Audio Player
✓ Using synced audio (timestamps accurate)
Starting at:
David Wolpert on the Monotheism Theorem, Penrose, Consciousness, Free Will, and Uncaused Causation
March 10, 2022
•
1:56:35
•
undefined
Audio:
Download MP3
✓ Synced audio available: Click any timestamp to play from that point. Timestamps are accurate because we're using the original ad-free audio.
Transcript
Enhanced with Timestamps
259 sentences
17,558 words
Method: api-polled
Transcription time: 115m 2s
The Economist covers math, physics, philosophy, and AI in a manner that shows how different countries perceive developments and how they impact markets. They recently published a piece on China's new neutrino detector. They cover extending life via mitochondrial transplants, creating an entirely new field of medicine. But it's also not just science they analyze.
Culture, they analyze finance, economics, business, international affairs across every region. I'm particularly liking their new insider feature. It was just launched this month. It gives you, it gives me, a front row access to The Economist's internal editorial debates.
Where senior editors argue through the news with world leaders and policy makers in twice weekly long format shows. Basically an extremely high quality podcast. Whether it's scientific innovation or shifting global politics, The Economist provides comprehensive coverage beyond headlines. As a toe listener, you get a special discount. Head over to economist.com slash TOE to subscribe. That's economist.com slash TOE for your discount.
This is Martian Beast Mode Lynch. Prize pick is making sports season even more fun. On prize picks, whether you're a football fan, a basketball fan, you'll always feel good to be ranked. Right now, new users get $50 instantly in lineups when you play your first $5. The app is simple to use. Pick two or more players. Pick more or less on their stat projections. Anything from touchdown to threes. And if you're right, you can win big. Mix and match players from
any sport on PrizePix, America's number one daily fantasy sports app. PrizePix is available in 40 plus states including California, Texas,
Florida and Georgia. Most importantly, all the transactions on the app are fast, safe and secure. Download the PricePix app today and use code Spotify to get $50 in lineups after you play your first $5 lineup. That's code Spotify to get $50 in lineups after you play your first $5 lineup. PricePix. It's good to be right. Must be present in certain states. Visit PricePix.com for restrictions and details.
This is part two of the conversation with the great David Walpart. Part one is in the description, and I recommend that you watch that first in order to gain an overview of the professor's work. In this podcast, we cover what the definition of an observer is, and we also cover many audience questions, including those from the inventive and prolific Chris Langen, Karl Friston, and Neil Seth, and Kevin Knuth, among others.
We also cover a topic virtually never touched on in the public discussion of the foundations of mathematics and philosophy called algorithmic information theory. David believes that the only unquestionable results in philosophy are those found in algorithmic information theory, so we perform a cursory examination of it.
Lastly, David makes a distinction between the what consciousness and the how consciousness. Personally, I'm generally a fan of disambiguation and think that many philosophical problems exist because of a befuddlement of our confected language, which later generations disembroil and then they think back and wonder why previous generations were so confused. In fact, John Verbecky also separates consciousness into adjectival consciousness and adverbial consciousness.
Click on the timestamp in the description if you'd like to skip this intro. My name is Kirchheim Ungel. I'm a Torontonian filmmaker with a background in mathematical physics dedicated to the explication of the variegated terrain of theories of everything from a theoretical physics perspective but as well as analyzing consciousness and seeing its potential connection to fundamental reality whatever that is. Essentially this channel is dedicated
to exploring the underived nature of reality, the constitutional laws that govern it, provided those laws exist at all and are even knowable to us. If you enjoy witnessing and engaging with others on the topics of psychology, consciousness, physics, etc., the channel's themes, then do consider going to the Discord and the subreddit, which are linked in the description.
There's also a link to the Patreon that is patreon.com slash Kurt Jaimungal if you'd like to support this podcast as the patrons and the sponsors are the only reasons that I'm able to have podcasts of this quality and this depth given that I can do this now full time thanks to both the patrons and the sponsors support.
Speaking of sponsors, there are two. The first sponsor is Brilliant. During the winter break, I decided to brush up on some of the fundamentals of physics, particularly with regard to information theory, as I'd like to interview Chiara Marletto on constructor theory, which is heavily based in information theory.
