← Tom Bilyeu's Impact Theory
Tom Bilyeu's Impact Theory
Is Reality Real? - New Science On How The Universe & Consciousness Aren't Real | Donald Hoffman PT 1 (Fan Fav)
01:21:57
Loading summary
Transcript151 lines
- [00:01]
VRBO/Grainger Advertiser
This summer, don't squeeze in. Spread out. Find homes big enough for your whole guest list on vrbo. From family reunions to trips with friends, VRBO has spacious summer stays for every group size and budget. That's vacation rentals done right. Start exploring on VRBO and book your next day. Now, When you manage procurement for multiple facilities, every order matters. But when it's for a hospital system, they matter even more. Grainger gets it and knows there's no time for managing multiple suppliers and no room for shipping delays. That's why Grainger offers millions of products in fast, dependable delivery so you can keep your facility stocked, safe, and running smoothly. Call 1-800-granger. Click granger.com or just stop by Granger for the ones who get it done.
- [01:03]
Tom Bilyeu
Today's episode is with legendary cognitive scientist Donald Hoffman. By the end of this episode, you are going to question everything you think you know about what is real life. Hoffman believes that this, your entire life, is all a simulation. You, your parents, the food you eat, your kids if you have them, your favorite restaurant, all of it. It is all a simulation. When you look away, it all ceases to exist and is only re rendered when you look at it again. His belief is that the mind creates the body, not the other way around. You heard that right. Ooh wee. We go deep down the rabbit hole of consciousness, what it means to be living in the real matrix, and how he can prove that all of this is correct using mathematics. This episode is wild and by the end you are either going to assume Donald has lost the plot or you're going to reevaluate your entire life, and quite frankly, life in general. And speaking of reevaluating your life, if you haven't already, be sure to subscribe to the ad free feed of Impact Theory on Apple Podcasts. When you subscribe, not only will you get ad free versions of new episodes, but you'll also get access to curated playlists of the best of Impact Theory in categories like Health, Relationships, Business, and Finance. And you'll get access to additional bonus content you won't find anywhere else. Subscribe now on Apple Podcasts Now. Buckle up for Donald Hoffman. I'm your host, Tom Bilyeu, and welcome to Impact Theory.
- [02:30]
Donald Hoffman
What we are are avatars of the 1. The 1. Awareness is exploring all of its possibilities through different avatars. So somehow there is this field of awareness that is, in some sense, deeply and fundamentally who you really are.
- [02:45]
Tom Bilyeu
What do you think about the AI scientists that signed the paper that we need to slow AI down Because, and I had one of them on the show because it passed a touring test faster than they thought. It's just moving faster than they expected and they're very worried. Do you think that AI will ever become conscious?
- [03:03]
Donald Hoffman
I'm actually not too worried about AI right now myself. So I'm not one of the alarmists that that says we need to stop and worry about it. The thing that would alarm me more would be if there were some kind of law that criminalized most people from doing it and let a few people do it, a few companies do it. That that alarms me. So if there's going to be any kind of laws, they should be universal and no one should be excluded. Is that.
- [03:28]
Tom Bilyeu
But why aren't you worried about AI?
- [03:30]
Donald Hoffman
It's pretty easy even with chat GPT to give it questions it can't answer right now. It's basically a good statistical analyzer. It's not deeply intelligent. It will find things that we humans won't find in medical searches and so forth. But that's because it just can't handle more data and do more statistical analysis than we can. But it's not deeply intelligent. And the founders would tell you that it's fairly straightforward kinds of algorithms. And in terms of consciousness, there is no theory right now of any kind that can start with physical systems like circuits, software, and explain even one specific conscious experience, how it arises. So I'll be very, very clear. There's no theory on the planet today that can start with an artificial intelligence and a description of some kind of circuit or some kind of software pattern of activity and can give you a specific conscious experience, like the taste of chocolate or the smell of garlic, where you would say this pattern of activity must be identical, must be the taste of chocolate. It could not be the smell of a rose. There's nothing on the table and there's nothing even close. So if AIs can be conscious, there are no theories right now at all that could explain how that could possibly be and nothing that makes it even plausible. So I'm not too worried about AI's being conscious. I think that they will eventually outperform humans in most everyday day activities, but simply because they'll have more compute power and can search more deeply than we can will.
- [05:14]
Tom Bilyeu
So for people that don't know you, I'm going to give a super brief synopsis and by all means put in where I go awry here. But you believe that this is all a simulation. We are living in a simulation. None of this is real. Space, time itself is not real. We are effectively living inside of what you call the headset that everything you've ever known or ever experienced is all effectively an illusion. It is a computer video game by way of analogy.
- [05:44]
Donald Hoffman
Right.
- [05:45]
Tom Bilyeu
Given that and an audience listening at home, you will notice he did not say no. So, and this is something I've. I have forever just dismissed out of hand that we're living in a simulation. And I say dismissed out of hand because I don't have any evidence to back it up. And I've heard all the arguments from a mathematical perspective that if you believe that humans are capable of creating photorealistic simulations and you give any rate of progress whatsoever, we will eventually create a simulation. We certainly with AI and how rapidly it's been advancing, I think people now really have a sense of, whoa, we really are going to be able to do this. Apple Vision Pro certainly gives an indication like you will really be able to create some very compelling, very realistic things inside of a visor. So I think people, more now more than ever, could see how we could get into a simulation. A simulation simulated world. That's convincing. I'll leave it at that. And if that's true, then why would we, then once we create that simulation, not create another simulation? And I will just tell you as somebody, the T shirt that I'm wearing is literally about this. We're building a. A game that we hope over time will be a truly simulated world that people will go in. They will have an identity inside that game. Okay, so if we know that loop exists, then once the game inside the game gets powerful enough and it will do another simulation. Once the game inside that game, inside that game gets powerful enough, it will do a simulation. And so you end up in this point where just mathematically it would make more sense to believe that you're in one of those, you know, conceivably infinite recursive loops of a simulation than that you're in base reality. But it just always seemed weird to me to say, no, no, no, we're in one of the simulations. But the more I research you, the more I'm like, maybe we really are in a simulation. And to that point, you talk about consciousness as being fundamental. And so I'll need you to explain that for people that that will be so jarring, it will take them a while to really grok that. But that consciousness is fundamental. So couldn't AI ever become a window into what you call a conscious agent in the same way that a human child is or a dog is or
- [08:05]
Donald Hoffman
whatever that I think is possible. Absolutely.
- [08:09]
Tom Bilyeu
So if you don't mind, walk people through how it could be possible that physicality, everything they see, touch, taste, the loves that they have, all of that is a simulation and not fundamental meaning it arises out of something else. But consciousness is the fundamental. Yeah, the foundation.
- [08:29]
Donald Hoffman
Well, there are two arguments for the idea that what we see is not an objective reality that exists independent of us and is there prior to when we look at it. So in physics, the Nobel Prize last December was given to three physicists for the experimental testing of a clean prediction of quantum theory that something called local realism is false. Local realism is the claim that physical objects like electrons have definite. So realism is the claim that electron has a definite value of position, momentum and spin when it's not observed. And locality is the claim that those properties have influences that propagate through space time no faster than the speed of light. And the conjunction of those two claims, the properties exist even when they're not perceived, even when they're not measured, and they have influences that propagate no faster than the speed of light. That's local realism. And local realism is false.
- [09:32]
Tom Bilyeu
How did they prove it?
