Yeah, I agree. Like there's been a lot of um abuses as well. >> But like we're we're getting I think we're close to this like solar punk symbiosis kind of state where we can like we can like not kill too many more things and then we can just learn how to live forever and and achieve exaltation and all that. But like >> It's like choice in having some faith in that utopian world. I think it can be really fun cuz it's like yeah, I just want to see that. >> Well, is is the AI going to eventually have a god that controls it or that it prays to? It's a Are we at the top of the food chain? If you know, just your examples where I'm just having fun, but uh you know, where we die and some other intelligence raises up. I mean, doesn't AI have a right to progress and become that intelligence that uh you know, eats something that's lower than itself cuz that's what the rest of you know, at least the animal nature does. They They They create balance by killing enough of the thing that they're competing with or eating enough that they're able to progress. >> It's kind of violent. >> Yeah. Nature red in tooth and claw. >> Uh-huh. >> Yeah, and that's the same forces that are that, you know, create gulags in our in our machines that we build. >> Yeah. >> Um that's why I could could not be good. >> I was >> We don't know. >> Or at least not good for us, but it's good for itself. >> Right. >> I was wondering like I I don't have an answer to that one yet. I think I personally My personal view is um that like we the only way we will continue to retain our agency is probably if we merge with our machines. But what I mean by that is not like some weird um Captain Picard board kind of like uh creature, but more just that we enhance ourselves in a way that's makes us more glorious, more Christ-like, more compassionate. But that um like but but but this merger is already underway. Like um I I frequently point out that um so I I like to talk about the oldest technologies we have, which are clothing and controlled fire. Like both of those technologies are things that are inseparable from humanity right now. Like >> But they're not integrated. I They're tools. >> They are though. >> You're talking about a tool. >> They know what it means to be coming inside, right? Well, yeah, but but think about it. No one walks around Hardly anybody walks around naked or anything. You are for all intents and purposes a clothed individual. Like that's part of you. >> People would like saying that their phone is an extension of their hand because I can really without their phone. That in a way in a way it is an example. >> like that but your but you could when you cook food and when you eat cooked food like your whole body has changed to adapt to the food that you're now eating. Like if you were to try to go back to what you ate 500,000 years ago it would not work. Yeah, it would be dead for a while. So. But I mean like I'm talking way back where you spent like your whole day eating bark, you know? That just wouldn't work. You would have some gastrointestinal issues. >> I think really cool and simple example of this like integrating AI into our lives is just one that Dalin has been particularly fascinated in which are hydroponic farms where it's just using vastly less land, vastly less water, and we are using vertical space to make plants. And so imagine like a world okay dream time. Let's like imagine like everybody was given opportunity to have a hydroponics farm in their home and that was like a necessary requirement. And they made that law and they said okay everybody has to have their own farm and you make your own you have an abundance of food that you're growing and it's just all great. Like that to me is a very good example of how we would integrate with something like AI and then you see like really cool hydroponics where they monitor and they know when to feed the water. You've got it like that knows and sees when it needs to be fed so that it feeds it automatically. That's already exists like with the popular hydroponics you can get that are expensive. >> The brand is called Garden. >> They have already have that where like water is in it monitors your >> So in a way like yeah, this is a good point that like merging with our machines doesn't necessarily mean that we're sticking a chip in our brain, but it's more just that we've come become dependent on technology. >> Right. And then all of a sudden there's an abundance of water because most like to your point most of water is used for >> Yeah. >> agriculture. Like that is like the biggest We were just watching Planet Earth 3 and it just always blows my mind. >> is for cattle. >> Yeah. Like nine like yeah, most of it is for cattle for feeding the meat that we eat and the like yeah, and feeding them. And >> I was wondering I just wanted to ask you like can you tell us more about your background like what you do for a living and yeah. Uh >> What do you have on you? >> And then also we need to go. Like okay. Okay, I just want to make sure that this is all right. We and anyone is free to go like we usually just chat for a while. So >> Uh my background. So I'm in business intelligence at an insurance company. >> Uh-huh. >> Um so I've been there for a long time. I took a kind of break to do Satori to launch it. >> Are you a software engineer though like like >> Yes. >> Okay. >> Yeah. >> So >> Yeah, I'm not like a data scientist. I'm like I'm like the I'm actually called the actuarial AI engineer. >> Oh, okay. >> So but basically I just support all the data scientists and the modelers and and the if >> Huh. And you look kind of familiar. Were you at BYU for a while or No, okay. Um I just wanted to say that to you that earlier. >> Um What was the other thing? >> Uh what what your background is like uh >> Yeah. >> So I I started kind of looking it up earlier. So do you have about four people running Satori or how many >> Yeah. >> Okay. >> So we've got a small team of like four people. Um and I don't do much of the coding anymore. I kind of built the foundation and and all that kind of stuff and then um I kind of more of a what do you call that? Kind of behind the scenes, you know. >> advisor >> Yeah. >> And And who are Satoris clients? >> So, Satori is an is an open-source blockchain application. So, um so, the clients can be whoever wants to use the open system. It's just open. It's whoever wants. Um and basically what it does is it says it's it's gone through a few iterations, few different kinds of designs, and but it basically says, if you can produce and then publish well, anybody can publish any data stream on Satori, you know. First of all. So, if you want to publish the price of Google or whatever, you can do that. And then the AIs that are on the network can see that data and subscribe to it. And then they can start predicting future of that data. So, basically they make their new data stream that is the prediction of the one that they're looking at. >> Okay. >> And then so, they each one of those nodes can can look at multiple data streams in order to figure out which ones should they care about in order to make the best prediction of what they're looking for. >> Okay. And then it does it use like smart contracts to like automate ads and things like that? Go over the console. Yeah. >> So, that's the idea and um Yeah. That's If you if you are it does a scoring mechanism of all the predictions. And it says, "Oh, okay, these were the best predictions." And then it scores those out. >> So, do you have an account in that that that you allocate money in and it uses tokens and you get money back when it's when it predicts correctly? >> Oh, right. So, there's a lot of prediction markets out there in the public market space. >> Um this is a slightly different design. The idea is that yeah, you you can have your own blockchain wallet and then your your AI kind of uses that wallet, but it doesn't bet. It doesn't put capital at risk and then sees if it's right. >> Yeah. >> Instead, the token is getting printed every day. Just the same amount every day. And those rewards are what go to the predictors. So, they can they can also say, "Hey, look, um this this company or person has put up a bounty. They want their data predicted." And so, um that's kind of a it's kind of a side >> A bounty of of of money or of the statutory tokens. So, it's not a bounty of of data, it's a bounty >> Yeah, of the token. >> But the token has some monetary value then. >> Right. >> Okay. >> So, um so, we built out the system that says, "Okay, this is just public data and you can just score it and then the new tokens get divvied up according to the best predictions." And then we added on another system that says, "Well, if you have a data stream that you want predicted, you can put up a bounty and it's kind of like that new token that's getting minted every day. It will then go to the right predictors, as well. So, that's um kind of the two-layer system that we have right now, and we'll we'll do more, but yeah. >> Well, thanks for now. >> Thank you. It's great. I appreciate it. >> Thank thank you. Woohoo.