Now, information theory is predicated on entropy, at least there's a fundamental formula for entropy. So, I ended up taking the brilliant course, I challenged myself to do one lesson per day, and I took the courses Random Variable Distributions and Knowledge Slash Uncertainty. What I loved is that despite knowing the formula for entropy, which is essentially hammered into you as an undergraduate,
It seems like it comes down from the sky arbitrarily, and with Brilliance, for the first time, I was able to see how the formula for entropy, which you're seeing right now, is actually extremely natural, and it'd be strange to define it in any other manner. There are plenty of courses, and you can even learn group theory, which is what's being referenced when you hear that the standard model is predicated on U1 cross SU2 cross SU3. Those are Lie groups, continuous Lie groups. Visit brilliant.org slash tau, t-o-e,
to get 20% off an annual subscription, and I recommend that you don't stop before four lessons. I think you'll be greatly surprised at the ease at which you can now comprehend subjects you previously had a difficult time grokking. The second sponsor is Algo. Now, Algo is an end-to-end supply chain optimization software company with software that helps business users optimize sales and operations, planning to avoid stockouts, reduce return and inventory write-downs while reducing inventory investment.
It's a supply chain AI that drives smart ROI headed by Amjad Hussein, who's been a huge supporter of this podcast since near its inception. In fact, Amjad has his own podcast on AI and consciousness and business growth. And if you'd like to support the Toe podcast, then visit the link in the description to see Amjad's podcast because subscribing to him or at least visiting supports the Toe podcast indirectly. Thank you and enjoy.
Professor, why don't you explain what Laplace's Demon is and then why there's an impossibility result about the existence of Laplace's Demon? Okay. Not surprisingly, Laplace's Demon is named for Simon Laplace, worked towards the end of the 18th century, mostly the early 19th century. He was, among other things, the first person to really use Bayesian statistics. He applied it to planetary motion.
to try to infer what the actual true or make more accurate estimates of the true parameters of the motion of the planets is a whole bunch of mathematics that we use that undergraduates will be using in the physics and math majors and so on has his name involved. So he's a very, very bright guy.
Now he was working at a time when people thought that everything was ultimately Newtonian classical physics, well before relativity, well before quantum mechanics. And he realized that in that paradigm, if you give a brilliant scientist or a demon sufficient information about the state of the world at present,
they can predict exactly what the state of the world would be in all of its details arbitrarily far into the future. And this was true before not only many notions of quantum mechanics or relativity, but also before people knew about chaos or any of the other kinds of phenomena that have been uncovered in the 20th century that might make one think that the Apostles Demon can't quite hold.
So that's his demon. One, the work that I've been involved with on inference devices, it's almost kind of a trivial, but one of the kinds of things that it uncovers is that regardless of what actually the laws of physics are, regardless of whether it's got anything to do with what's currently called the standard model, whether it's quantum physics,
Whether it's classical physics, it could have been Newtonian. It could have been Newtonian where there was no chaos. It could have been exactly the paradigm that Laplace was thinking in. But even given that kind of a hypothesis, his demon could not exist. There's a logical fallacy embedded in the very notion. And that is actually intuitively, it's quite simple. You can set it up in terms of
What's sometimes called a Cretan liars paradox. Some people say Epimenides was the person who first wrote it down. Others say it actually has earlier roots still. But basically you can set it up in such a way that even in a classical universe where it's a finite one even, let's make it be finite, with no chaos,
You can very easily construct possible questions concerning the future state of that universe that Laplace could not possibly get correct. By definition, whatever his prediction is, it would be exactly wrong, precisely because Laplace is part of the universe. That is the key. So it's not too hard to convince yourself of why that might be true.
And we can work through the details if you want. I don't know how you might actually get this out to your audience, but there's some work. I can show a slide later. You can just send me those. Yeah, maybe we can do that later. Yeah, let's get to all that later.
So in any case, so that is part of a general inference devices really grew out of the desire to quantify and build a formal structure that goes beyond this impossibility argument of the positive demon, which by the way, I have not actually proven it to you, certainly not in the generality that I stated it about how it's impossible where it doesn't depend on the laws of physics.
So what inference devices is about, that whole formalism, at root what it says, what it first notices is that let's go for, let's adopt the one attribute of reality that most card carrying physicists, certainly cosmologists, would ascribe to independent of the detailed physical theory.
which is that the human beings themselves are physical systems. So it's really a perspective that turns conventional philosophy on its head in that rather than try to understand what our scientific theories might be in terms of wealth,
we just discovering the external reality or like an instrumentalist would say we're discovering useful variables for making predictions about the external reality and so on. Instead to practicing physicists, as I said especially those who deal with for example cosmology, rather physical reality is the starting point and then you derive human beings
as an approximation to some of the equations of the actual true world. So to whatever degree those epiphenomena of human beings can themselves make accurate theories about this physical system they're embedded with, that's a secondary question. So rather than starting from the perspective of humans saying what is it that we're talking about when we talk about science,
How does it relate to the external reality? The physicists make that be a very, very secondary, minor question saying, let's start with the physical reality. Then human beings are just one of the phenomena in that physical reality that we can describe completely. And if we wanted to, we could drill down into even more minute, semi-trivial questions like, well, some of those human beings, they call themselves philosophers of science.