- [09:33]
Donald Hoffman
So that's why you get a Nobel Prize. So John Clauser, Anton Zeilinger and Alan Aspic, over decades, there's a string of experiments that were tighter and tighter. Each experiment closed loopholes in the previous ones. So the experiment experiments have to deal with. They're complicated experiments. I mean, Zeilinger was actually using photons from outer space to get entangled particles that they could use, that you could, couldn't argue that they were somehow, you know, being connected or correlated some in some deep way. But basically the, the, the experiments are set up to show that properties like position or momentum or spin, typically they like to use spin in principle, could not have definite values until you actually measured them. So one way that they do this mathematically are there these bell inequalities. And so if, if the statistics of the correlations between the particle spins, you have two different particles that you're measuring, the spin axis, for example. And if they had definite values even when you weren't observing, you'd have certain pattern of correlation. And if quantum mechanics is right and those values don't exist until you measure them, then you have a different pattern of correlation. And so that's what they, they do. They have to look at a bunch of different measurements, look at the correlations, and the correlations come out to be what's what quantum theory predicts and not what our classical intuitions would tell us. And so the, this was done by Clauser decades ago. But it's so counterintuitive that people are going, okay, well there must be a loophole here. So then they closed a series of loopholes and finally they started getting photons from like distant galaxies where the photons couldn't possibly have certain within space time causal connections and close that loophole. And so that's one, one, one direction. So physicists tell us that local realism, at least for microscopic, you know, subatomic, recently they've gotten up to groups of 700 atoms, I believe. So they're showing that these effects, these superposition effects of quantum theory are not just at the very, very small end of things. So local realism is false. Now, one can still try to say, well, but that's for really tiny things. But at the macroscopic level, maybe local realism is true. And that leads to a problem because there's no principal distinction in quantum theory between the microscopic and the macro. You can't say it 10 to the minus 20 centimeters, that's the limit. There's no boundary between micro and macro. And this is a well known open problem. So that's one direction. I'll just go with that. Now the other direction of argument is from evolution by natural selection, where you can ask a technical question. Evolution shapes sensory systems to guide adaptive behavior. So that means to keep you alive along long enough to reproduce, right? So you have vision and touch and hearing and smell, and they've been shaped so that you're able to get the food you need, mate and stay alive at least long enough to reproduce and pass your genes onto the next generation. That's the standard story of evolution. Many theorists also think that evolution shapes our sensory systems to tell us truths about objective reality. Like when I see an apple, that's because there really is an apple. And the red color and the shape really exist even when they're not perceived. And so that's notice. That's a step beyond just saying that our senses evolved to guide adaptive behavior. They want to say more than that. They want to say that if you guide adaptive behavior, you're going to see the truth. So I decided with my colleagues Chetan Prakash and Manish Singh and Robert Prentner and others, my graduate students Justin Mark and Brian Marion, to test this evolution is a mathematically precise theory. We have evolutionary game theory. So there's a technical question. What is the probability that evolution but natural selection would shape any sensory system to see truths about objective reality, the structure of objective reality, and it's straightforward to prove. What we do is we look at various kinds of so called fitness payoff functions, maybe payoff functions that are, that are. And we can ask, do these payoff functions preserve certain kinds of structures in the world, like orders, a total order or a partial order, or a metric or topology, or a measurable structure? So we can say we don't know what objective reality is, but suppose it had this structure. What is the probability that fitness payoffs which govern our evolution would actually have information about that structure in the world so that we could actually be evolved to have some insight into that structure of objective reality? And in case after case, the answer is the probability is zero, There are payoff functions that would preserve the structure, but those payoff Functions have probability 0 in the set of all payoff functions. So that means if you're a betting man, you would bet long odds against it. So it doesn't mean that it can't happen, it's just that the probability is zero. And so I take this as a convergence between two of our big theories in science. Evolution for natural selection and quantum theory, quantum field theory. Both are telling us that local realism is false. And so I think a good metaphor then is, as you were saying, like a user interface or a video game where you render on the fly what you need. So I'm looking at you, I'm rendering a tom face and I look away and I'm not rendering it, someone else might be looking at you and they're rendering their tom face. But, but their tom face is not the same as mine. It's going to be at a different angle and so forth. So we render on the fly. And that's what physics is telling us basically, that local realism is false. We render on the fly.
- [16:05]
Grainger Advertiser
If you work in university maintenance, Grainger considers you an MVP because your playbook ensures your arena is always ready for tip off. And Grainger is your trusted partner, offering the products you need all in one place, from H VAC and plumbing supplies to lighting and more. And all delivered with plenty of time left on the clock. So your team always gets the win. Call 1-800-GRAINGER visit grainger.com or just stop by Grainger for the ones who get it done.
- [16:36]
Tom Bilyeu
And so the where you're taking that from is the quantum uncertainty principle. Basically everything has a probability of being in a given state. And the reason that it's just a big question mark is because nothing's looking at it. So it does not need to render that. It doesn't need to decide the System, which is the simulation which people think of as space time. But they're almost certainly. I've interviewed you so many times and I know how hard it is to escape this matrix. But they're thinking of things within space time being real. But once you start looking at space time as purely a simulation, and that the then rendering only happens when you look at something. So that to me makes a hypothesis that I think your data backs up, which if that were really the case, then I understand why big things would adhere to what seemed like a different set of rules where things are static and small things would not. Because you're far less likely to observe a first order consequence of something microscopic. You may be observing a second or third order consequence, which raises questions for me that I'm sure we will get to at some point. But just to close the loop on that. So first order consequence, I can look up and see the moon. I see planets, I see stars. And so for that to be persistent, which is going to be a big thing in our discussion today, this is like the prime thing I want to talk to you about is persistence and what that means. But big things will need to be persistent. And therefore there has to be. There is a constant collapsing of its probabilities because there are so many things that require. Even if it's just its effects on gravity, there's so many things, quote, unquote, witnessing that or measuring that. So I get why those would be stable. But then things where they're so small that there's very little that hinges on that, that that would need to be directly rendered. That would need to. Because you can get away with sort of the probabilistic rendering of the big things and their influence by these smaller things. But you don't need a direct representation of the spin, for instance, of a particle that, that all things that will quote, unquote, measure it don't see, don't interact with or whatever, because nobody's effectively looking at it. It does not need to be rendered right.
- [18:47]
Donald Hoffman
So good.
- [18:49]
Tom Bilyeu
So that'll feel right just to.
- [18:51]
Donald Hoffman
No, that's a great question. And so great question.
- [18:53]
Tom Bilyeu
I was not asking a question. I was stating a hypothesis about crazy
- [18:57]
Donald Hoffman
or no, I think.
- [18:58]
Tom Bilyeu
Does that make sense to the macro, to the micro level?
- [19:01]
Donald Hoffman
Well, it, it does. But I think a good analogy here that might help clarify the issue is, is so in say, Grand Theft Auto, right? I look over, I'm playing with somebody who's, you know, in Canada and somebody else is in Europe and someone else is in China. We're all playing a remote version of it in virtual reality. And I look over and I see a red Porsche to my right. And so I say, is there a red Porsche? Am I right? And the guy in China says, oh, yeah, I see a red Porsche. And the guy in Canada agrees, and the guy in Europe agrees as well. So, of course, each of them is rendering their own red Porsche. So there is some reality that's coordinating all of these perceptions. Right. So the guy in Canada didn't see a red Porsche until he looked, but when he looked, there was this whole world, you know, of circuits and software that you don't see. There's some supercomputer that's coordinating the whole thing. How's it coordinating in that particular metaphor? Right, the. There's a supercomputer that's. That's taking the inputs from, like, your headset. What, what direction are you looking with your headset? Maybe you've got a bodysuit, so it's looking at your arm movements and so forth, and it's feeding all that into a supercomputer where it's got a model of the game. And in that model there's some red portion model. Of course, there's no red portion in the computer. And it knows then how to coordinate and send the photons to your headset in Canada and my headset and Irvine and someone else's headset in China. So that we have this notion of a persistent reality of a Porsche, even though individually for each one of us, local realism is false. The Porsche doesn't even exist until I render it, and there's no red Porsche inside the supercomputer. So that's sort of the idea is that space time is just a headset and there's behind spacetime, there's going to be an incredibly complicated realm to explore that's as least as complicated, more complicated as, like, the supercomputer is to my little headset. Headset is sophisticated, it's beautiful technology, but the supercomputer is really, really powerful thing. The same thing will be true of space time. It's just our headset. But if we look beyond that headset, we're going to be finding a realm that's far more complicated. So in some sense, science up till now has only studied our headset. We've studied inside space and time. We're taking our first baby steps to start to explore. We've cut our teeth in science on studying our headset. We learn the tools in the last three or four hundred years about experiments and clean mathematical Theories and the loop between experiments and theories. But we thought we were studying objective reality. We were studying our headset. But now we have the tools to actually take a first step beyond space time and start to find structures beyond space time and their projection back into space time. And so from that point of view, our view that objects in space time, we've taken that to be the fundamental reality, will look sort of parochial. Hopefully in just a few decades. I think the next generation, where many people will have spent a lot of time in virtual reality. My generation didn't spend a lot of time in virtual reality. So this is hard concept. But if you've spent.
- [22:13]
Tom Bilyeu
I've heard you say that before. I don't think that's going to get people where you think it's going to get them. Maybe not, but in this episode I want to try to explain why I think that and. And get your take. So here's what I think we need to do first, and then we'll go even deeper. There's two things we need to do in the near term. One, I think we. We need to. In. In our previous interviews, we spent a lot of time dealing with the headset. So for anybody that's sort of confused on that idea of you're living in a simulation, everything that you know and love and touch and have ever experienced, it is all a simulation. You have never existed outside of the headset. So if right there your brain breaks, go watch the other episodes. We spend a tremendous amount of time building that up. But for now, what I want to do is say, okay, I'm going to assume that you get it, that your whole life is basically Grand Theft Auto.
- [23:08]
Donald Hoffman
Okay?
- [23:08]
Tom Bilyeu
And people understand it. You've been in there playing the game, and they understand the difference between playing the game and the computer rules and things that give birth to that game. And so that's. That's the difference. What I want to do now is map that one layer back. So I want to take that idea of your life is Grand Theft Auto, but there's this thing called space time that's outside of it, and get to what you're actually saying, which is that same relationship, but move back one very profound level. Because what it does is it inverts everything. And what it says is that the universe, the universe, space, time is an emergent phenomenon from consciousness. That consciousness is in this, to use that analogy, just to map it back. That consciousness is the quote, unquote, computer and rules of the system. And then the simulation is what we all think of as real Life. Okay. So that's where we're mapping. So one does. Does that track for you, that we can move that analogy sort of one rung deeper is probably the word you'd be most comfortable with.
- [24:19]
Donald Hoffman
Right? So absolutely. A model in which we take consciousness as fundamental and we have a mathematical model of consciousness, and we then try to show how space time gets rendered from that.
- [24:28]
Tom Bilyeu
Okay, perfect. So now in this interview, instead of making our references to Grand Theft Auto, unless we need to, for whatever, for an anchor point, I want to talk about space time like a simulation.
- [24:38]
Donald Hoffman
Okay.
- [24:38]
Tom Bilyeu
I want to talk about space time like it is Grand Theft Auto. Because researching you this time, I. I want to sit with it for a while before I start saying I'm 100% behind it. And I mentioned one of our previous interviews that I do revert to the mean after I spend time with you, but each time you're. You're shifting me farther where my mean is sort of closer to you. This time, at least in the research, I had a real sense of. He's right. I don't know about the consciousness is the only part that we may disagree, but that you really gave me an internally consistent set of logic points for why space time is the simulation. And when I grant you a few base assumptions that will go through my own worldview makes more sense. Okay. And so I realized for the first time again, fully acknowledging that I may revert to the mean once I've interviewed three or four other people on totally different topics, and this is sort of cleared in my system. But right now, as we do this, I really felt like you improved what I consider a prediction engine. I think of the human mind as a prediction engine. And the closer you get to baseline truth, the more you're able to predict the outcome of your behaviors. What I'm watching happening with AI, which is why I wanted to start there, I can't make sense. I don't. When I think about a hallucinating AI, I'm like, I don't understand. When I think about AI pulling patterns out of noise, I don't understand. When persistence is difficult for AI, I don't understand. And then I research you, and click, click, click, those pieces fall into place when I assume that it's all already a simulation and that AI is simply revealing to me how the simulation works. And so. But the fact that we disagree, or maybe we don't. I think AI will be windows into consciousness. I think AI is leveraging your own theories to create AI Right now, as we're talking about It, I think, I'm a layperson. Everybody needs to take this with a huge grain of salt. Trust me, I am well aware of my limitations. But I think right now that what we're witnessing with things like stable diffusion, where AI is creating an image out of the infinite possibilities that exist within this, the possibility space of noise. Okay, for people that don't understand how stable diffusion works, that's how it works, is it dips into the noise to find a pattern and then solidifies that pattern to reveal, is this what you wanted? And what I'm saying is, when I research you, I realize, oh my God, that's precisely what your theory predicts in the idea of Godel's incompleteness theorem, which I have struggled with so hard in the previous interviews. I feel bad for everybody that has to watch me go through that. But the more I feel like I can grasp why you keep coming back to it and why this sort of infinite possibility space is so important to understand. When I watch AI pull a static image out of infinite possibility, I'm like, oh my God, that's exactly what you've been trying to describe. Okay, put a pin in that. Because what I want to talk about now is consciousness as fundamental. Because this is the part, if people are really paying attention, this is the part that will change your worldview to, to get into the, the space time as a, as a construct, as a simulation. You first have to understand that you think that's born of the. As born of consciousness itself. And I. Please, dear audience, stick with this because this point is going to be very important as we piece together the predictions that your own model is going to make. But they have to understand this first. So how is it possible that consciousness, the thing that I think everybody intuits, comes from stacking neurons, neurons, neurons, neurons, neurons. And you pass through a cricket, an ant, a mouse, a cat, a dog, a dolphin, a gorilla, and humans, it just feels like, oh, just stack more neurons and then you're ultimately going to get these more sophisticated neurons which give you a more sophisticated consciousness. That seems so self evident. And you're. To me, but you're saying, nope, no,
- [28:56]
Donald Hoffman
and, and by the way, I'll just. On the pin, I'll just mention that I agree with you that AIs could actually give us a window into consciousness, but they won't create consciousness. That was all. That was all I was saying.
- [29:08]
Tom Bilyeu
Interesting. So I think we disagree about that.
- [29:10]
Donald Hoffman
Okay, so you're. So we can go into more thoughtful.
- [29:13]
Tom Bilyeu
So much farther ahead. When we get there, I will Lay out my absolute ignorant perspective.
- [29:18]
Donald Hoffman
So on consciousness being fundamental meaning. That's all there is, that, that's right. So the idea would be, and this is by the way, in some sense not new. Leibniz in his Monadology had the same idea.
- [29:37]
Tom Bilyeu
So I really appreciate that you assume I know what that means and from context I can tease it out. But can you tell us what that means?
- [29:44]
Donald Hoffman
Oh, so, so Leibniz was this genius contemporary of Newton, sort of a, antagonistic. They, they both invented calculus roughly the same time. There was a question about who was first and so forth. And they, they, they, they were, you know, sort of at each other. But, but they were, they were contemporaries. But Leibniz had this idea that, that consciousness couldn't emerge from physical systems. He has a famous argument of the mill where he, he in one paragraph basically dismisses the idea that objects inside space and time, like neurons for example, could create consciousness. For him, it was so obvious that he spent a paragraph on it and moved on. And, and then he's got a, a book called the Monadology where he was proposing essentially that consciousness perceiving entities are the fundamental reality and that they were interacting.
- [30:38]
Tom Bilyeu
All right, if I break down the
- [30:40]
Donald Hoffman
words mono, dology, monad, so M O N A D is a technical term for him. It was he, it was a new term for him. Monadology is then the book's name, monadology. And it was basically, it was a dynamics, it was a strange dynamics we called a pre established harmony where God, so he had, he brought God in on, on his thing I believe, to, to sort of coordinate all the, the perceptions of these.
- [31:10]
Tom Bilyeu
So meaning God was the first mover, the fundamental.
- [31:13]
Donald Hoffman
Yeah, the fundamental.