And they're looking at these various things that other human beings called scientists that come up with and they're trying to worry about how they might relate to this underlying physical reality. And yet that's all well and good, but it's very, very secondary to the presumption that physical reality is just the set of mathematical questions and we humans and all of our conundrums are just something that comes out from it. So if you adopt that perspective,
Then you can go very, very broad scale on some of the issues that are addressed in philosophy of science and associate branch of philosophy called epistemology. You can say, given any agent, doesn't have to be a human being, that is situated in that physical reality. So the states of its brain are simply a subset of, they only, how to phrase it, so the physical reality,
to allow it to have any kinds of laws, we will reduce it to a set, which is the set of all possible histories of reality through all time, what is called a world line in relativity. So there's going to be one world line in which you, Kurt, are wearing a reddish jacket. There's another world line that would be the same laws of physics, but it's another element in the set
which you're right now wearing instead of a more brownish jacket. These are both world lines that are elements of the set of all possible world line histories. So if I then look at the particular world line in which Kurt is sitting in that, so if I look at the particular event that Kurt is right now sitting in that chair, that actually selects out a subset of all possible world lines. There will be some in which you are sitting in that chair and some of which you are not.
So that event is really just a subset of the set of all possible world lines of the universe. In addition, the event where you pose to yourself a question like what happens when two bodies are gravitationally attracted to one another and have the following masses and momentum come within a certain distance of one another. The set of realities in which you pose that question
is a subset of all possible world lines. The set of realities in which you give a possible particular answer to that question is another subset. And also we can presume that the set of physical realities in which your answer is correct and which your answer is wrong are yet other subsets. So what this sets up, so to speak,
is that there's a relationship between these sets where one subset of all realities where you pose a question, another one is where you give a particular answer, both of which are just physical states of your brain, and then there's a third one which has to do with the actual physical system you are asking this question about.
The answer to your question may in some of those describe that physical system accurately and they may do it inaccurately. They might be wrong in other words. It turns out that just by making that decomposition that everything about you are making a prediction involves questions, answers, and truths. We can also say everything about you're making an observation. You are right now observing something
which you can be then posing a question to yourself, what am I observing? You can then have an answer to that. What am I, what do I think in response to that question? And then it can be the actual truth. And these again are just three subsets of all possible world lines. Whether your observation is correct or is distorted will be determined by the overlap between these world lines. So everything you observe, everything you predict,
everything that you remember, which is another example of this, everything you can control. You could reach out your hand and try to be having the question, where is this pencil going to be one minute from now, and you are making it so. Anything in this most broad sense imaginable of what in philosophy is called epistemology, of your knowledge about this world, about whether your knowledge could be correct or wrong,
we can reduce all analysis of it to saying how is it that these three subsets relate to one another. Based upon that, so notice that this level of formalism we're far beyond any specifics like is it a universe where quantum mechanics hold or class mechanics hold. The universe could be finite, the universe could not have chaos on it, it doesn't matter, nor does it matter how you are
making your questions and answers, whether it be by prediction, predicting whether the sun will come up tomorrow, or what's going to be happening with those two gravitation gravitating bodies I alluded to before, or if it comes via observation, or even memory, or directly control. Any form of these is being covered by the structure of three overlapping subsets. It turns out
That once you do that, that all that is needed is that formalism of three overlapping subsets that you can get a whole bunch of impossibility results. And one of them, to return to the point that you were raising before, has to do with Laplace's demon. You can say that Laplace is somebody asking a question. We can say that what that question involves is his future answer itself.
Then there can be the answer he gives, and there can be reality. And basically, if Laplace's question is of the form, will I not be thinking that the answer is yes one second from now? Whatever answer he gives now, you can set it up in such a way that his answer to his question will actually be the opposite of the truth.
and it doesn't matter what kind of a system he uses to make that future prediction, it doesn't matter if the paradigm in which he was working of classical non-chaotic finite physics was correct, his demon is in this very formal, very broad sense going to be wrong. He as a demon will be wrong and there will still be instances of such erroneousness even if he builds a
So that's the way that we can actually formalize the very simple notion of Laplace's demon. It is a flawed concept in its very presumption, in its very structure, independent of the rules of reality, in such a way that actually allows us to then derive things like the monotheism theorem,
It sounds like what you said is that there exists a question such that he will necessarily be incorrect. So when I was first reading about this, I thought that it was going to go along the lines of there exists configurations of physical space such that he won't necessarily be correct.