- [31:14]
Tom Bilyeu
Right, okay, but he saw it as a creator touching things with like a divine spark of consciousness.
- [31:22]
Donald Hoffman
Yeah, but his ontology was that, that, that the fundamental reality beyond space time was these monads, these perceiving entities basically. And, but, but God I think was, that was the deepest reality for, for Leibniz there. I'm less secure the monologue. I'm not sure exactly what his thoughts were on God, but I believe that's what he said. So I just brought that up just to say that, you know, we're not the first to have this kind of idea. Centuries ago, Leibniz with his mona knowledge had an idea that perceiving entities, experiencing entities could be more fundamental than the physical space, time, world.
- [32:02]
Tom Bilyeu
All right, you talk about conscious agents. Do you mean exactly that same thing?
- [32:07]
Donald Hoffman
That's right. So conscious agents are a mathematically Precise statement of what we mean by consciousness. Right. So as a scientist, it's not enough for me just to say, okay, there's consciousness beyond space, time and is fundamental. I have to write down a mathematical description of what I mean by that. So what aspect of consciousness do I take to be fundamental and what's the mathematical description? So if you think about it, think about consciousness, there's of course, experiences, there's learning, memory, problem solving, intelligence, maybe free will. There's lots of things, the notion of a self, all these things that you might think a theory of consciousness needs to, to, to incorporate.
- [32:52]
Tom Bilyeu
I'm so sorry, and I should have done this before. And that apology goes to the audience. If you're new to Donald, it's probably worth just a quick sentence about what consciousness is.
- [33:02]
Donald Hoffman
Oh, well. So I would say consciousness is the ability to have experiences like the taste of chocolate, a headache, emotions.
- [33:15]
Tom Bilyeu
So this thing feels like something.
- [33:18]
Donald Hoffman
Yeah. The way a lot of philosophers will talk about it's is to have conscious experience. There's something it's like to be a conscious entity. It's something. There's something it's like to have a headache. There's something it's like to have your, you know, to eat, to have a nice cup of coffee or something like that.
- [33:33]
Tom Bilyeu
Okay, and so let's call that qualia again, me stealing directly from you. But just so we have words, because qualia is going to become very important as we get into your paper and all of that. Okay, so back to conscious agents.
- [33:46]
Donald Hoffman
So what we decided to do was we don't want to throw the kitchen sink in our mathematical definition. So we took what we felt was the bare minimum starting point. There are experiences like the taste of chocolate, smell of garlic, and so forth. And those experiences affect the probabilities of other experiences occurring. So there are experiences and probabilistic relationships among experiences. That's it. So we're not bringing in the notion of a self learning, memory, problem solving, intelligence, none of that. What we're saying is, yeah, all that stuff is important, but we have to prove how it arises from just experiences and probabilistic relationships among experiences. So that's. As a scientist, you try. It's what we call Occam's Razor. You want to have the minimum number of assumptions at the start of your theory. Every theory has assumptions. There are the miracles of the theory. We want as few miracles as possible. Right. So our only miracles are there. Well, it's a big miracle. There are experiences and probabilistic relationships among experiences. And we formalize that the experiences we just write down what's called probability spaces. We can, if you want, we can talk about probability spaces. And the relationships among experiences are what we call Markovian kernels. And we get what's called Markov chain, so very simple dynamics.
- [35:12]
Tom Bilyeu
So we'll, we'll explain what Markovian dynamics are in a second. I don't. Now that I finally have at least a tiny bit of a grasp, I don't know how important it is that people understand that.
- [35:21]
Donald Hoffman
Sure.
- [35:21]
Tom Bilyeu
But I do want to know how important is it that one bit of qualia impacts other qualia? Like, does that, does that relationship play heavily into the idea of consciousness as a fundamental agent?
- [35:37]
Donald Hoffman
Yes, we, we stipulate that as a fundamental property that, that experiences aren't in a vacuum. Experiences probabilistically lead to other experiences.
- [35:47]
Tom Bilyeu
Okay. It's very interesting that you said not in a vacuum because that my whole thesis is that the construct of space time. The simulation. Let's just be very clear. The simulation that is this real world. Sorry, that's a terrible use of word real. The simulation that everybody lives in and experiences is required. This is, this is my pitch. The simulation is a required constraint in order to give context that something can be like anything but that. For consciousness to explore the possibility space of qualia, you have to have a rule set and the rule set that we're all in, which may be one of a gazillion headsets. But the rule set that we're all in creates the possibility for the subset of qualia that we as human beings or lizards or whatever experience. But without that rule set, that is space time.
- [36:54]
Donald Hoffman
Right.
- [36:54]
Tom Bilyeu
We would not have enough limitations to give us the context in order to feel a certain way.
- [37:01]
Donald Hoffman
Exactly. That's. That's a very good way to put it. So that a lizard presumably sees things very, very differently than I do. Pigeons have four color receptors. We only have three.
- [37:14]
Tom Bilyeu
Pigeons have four.
- [37:15]
Donald Hoffman
Yeah, that's right. So they see more color than we do.
- [37:17]
Tom Bilyeu
Birds. And I feel cheated now. I knew 15 of women do. I did not know pigeons.
- [37:23]
Donald Hoffman
Yeah. The mantis shrimp has more than 10 photoreceptors. Yes, that's right. Different kinds of or pigments that are used for the photoreception process. So we may be cheated in many, many ways.
- [37:38]
Tom Bilyeu
That's for sure.
- [37:41]
Donald Hoffman
And we don't, for example, perceive polarization of light. And birds and baby bees do as they can perceive the polarization of light, we can't directly experience electric fields. And there are, there are animals in the Water that can do that. So some to see infrared, some that see ultraviolet that we can't. So, so we're, we have a very, very small window and other animals are not restricted to the windows in which we, we see. So I like your idea that there, there's an infinite space of conscious experiences to explore. And when we look at different animals, we're seeing different explorations with different headsets and, and different, as you say, different constraints. And it's, it's in some sense consciousness exploring all of its possibilities, all the possible ways that to explore. So in some sense we're here for the ride and we should enjoy the ride. We're, we're, you know, we're exploring. We thought this was the final reality. No, this is just one of countless possible headsets, just one of countless. And we'll enjoy this ride. And then consciousness will, then it's looking through other headsets. So I like your idea that it's, you know, there's some kind of consistency, some kind of coherence, but it's a subset of the experiences. There's an infinite number of experiences to explore. So this ride never ends.
- [39:08]
Tom Bilyeu
Okay. So when I think about consciousness as fundamental, I cannot help but imagine a blob that then takes shape in the form of a human or a lizard or an avocado, whatever. Help me understand what. Do you have an image in your head of what the, what consciousness is? Is it just completely non physical?
- [39:35]
Donald Hoffman
Well, maybe the closest I can get that would be the way that we communicate to people would be if you go into an entirely quiet room, shut off all the lights, close your eyes and get very, very still and don't think.
- [39:50]
Tom Bilyeu
Good luck.
- [39:52]
Donald Hoffman
That's right. Usually letting go of thought is not easy, but, but if you can go for a few seconds or a minute with absolutely no thought, and now you're just aware, you realize, yeah, I can be aware without being aware of anything in particular. I am fundamentally awareness. And into that awareness right now are coming a cup, a microphone, a table. I can close my eyes and those are gone from awareness. So somehow there is this field of awareness that is in some sense deeply and fundamentally who you really are.
- [40:36]
Tom Bilyeu
That so that it seems like your theory would say that's false.
- [40:42]
Donald Hoffman
Well, it's going to say that the conscience. So the reason why I talk about this awareness is that when we talk about all these specific conscious experiences, we have to write down something that's called a probability space first. We're required mathematically to do this. We write down a probability space in
- [40:59]
Tom Bilyeu
which probability of qualia.
- [41:02]
Donald Hoffman
That's right, probability of qualia. So you have to write down the space of all the potential qualia that this particular conscious agent could experience. So here is this space and there's the mathematical structure. It's just sitting there prior to any particular experience happening. It's just sitting there. And it took me a few years to ask myself the question, what is that space? I had to write it down. I couldn't do the math. I couldn't write down my Markovian dynamics until I wrote down the probability spaces. But as you know, the way we do it is we just of course have to write that down. So you don't even think about, you write down the probability space and you go on to the fun stuff. You write down now the dynamics and so forth. But a few years later I came back and go, well, wait a minute, I went too quickly on this first part. I had to write down a probability space. What does that mean? Because this is a space prior to any specific conscious experiences happening. And so the best I can say right now is that perhaps is the mathematical counterpart to what I was just describing, which is the awareness that you can experience prior to having any particular specific conscious experience arise in that awareness. So that's, that's why I talk about it in that way.
- [42:20]
Tom Bilyeu
Can I just restate that to make sure that I understand and linger on a, for a second for the audience? So you're using words that I know you know are dangerous, that Annika Harris has warned you about letting people carry the sense of self into all this because you said you are the awareness. But really consciousness is the awareness that animates me in some way or it needs my constraints in order for it to experience the qualia. I think that's the right way to think about it. And so in those moments where either through meditation I get to true, where I am simply aware of the qualia of being aware, but when it's not aware of anything in particular. So I'm not aware that my foot hurts. I'm not aware that my, my stomach is churning on food. I'm not aware of something I need to do later in that day. I am just the, the potential to point that awareness at something is the thing that I'm sitting in that that's who we really are. So that, that feels right, but I know it. It's re trapping me in my sense of self that I am a real thing. Your whole thing clicks into place for me when I realize that according to your theory and this makes a lot of things make sense. In my own life, I am simply one instantiation that creates a set of what I call biological limitations that then once I have those constraints, now the fundamental element of consciousness can begin to explore its qualia. The, the different things that, like, oh, in this human form, I can experience these things with all the context that this person has. He responds to this thing in this way.
- [44:19]
Donald Hoffman
Right, agreed.
- [44:21]
Tom Bilyeu
There are some deep complexities with that, but we'll push those off for later. Okay, so if, if that's where we're at, my fundamental question is why does consciousness, why is it compelled to explore these qualia states?
- [44:36]
Donald Hoffman
That's the $64,000 question. That's so I don't know, but I can. Of course, that's the very natural question to, to ask. And I agree with what you just said. I mean, I don't want to reify the self. What we are are avatars of the one, effectively. And the one consciousness is the one. Awareness is exploring all of its possibilities through different avatars. Why there? I, you know, I think there may be some deep mathematical reasons. So it may be that, I mean, there's, there are theorems to the effect that no system can completely know itself. It's impossible. So because, for example, if I, if I have a computer and I want the computer to explore itself, how is it going to know itself? Well, it's going to have to build a model of itself and write down what. Well, in the very process of building a model of itself and writing into its memory things about itself, it's becoming more complicated. It's changing itself. So now to really understand itself, it's going to have to now describe what it just did. And now to. So you get this infinite loop. And so there are, there are problems with self understanding. It's not possible, in many cases, provably not possible to have a complete understanding of yourself. You get into this infinite loop of, now I have to get more complicated to understand myself after I just understood myself. Right. And so that's one direction of this. Another direction is there's a whole hierarchy of infinities. So the integers like so 1, 2, 3, up to infinity, that's an infinite number of integers. We call that accountable infinity, or aleph 0, the Hebrew letter aleph and 0, just meaning the smallest infinity. But there are other infinities. So the next, if you take the set of all subsets of integers, so like 1, 2 and 15 and 2, 3, 4, look at all the possible subsets of integers and ask how many subsets are there? How many subsets of integers can you come up with? It turns out that of course there's an infinite number of these subsets because
- [46:54]
Tom Bilyeu
every number is divisible by an infinite number of subsets.
- [46:57]
Donald Hoffman
No, we're just grouping them together. So I'm saying think about the group one and two. So that's a group. Now one and five.
- [47:03]
Tom Bilyeu
Got it. So we can group an infinite number that's an infinite number of times.
- [47:07]
Donald Hoffman
So those are called the. All the different possible subsets of the integers.
- [47:10]
Tom Bilyeu
Got it.
- [47:11]
Donald Hoffman
And there's of course an infinite number of them because one is a group, two is a group three. So we already know there's an infinite number. But there's more than that. How much more? It turns out it's a bigger infinity. So the.
- [47:20]
Tom Bilyeu
It's a bigger infinity.
- [47:22]
Donald Hoffman
It's a bigger. Well, that's what mathematicians said when Cantor, the mathematician who first came up with this when he first proved this feels
- [47:31]
Tom Bilyeu
a bit like my speaker goes to 11. Why not just take. Make 10 louder?
- [47:37]
Donald Hoffman
But this one goes, this one goes. It's actually a different size of infinity.
- [47:41]
Tom Bilyeu
And so I literally can't wrap my head around that.
- [47:44]
Donald Hoffman
There is something called Cantor's diagonal argument. So there's a simple diagonal argument where you can actually show on paper, pen and paper that it's impossible to capture all the power set, this bigger infinity, with the smaller infinity. So he gives what's called Cantor's diagonal. So if people want to, you know, check me on this, you just look up Cantor and Cantor's diagonal argument for a proof that there are these bigger infinities, and you can actually, I think most people can actually follow the proof. I mean, it's mind bending, but you can follow it well. There's not just one bigger infinity, that's Aleph one is the bigger infinity. Now take the power set. So by the way, taking the set of all subsets is called taking the power set. So the power set is all the possible subsets. So Now I've got Aleph1, which is the bigger infinity, which is the powers, all the power sets of Aleph zero. But now I can take all the power set, the power set of Aleph 1 that gives me Aleph 2. Take the power set again at Aleph 3, Alep 4, and this goes forever. So infinity is not one thing. There's an infinite unending hierarchy of ever larger infinities. So we have, to my view, take this into account in our theory of consciousness that all of these different infinities are valid directions for projection of this one deeper consciousness. And so we're going to. So the answer to your question may again be because Cantor's hierarchy never ends, this exploration never ends. The exploration of the possibilities of consciousness, of qualia, is in principle, never ending.
- [49:31]
Tom Bilyeu
So I get the never ending.
- [49:32]
Donald Hoffman
But again, I would just say I'm in deep water here and I'm maybe over my head.
- [49:39]
Tom Bilyeu
Fair enough. This is the fun part of exploring this is what are the predictions that are made based on the hypothesis? Right? So every hypothesis makes a prediction, and then you have something you can test and it becomes verifiable. So this is where it gets very interesting to me, is what the predictions are that it makes. So going back to the hypothesis that I have that, okay, maybe this really is all a simulation, because as we go to build the next simulation, simulation, it actually tells us more.
- [50:13]
Donald Hoffman
It.
- [50:13]
Tom Bilyeu
It gives me a better way to understand what's already happening. Now, again, I'm a layperson, so I may be way out of my own depth here, but I think people will be able to follow the internal logic. So this is what I was stating earlier about AI. So the way that AI works is there is an infinite possibility space in noise. So you can just think of it as a screen. And that screen can have think of every conceivable pixel that's there. And depending on what color you make any one of those pix, if you have like a grand enough resolution, meaning enough pixels in a finite space that you can recreate any image that's ever been seen or created or even just what's possible. So if anybody's seen what they call an AI hallucination, where the AI will just continually, like, push into itself. And every time it pushes in and a pattern begins to emerge, it then crystallizes that pattern and basically says the most likely shape to emerge out of this would be a staircase. But as you push in, the most likely shape to emerge out of that would be a cathedral. And. And it just keeps going and going and going and going, and it never runs out of sort of most likely things to emerge out of this pattern is because it's looked at all of these things, and so it will create things that it's seen before. So the Mona Lisa would be one representation that is very predictable, especially given how many times the Mona Lisa has been replicated. So one of the things in the possibility space is the Mona Lisa is a Rembrandt. Is David, is you looking at your wife this morning? Is one of the possibility spaces that it could eventually draw out of this thing. So it's, it's constantly searching for what is the next potential pattern. Now my whole thing is what really starts to make this interesting. And the reason that I think that the simulation isn't something to be brushed aside as being trivial, but is critically important. And if you're right, that what consciousness is doing is it has some motivation for some reason that neither of us know why, but that it is cycling through all of its permutations, if that's what's really happening, then to do that you need a set of rules. And so what I realize is I'm building the. Going back to the Grand Theft Auto, so we're building a simulated world. And I realize as we build it, all I'm doing is making the most detailed if this, then that statements. And so I'm trying to create these algorithms that then not trick you, but they give you a set of rules by which you now must adhere. But by doing that, by actually limiting the possibility space, I can make a game that's quote unquote fun. So it is in the limitation, it's in the setting of rules that this becomes a useful space. So what I want to know is you, you talk a lot about like, hey, we want to get out of the headset. Do you really do you want to get out of the headset or do you want to manipulate the headset?