But you're saying, no, there's a question such that he will necessarily be wrong. No matter what, no matter what question, answer pair he comes up with, it will be wrong if it is supposed to be related to physical reality. No, there are three things. There's your question, your answer, and then whatever the truth of the matter is. And so there are ways of setting up that triple. I see, I see. So that we can be guaranteed.
that if that question answering reality is supposed to be something that concerns the future state of his brain, then no matter what the details are, we can always set up a question in such a way that the answer
will not actually be the same thing as reality says, but will be the opposite. Question. When you stated it, it reminded me of the liar's paradox or Gerdelian question. Have you found any impossibility results that are on more trivial questions? And when I say trivial, I mean, recall back in the early 1900s, when Gerdel came up with this, some mathematicians said, sure, in some special cases, there exist statements that you can't prove.
But it's not like those are useful statements or statements that we talk about. They're incredibly contrived statements. That was a critique back at the time. It's no longer a critique. But I'm wondering, so it's your impossibility result. Does it cover those non-trivial statements? Sorry, those trivial statements. This particular one about the plosus demon, I would say is one of the more of the trivial ones. Here's a more interesting one.
And I'm going to state it in very loaded language. We were discussing in the previous podcast about you should not use terms. Afterwards, you and I had this back and forth where it's called the persuasion definition, I guess, where you define something some way in a particular way, so you know that you can define out of it everything that makes interesting in the first place, but be that as it may, this is actually not completely an exercise of that. But so let's define an inference device
As a question answer pair. So you are an inference device, a computer can be an inference device. My dog, whose name is Chaco, and he's the most amazing dog in the universe. I'm just to make sure that people are aware of that. Is that a true fact? Is that true? You have a dog named Chaco? Yeah, I named him Chaco. He looks somewhat like the Canyon. He's a good mutt somewhere off of the streets of LA. He's got a bunch of long haired chihuahua in him and other things, and he is
It is really unfortunate that he's been sterilized because he is so gentle. He's so good with people and he is so smart. I never trained him to do anything, but he just can tell by my tone of voice. He took after the gentle and smart nature of his owner. I wish I could say that and rather everything that I took my deepest wisdom about life, all of it I learned by watching him.
Say, no, no, no, be more like him. And then you'll be able to get through this as well. But in any case, these are all inference devices. So you can say that, well, some inference devices and different inference devices are going to be able to make correct inferences about different things. We have this problem of them trying to make these inferences about themselves. Cretan-Liage paradox apply to themselves.
But we can assume that they can make good inferences about almost everything else. And notice that the notion of an inference device, it's not just being able to make predictions, it also had it even in its control. So, for example, a really strong person who's able to move things physically, part of what they can do successfully, them as an inference device, is actually move those objects
That way, Atlas, the Greek god, would have been a… Titan was he? I forget which. Atlas. Atlas. The one with the world on his shoulders? Yeah, yeah. I just can't remember. I don't think he was a god. I think he may have been a titan, but I don't remember. The titans were the predecessors of the gods. In any case, he would have been an inference device, a very strong one. So basically, you can conjecture
that the universe might have a god in the sense of an inference device that can control and accurately predict everything except for in the case of Laplace's demon, that it can go after everything else and do it correctly. So as long as you're not asking god, this putative god or goddess, whatever it would be, deity, deity dude, we'll call it deity dude, deity,
So long as you're not asking, so long as Didi is not asking itself questions along the lines of what will I not be asking a second from now? So long as Didi is asking things like, oh, well, is Adam going to eat this apple that was just proffer to him by Eve? And so long as Didi is doing things like
reaching down his hand and writing on walls, parting red seas, doing things that are more interesting from the non-Judeo-Christian religions. They, in some ways, had much more interesting deities than just the standard Yahweh from Judeo-Christian Islamic religion, but in any case, Abrahamic. We can suppose that there is a universe that does have such a deity in it.
And so where it would be actually correct in everything that it says about everything in its own universe, except that doesn't have this weird kind of structure related to itself, that it's correct. Now let's consider the question though, could we have two Didis?