- [53:23]
Donald Hoffman
Well, when I say we want to get out of the headset, that's as a scientist trying to look for a deeper theory. So as a scientist, I mean we've sciences.
- [53:31]
Tom Bilyeu
But let me ask you. So the reason that Einstein, his breakthroughs were so useful is within the headset, they let us do something. Are you trying to do something in the headset or. So if you understand how the headset works, you can either manipulate the like Einstein, bend space time, right? You can create gps, which if you didn't understand relativity, you would not be able to do. And that made the atom bomb possible. It made nuclear energy possible, made GPS possible. His breakthroughs. Are you trying to do a breakthrough that has. Has headset implications, or are you searching a breakthrough that has get out of the headset implications?
- [54:10]
Donald Hoffman
Both. So what I want to do is get a theory of what's beyond, at least a baby step beyond the headset. Presumably, as I mentioned, there's a Cantor's hierarchy of infinity, so we have infinite job security going beyond the headset. It's literally an unending job. But to take a step entirely outside of the headset Then as, as you point out, as a scientist, I need to make predictions back in the headset because that's the only place we can do experiments to prove that you're to get. Right. Well, to, to. I don't. You can never prove that you're right. But, but to, to sort of what we say science will say to, to get confirmation of your theories, which is not proof. But to, to say you're not stupid, you seem to be on the good tracks.
- [54:57]
Tom Bilyeu
The things that we already understand.
- [54:59]
Donald Hoffman
That's right.
- [54:59]
Tom Bilyeu
Hopefully makes no things that we don't currently understand.
- [55:04]
Donald Hoffman
That's right. We should be able to get quantum field theory back as a special case. We get Einstein's theory of general relativity as a special case. Evolution by natural selection as a special case or generalizations of these theories within space time. So yes, we're going for the first baby step outside of space time in terms of a scientific theory, but of course we have to project it back into space time where we can do experiments and a better look like evolution by natural selection and quantum field theory or understandable generalizations of those theories. Or we're wrong. Right. So, so you might say, well, yeah, if you go outside of space time, you can do anything. You have all the fun you want. You can do anything you want to. No, you can't. You can. You need to tie it back to what we can perceive inside our headset so that that's where we're headed. But as I said, there's infinite job security. And so I view myself as just looking for a first baby step outside of the headset. Science for centuries has only studied our headset because space time is our headset. But in the last 10 years, physics has gone beyond. We've talked before about the amplitude and decorated permutations and other structures that physicists are finding. These are not the final word. Again, these are the first baby steps outside of our headset. And they will be, of course, refined and eventually superseded.
- [56:25]
Tom Bilyeu
All right, so there's one of these things that I think I've, I've grasped enough that I can present it to people as one of the first baby steps. So in physics, one of the things they're constantly doing is smashing particles together to try to see what happens when those particles collide in the hopes that it will reveal smaller and smaller elements of the building blocks of the universe, which will then help us understand what the, the sort of fundamental makeup of space time is. And as they look at this data, what they found is that there are patterns in that Data that replicate endlessly. And you smash these together and the, the collisions. There's so much data. At first it seems impossible, just so much data to wade through. We'll never understand anything. And then all of a sudden you realize, wait, there's only so many patterns. Once you take those, like, once you group those shatters. Like, if you think of it this way, if every time you broke a mirror, it broke into the same pattern, you'd be like, wait a second. And am I understanding it correctly that that's what happens when you collide particles statistically?
- [57:37]
Donald Hoffman
Yes. Right. So it's not exactly. But, but you, you can use statistics to show that there are the statistical commonalities to the interactions? Absolutely.
- [57:46]
Tom Bilyeu
Okay, walk us through that. And why does that matter?
- [57:49]
Donald Hoffman
Well, for physicists, of course, this is some of their most fundamental data. So there, what are particles? Particles, Eugene Vigner taught us, are what he called, you know, irreducible representations, unitary representations of the group of symmetries of space time that what they call the Poincare group. Essentially, particles are like the, the simplest things allowed by the symmetries of spacetime, the simplest entities allowed. And so in some sense, by studying these particles, we're really studying the nature of space time itself and the structure of space time. And so when they, for example, in the Large Hadron Collider, they will smash protons together or they will, they're. They'll also, you know, sometimes have an electron and smash it into a proton. And at high energies. And when you do that at high enough energies, you destroy the proton, it actually falls apart. And you see all these particles scattering up things like quarks and gluons and mesons and so forth. And so you can look at the angles that these particles are spraying out at. And look at, at, for example, do they have, you know, a spin, a magnetic charge? What's their. Do they have a mass? So you can sort of, you can look at all the. And then when, when you start looking at all the data, you begin to see patterns in the data. And, and so we see, you know, for example, it was a big surprise to physicists that inside the proton there were these things that they now call quarks. But the quarks, in some sense, at least at the energies that, that are available to us, can't be on their own. You can't have, like, quarks flying out on their own. There's something called quark confinement. And that was a big, big discovery. So quarks, like in a proton, there are three Quarks two up and one down. A neutron has two down and one up. And but if, if you, if the quark escapes, if it's trying to get away, the, the force of attraction between two quarks grows with the distance and the energy. Well, the force doesn't grow the energy. So the force doesn't. Normally we think of the force, the force. So the force doesn't grow, the force remains constant. And so the energy, the potential energy keeps growing and growing as you move these particles apart. And so at some point they snap and you, you create all that energy goes and creates a new quark, say so, so then they pair off. So it's very, very strange this court confinement thing. So one reason we do experiments is because, I mean who ordered that? We, we wouldn't have like guessed, you know, quart confinement. And so but we, we found quart confinement and is still being studied. I mean, trying to understand that there's a theory that if we get at really, really high energies, they won't be confined. But, but those are energies that we currently are nowhere near. And we have no analytic proof right now of quart confinement for what are called non immediately engaged theory. So, so one of the big open questions in physics is to actually prove this analytically that, that, so they, they have lattice gauge models that, that of this that show it. And, and, and, and they have other cases where the experiments and the theory convince them it's the truth. But we don't actually have the final analytic proof of this in what's called non abelian gauge theories. So that's still an interesting open question. But that's why physicists are doing this. These particles are really probing in some sense the fundamental nature of space time itself. And so they look at, at patterns, they look at the cross sections for interactions. So this was for example, way back in the early studying of the atom. So there was a plum pudding model of the atom, right? So there was electrons were these negative point particles inside a positive field. And then this one experimenter started shooting particles at atoms. And the plum model would say that most of these particles would just go, go straight through. And most of them did, but every once in a while one would bounce back very, very small percentage of the time. And so that, that gave them the idea, okay, there are point like particles, we now call them protons and, and neutrons. These particles that were, they were hitting that, but they were a very, very small space within the, the atom. So the atom was mostly empty space. The electrons were way far away. So to speak from the, the much smaller protons and neutrons. And so, but then we look inside the protons, we find that the proton itself and the neutrons are composed of even smaller particles, quarks and gluons and so forth. And who knows, even the quarks and gluons might be, you know, composed of smaller particles. But we don't have the resolution in our colliders right now to test that. We can only go to, you know, a thousandth or ten thousandth the diameter of a proton, I think. And at that resolution, the quarks and gluons still look like point like particles.