And it turns out that we need to, in addition to these two deities... As an aside, David, I want to make sure these deities, are they embedded in the physical universe? Yes, that's the key. That's the key. But this physical universe itself, you can define it however you want. And whatever the deity is, so long as it itself, it does things at various times.
And we're not even saying requiring it to be lawful in any sense, just that we can write down a set of world lines. And those world lines are by definition, they are the record of what the deity state is across all space and time is in one of these world lines. So this might very well extend to supernatural variables that are not embedded in the laws of physics. It doesn't even matter.
There's no presumption here that the deity is in any sense lawful, bound by laws. It's simply that its state can be recorded. And in a certain sense, if not, then that almost means by definition non-consequential. Wittgenstein's platitude about that which you cannot talk, shut the hell up, certainly would come to play there because we're in any case. But so now let's consider the
And so this is what I'm going to say. I'm going to say, Kurt, you're an inference device. I'm an inference device. We both have free will if the following property holds.
Whatever state David might be in, it's possible for Kurt to, it's not somehow, there is a element of the states of the set of all possible world lines in which David is in that state and Kurt is in state A. There's also another one in which David is in that state and Kurt is in state B. There's another one in which David is in that state and Kurt is state C. So basically,
The definition of free will is that the state that David is in does not rule out the possibility of Kurt being in various states. In that sense, it's not a matter of I'm controlling you, it's not even having anything to do with statistical correlation. It's simply saying that I cannot be in a state, like for example, where I have in my head the edict, Kurt is not wearing a jacket.
You could, that either could be true or it could be false. In that sense, you are free. And conversely, I am free in that sense of you. So as long as we have free will in that sense, in that definition, and I, it's very, very much tongue in cheek for me to call that free will. But if we have free will in that sense at at least one moment, then it is impossible to have two deities. You cannot have two deities
in the sense that they can both have perfectly accurate inferences about everything other than themselves if they both have free will at least once. That's incredibly interesting. So it's an impossibility result against more than one god? Yes. Yep, that's why it's called the monotheism theorem. It might be that we are in a universe in which you could have one god, who knows? I'm not going to go there. I mean, my personal feeling
There could be universes, I would say, there's no sense in which we can actually rule them out, in fact, in which there are deities. But there cannot be any of them that support two deities.
So we might be in one of those ones in which there is a deity, who knows, but we can't be in one which is more than one. When you say deity, there can't be two omniscient deities? There can't be two that have free will. So you could have one where there's Zeus and Hera, but Zeus can always be in a particular state that restricts Hera from being in some of her particular states. Interesting. In that sense, she does not have free will of him.
At one particular time, one particular state of Zeus, it's just not going to be any possible world line in any of the universe, in our universe, in which Hera is in one of her particular states. There's some limitation. My being in one particular state at one particular time causes a restriction on the possible states that you could be in at that time. And if that's true, then we have no free will. Have you heard of Norton's dome?
Norton's dome. I think I did a while ago, ringing a bell, but I can't bring it up. Sure, it's a thought experiment about Newtonian mechanics and it's to show that Newtonian mechanics isn't deterministic, even though it's often said it is, and the reason is there are certain configurations you can set up such that there's not a unique answer to the differential equations.
You know, ordinarily in physics, just for people to know, one of the reasons why mathematicians quibble with physicists is that physicists hand wave and gloss over many details. And so one of them is whenever we have an ordinary differential equation, we tend to say there's uniqueness in existence. However, that's contingent on something called the Lipschitz continuity. And if you don't have Lipschitz continuity, you don't necessarily have a unique solution. So you basically set up a certain situation with a ball on a dome, and the equation for the dome is fairly simple. It's almost like a parabola.
And then it turns out one solution is it stays there forever, the zero velocity initially. And then another solution is at some point t and the time t is not specified. It goes down some route and any one of them. So that's extremely interesting. Let's imagine we live in a Newtonian world. Is that related to free will? Would you say, or is that not related to free will? That's something different. I know that free will forget about the sense of intention interacting with the laws of nature to produce that effect. Yeah. The Norton's dome, it's also, I think it was actually Sabine.
who has one of her FQXI essays that she points out that chaos is, in some sense, it can be a much stronger phenomenon than people understand, and that you can set up physical systems in which the chaos is to such a degree that actually passed a certain point in time, it is not defined, but the status system will be after that. So I think that's the context in which I ran across it.