- [63:23]
Tom Bilyeu
It doesn't seem self evident to me that just because again, I'm granting you the conceit that consciousness is the fundamental thing, but it does not seem self evident to me that even if consciousness is the fundamental thing that gives rise to this constricting rule set as I describe it, that we call space time, that you couldn't have a theory of everything regarding space time. Why do you think we have failed to get a theory of everything in spacetime in space time, knowing that it's the simulation? But going back to Grand Theft Auto feels like even if I just said, oh, all I can tell you is cause and effect, that when this pixel goes here, it has this effect. And so now I can play everything's forwards or backwards. And you could. In Grand Theft Auto, it has a set of rules and it adheres to those rules, period. Plain and simple. And so even though it is the, the, it is the headset, a computer program, assuming that a simulation acts like a computer program, space time, in this case, it adheres to rules. And so when you get a quote unquote bug, it is what the program is programmed to do. You just didn't intend to program it that way.
- [64:39]
Donald Hoffman
Well, in that framework, yes, I agree with you that I think we could get a complete theory of space time. Not a complete theory of everything, but a complete theory of space time. So the theory of everything for me would be, you know, space time is a trivial aspect of everything.
- [64:53]
Tom Bilyeu
Right?
- [64:54]
Donald Hoffman
So, but, but absolutely, I think we can get a complete theory of space time and we'll see its limits. It, it falls apart at 10 to the minus 33 centimeters and 10 to the minus 43 seconds. So we'll, we'll see that, we'll understand that. Yeah. So it's, it's quite, quite possible. I would say though, and I like your idea about the, the program and the rules and setting up a framework in which you can Explore experiences. I'll throw in a little wrinkle. You're writing computer programs. And so Alan Turing, you know, is sort of one of the fathers of modern computer science. And, and Turing machine is like the first, like really good theoretical framework for computer science. And the universal Turing machine that, that Turing described in some of his papers is sort of our notion of a universal computer. But there's a well known limit to what Turing machines can do. Take again all the integers, you know, 1, 2, 3, up to infinity, also minus 1, minus 2 and so forth. And ask, think about all the functions from the integers to the integers. For example, the square function. So you know, the square of 2 is 4, the square of 4 is 16, and so forth. How many functions are there? It turns out it's a bigger infinity. It's not accountable, it's a bigger infinity than the integers. But Turing proved that the set of computable functions is countable. So when you're programming, you're using only computable functions, but they're, they're a much smaller infinity than all the possible functions. So right now in our current technology where when we build these computer simulations, we should know that we're using a probability 0 subset of all the functions that are actually available. And maybe later on we'll figure out how to do something more interesting with all these other functions. But then as we go again, Cantor's hierarchy, I think that in other words, the, the kinds of rules they're going to, are going to be very, very hard for our heads to understand. You can write down if you take a class in theoretical computer science, you can study non computable functions so that you and almost every function is non computable. Okay, As I just said, the computable functions are probability zero. The set of all functions is, is all. Most, most of those functions are not computable. But in a theoretical computer science class, you will, you will actually spend some time actually studying how to construct and prove that a certain function is not computable. Like the halting problem is not a computable, it's not a computable function. It doesn't. But it's really hard for us. Even though almost every function is not computable, almost every function we can think of is computable. So here we are stuck with the limitations of our headset. And so thinking out of the box in the simulation idea is really going to be mind numbing because to really think out of the box, you're going to have to learn how to think about non computable functions. And that is not trivial. That's not, but that's. So I just wanted to throw that out there to just open up how complicated this can be and why the exploration could be.
- [68:46]
Tom Bilyeu
To get a theory of even just the everything of space time, we have to get into non computable functions.
- [68:52]
Donald Hoffman
I don't know if we will or not, that's an open question. But we should be open to that possibility.
- [68:57]
Tom Bilyeu
Very interesting.
- [68:58]
Donald Hoffman
And certainly to explore consciousness. I see no reason why we should. A priori, I would say this. If someone claimed that the computable functions were all we need, I would say the burden of proof is on you
- [69:13]
Tom Bilyeu
talk about something I have not even considered. I don't know that I can wrap my head around that one yet.
- [69:17]
Donald Hoffman
I have a hard time. I mean I took a class and I, and I looked at that non computable function, the halting problem. And you have to really, I mean you have to be sober, you have to be well rested and you have to think really hard. At least with my apparatus you have to think really, really hard to even grasp it. It's not trivial, intense.
- [69:34]
Tom Bilyeu
Okay. So when we have a hypothesis that makes predictions, we need to be able to solve. We were talking about this a few minutes ago. We need to be able to solve problems or our hypothesis needs to predict outcomes of things that we can observe but not yet explain in. I can't remember if you mentioned this in your paper, but I have heard you talk about this. So dark matter, dark energy, we don't know what the hell it is, but we know that the universe would not hold together if it wasn't for that. Or it wouldn't be racing apart at the way that it's racing.
- [70:07]
Donald Hoffman
Whatever.
- [70:08]
Tom Bilyeu
It wouldn't function the way that it functions. Now. What does your consciousness as fundamental agent tell us about dark energy?
- [70:18]
Donald Hoffman
Well, nothing specifically. Right, so that's, that's a big open question. In fact one, one of the, my collaborators is a, is a student working right now on dark energy experiments. A brilliant student named Ben Knepper.
- [70:32]
Tom Bilyeu
Because he thinks it will yield results tied to consciousness is fundamental.
- [70:36]
Donald Hoffman
No, I think it's just because it's a good thing to do at this stage in your career to get that kind of experience and you know, actually spend time hunting with real experiments for dark matter. So you learn the ropes. I think it's. It was. And so he's doing that and who knows, you know, our, our current techniques may or may not find dark matter. We, we just don't know but it's no surprise from a point of view that says that space time is not fundamental to say that there could be influences on our headset that are not explicitly represented by the headset itself. They're only seen as influences on the headset. But. And so in one way that we're going after this in our own mathematics is we have this Markovian dynamics of these conscious agents.
- [71:30]
Tom Bilyeu
Can you take a second to explain to people what Markovian dynamics are?
- [71:33]
Donald Hoffman
Yeah, Markovian dynamics is fairly simple in concept. It says that, that what you do next. Suppose I'm on just say a sidewalk. And there are different. I could either step one step to the right or one step to the left and there's some probability. Maybe I choose to step to the right with probability of two thirds and to the left probability of one third. And so you can see where would I go over time. But the key thing about it is that, that the step I'm going to take now only depends on where I am now. So where I'm going to end up next only depends on where I am now. So there's a finite memory. I don't have to know everything I've done in the past to know what's going to happen next. I only need to know where I am now. And that's the key Markov property, that you only need to really know the current state. Don't have to know the whole history to have all the information about the probabilities for what's going to happen happen next.
- [72:38]
Tom Bilyeu
The. An analogy that I heard that I was really helpful in understanding is if you think of it as airports, some airports have more connections to other cities than other airports. You're. So if you're asking, let's say that there's five airports in question. One is isolated and one is a hub to all the rest. And then the other ones only have one or two links. Whatever. Going back to your idea of if I'm on the isolated airport, there's only one option. So you don't need to know where I was before all of that. If you know I'm on the isolated one, you know, I'm flying back to the only thing it's connected to, which is the other hub.
- [73:17]
Donald Hoffman
Right?
- [73:18]
Tom Bilyeu
Now, when I'm at that hub, that has, let's say, five options. Now it's just a probability curve of which one I'm going to go to. But once I go to another one of those airports, then it's like, okay, well, I could go, you know, to Cincinnati. I could go to New York. I could go to la, or I could go, let's say those are the only connections. But when I'm in Hawaii, if Hawaii forces me to route through la, then you know where you're going to go. I was like, okay, that, that at least gives a simple understanding of oh, this is a relatively simple concept that sets aside all the history. And so from a computational standpoint that becomes very important because when people talk about booting up a simulation of the universe, you very quickly to track every element that could possibly interact with every. If everything could interact with everything, it becomes impossible. And you would have to have a computer the size of the universe itself in order to track like a one for one atom, basically. But I think I'm understanding this right, that Markovian dynamics eliminates a lot of that computational need. Need. Because I don't have to. There is a small set of things and once I know the probability distribution over time it completely stabilizes. And so when I, I know if I'm at airport C, I know the exact probability of where they're going to go next.
- [74:39]
Donald Hoffman
That's right. So Markovian dynamics is help simplify things by demanding only a finite memory instead of an infinite memory of the past history of what you've been doing. But you can make the memory as big as you want. So it's really not too much of a limitation either. So it's a nice formalism.
- [74:58]
Tom Bilyeu
Why do we care about it?
- [75:00]
Donald Hoffman
Well, most of us don't have to deal with infinity anyway in terms of past history, so we can just use finite histories and that's quite good. And another reason to be interested Markov dynamics is we talked about computable functions. Well, Markovian kernels are computationally universal. So anything that can be computed with a neural net or with a universal Turing machine can be computed with Markovian kernels. So they form a. They give a nice network kind of modeling for dynamics, but they also give us universal computational abilities. And they're not limited to computable functions because the sets on which the probabilities are defined need not be computable sets. So they actually give us a window toward going beyond computation. I'm not there right now, but that window is there in the future if we need to go there. Hopefully that will go there. But so our current model is a Markovian model of conscious agents. And then what we have to do is we can then show that space time is just a projection of this dynamics. And so you only. There's a lot of states really fast
- [76:16]
Tom Bilyeu
before you Move on. So just re anchoring people that these conscious agents, the states that they can be in are coffee, elation, desire, headache. So when we're talking Markovian dynamics, we're talking about moving from one of those qualia states to another. A human headache versus a dolphin headach headache, etc, etc. Right. So help me understand why that's important that I can like, if I'm in the state of blissed out coffee, taste that I have a certain probability of going somewhere else that, that feels counterintuitive. It feels like my wants and desires are really what's going to drive the next state, not, not the state that I'm currently in.
- [77:09]
Donald Hoffman
That's right. So, so now we're just talking about the consciousness, not about space time. For, for this question.
- [77:16]
Tom Bilyeu
Yes. Right.
- [77:17]
Donald Hoffman
So there when we write down a Markovian kernel and say, okay, whatever your conscious experiences are now this Markovian kernel describes what your next conscious experiences will be probabilistically and also what, how you're influencing the conscious experience of others. So, so now we can ask the kind of question you're asking.
- [77:39]
Tom Bilyeu
So is that's happening outside the headset.
- [77:42]
Donald Hoffman
This is all outside the headset, right? This is all out. This is.
- [77:44]
Tom Bilyeu
So the probability of what I do next is determined outside the headset by Markovian dynamics.
- [77:49]
Donald Hoffman
That's why we're going to get to this dark energy and dark matter stuff.
- [77:52]
Tom Bilyeu
You are breaking my brain right now.
- [77:53]
Donald Hoffman
That's. So that's, that's why I brought this up is because your question was about dark energy and dark matter. So what we have to do to get at that from this point of view, what we're going to say is, look, most of the states of this dynamics are states that are not represented in space time. They're dark. So there are these influences that you're not going to see. When you count up all the matter and all the energy that you can see inside space time, you're going to be missing all the stuff that didn't project into space time. So in fact probably the dark energy and dark matter is much more than we've discovered so far. So, so that's why it's important.
- [78:26]
Tom Bilyeu
But, so okay, hold on. This all really does start to feel weird when I remind myself that this is about qualia, the sense of it being like something. And so I'm gonna make something up. Dark energy is the energy created. This is why I don't understand how it could be energy. But dark energy is the energy of a qualia that I will never be able to Experience. So it's something like an alien drinking blood wine, making that up. But it has to be qualia, so it's got to be something to be like that thing. Is that right?
- [79:08]
Donald Hoffman
It's even more complicated than that. It's not just one qualia. It's probably, who knows how many countless infinities of qualia, things like that. Exactly right. That are interacting and affecting the dynamics that we perceive inside of our space time headset. But notice that among the qualia are, for example, the qualia that you are about four feet from me, so your position. So position there's a quality mean it's very, very different to experience you four feet from me than four inches from me. Those are very. So depth and space is quality. And in fact, our quality of the air sort of compresses. If I look at the, like a distant mountain and the moon rising over that mountain, the moon looks a little further than the mountain, but not much, right? Yeah, the moon's a little further. But if you were to, you know, that mountain might be, you know, 20 miles from me. The moon is a quarter million miles from you. So that means you have no idea that it's like orders of magnitude further away. So, so our qualia space of depth is quite compressed compared to what we might call the measured world. So like, when you actually, and you see that in, in your, you know, like a Grand Theft Auto, when you're actually looking around, you only see the roads around you in a little bit, but the Grand Theft Auto world, you might be able to drive thousands of miles in, in a really complicated simulation. You don't see thousands of miles in any one time. You only see a little bit that your headset allows you to see. But, but because you use that same headset, you're, you're not stuck in that world. You're. There's actually a supercomputer that has a much bigger world than your headset, right. Than what you see right now in your headset, but it's rendering a little bit in your headset right now. So that's why the, the mountain and the moon look about the same, because their headset we can now, of course, when we go to the moon on a rocket now, it's like going through Grand Theft Auto with your headset on and going places that you couldn't see because they were too far away in your current headset view. But you can get there eventually. And, and so that also is pointing to a world outside of your headset. Your headset is just what Little bit of that world that you're rendering at, at any one time. Now, dark energy and dark matter, you're
- [81:33]
Tom Bilyeu
not really getting outside of your headset to go to Mars. You're getting outside of what you. You rendered previously.
- [81:39]
Donald Hoffman
Well, so at any moment you're only seeing in your headset. Right.
- [81:44]
Tom Bilyeu
But if I go to Mars, I'm still seeing in my headset.
- [81:46]
Donald Hoffman
Yeah. And in Grand Theft Auto, for example, there might be a Porsche that's a thousand miles away and you're gonna have to drive like three hours in the game to get there, so you're not going to see. So it's in the simulation outside of your headset. Right. Now to get it in your headset, you're gonna have to do all this work to get it inside your headset. But it already existed in the software, in the computer prior to that. You just don't see it in your headset.
- [82:08]
Tom Bilyeu
Understood.
- [82:10]
Donald Hoffman
So all the stuff inside space, time, the galaxies that we see that are far away from us and so forth, that's not dark matter, dark energy. That's more like the headset stuff that you see in Grand Theft Auto if you go far enough within the game. But then there's this deeper notion that there are some states in the computer that you'll never see see in. In Grand Theft Auto, but they could, you know, subtly influence what you are seeing in Grand Theft Auto.
- [82:36]
Tom Bilyeu
Doesn't your thesis necessarily. No, you're not going to say yes to this, but I'm going to finish. Doesn't your thesis necessarily mean that that is some element of the. I like to think of it as a blob that is consciousness cycling through. Why would it be in the same simulation cycling through different qualia. But then I don't understand why it would be in the same simulation if it's going to be something I could never possibly interact with.
- [83:09]
Donald Hoffman
Right. I mean, almost everything that the real consciousness is doing is not in our. In our headset. We have this. What we're perceiving is probability zero of what's going on. It's. It's basically if. Yes. Of all the things that are being experienced in consciousness, what percent of it do we experience? 0%.
- [83:28]
Grainger Advertiser
If you work in university maintenance, Grainger considers you an MVP because your playbook ensures your arena is always ready for tip off. And Grainger is your trusted partner offering the products you need all in one place, from H vac and plumbing supplies to lighting and more, and all delivered with plenty of time left on the clock. So your team always gets the win. Call 1-800-GRAINGER visit grainger.com or just stop by Grainger for the ones who get it done.