It's also their related things, the work that goes back to two people called Porel and Richards. Those were physicists who is well known that, for example, the three body problem, where even if you do, so there you do have the standalicious continuity and so on, you've got gravitational traction and so on. You can set it up to be a, what's called a universal Turing machine.
you basically what you do is you encode the input tape to that Turing machine into the actual precise initial conditions of these three bodies. And then by reading the appropriate bits of the state of the system at some future time, you can figure out what that universal Turing machine state of its tape would be at that time.
What this means is that you can feed in a configuration that's actually the halting problem, so that that physical system, in fact, it violates the Church-Turing thesis. That physical system, it's a state in the future would not be computable. So it's a very closely related kind of result. The impossibility results that I'm describing are stronger in that they don't presuppose differential equations.
they don't suppose any of that kind of lawful structure whatsoever. So that's the sense in which they would hold no matter what the laws of physics are. And that is for that reason as well, that they encompass any possible supernatural phenomenology, impact on the real world, no matter how you might even think about it,
because they're just subsets of the set of all possible world lines, and world lines, in this sense, they're just themselves are a set of events in a completely arbitrary sense of what the word event means. It doesn't even presuppose space-time. Really, the mathematics just says, here is a set, I'm going to call the elements of the set world lines, because that's what they happen to correspond to in our kind of world,
But there are things. And everything that we as humans would call rules and laws concerning them has to do with things like observation operators on them and projections and things like that. But that's all extra structure, just the topology of it. The fact that you have these interlocking subsets, that itself has many consequences. One of which is the monotheism theorem. You could maybe have a god, but you can't have more than one.
I play a bunch of fast and loose games when I describe this result. On the one hand, almost to protect myself in intellectual discourse, I want to say that I'm phrasing that theorem very much tongue-in-cheek, and there's an element of truth to that. But without making a firm
declaration, I would say quite honestly that this does prove that by any notion of a deity in any of the world's religions and any philosopher might ever have come up with, no, you can't have more than one. This actually does prove monotheism. So if you push me to the wall, I'm going to go tongue-in-cheek and say, well, I won't be using my definitions.
But there's a large part of me that's thinking that no, in fact, this does prove that you can't have more, you can't, polytheism is ruled out. Sorry. This sounds like a great time to talk about what an observer is. So an observer in this sense is basically an inference device. If we're talking, so an observer, you can be thinking about there's many different frameworks. The way that I, how to phrase this. Okay. This is a digression. Um,
or a diversion. I'll come back. I'm circling around. So when I do my own personal philosophizing, if that's the word, when I construct, configure my own internal... Hear that sound?
That's the sweet sound of success with Shopify. Shopify is the all-encompassing commerce platform that's with you from the first flicker of an idea to the moment you realize you're running a global enterprise. Whether it's handcrafted jewelry or high-tech gadgets, Shopify supports you at every point of sale, both online and in person. They streamline the process with the Internet's best converting checkout, making it 36% more effective than other leading platforms.
There's also something called Shopify Magic, your AI powered assistant that's like an all-star team member working tirelessly behind the scenes. What I find fascinating about Shopify is how it scales with your ambition. No matter how big you want to grow, Shopify gives you everything you need to take control and take your business to the next level. Join the ranks of businesses in 175 countries that have made Shopify the backbone of their commerce. Shopify, by the way,
powers 10% of all e-commerce in the United States, including huge names like Allbirds, Rothy's, and Brooklinen. If you ever need help, their award-winning support is like having a mentor that's just a click away. Now, are you ready to start your own success story? Sign up for a $1 per month trial period at shopify.com slash theories, all lowercase. Go to shopify.com slash theories now to grow your business no matter what stage you're in.
Razor blades are like diving boards. The longer the board, the more the wobble, the more the wobble, the more nicks, cuts, scrapes. A bad shave isn't a blade problem, it's an extension problem. Henson is a family-owned aerospace parts manufacturer that's made parts for the International Space Station and the Mars Rover.
Now they're bringing that precision engineering to your shaving experience. By using aerospace-grade CNC machines, Henson makes razors that extend less than the thickness of a human hair. The razor also has built-in channels that evacuates hair and cream, which make clogging virtually impossible. Henson Shaving wants to produce the best razors, not the best razor business,
So that means no plastics, no subscriptions, no proprietary blades and no planned obsolescence. It's also extremely affordable. The Henson razor works with the standard dual edge blades that give you that old school shave with the benefits of this new school tech. It's time to say no to subscriptions and yes to a razor that'll last you a lifetime.
Dwelling that's got to do with that I call the philosophy house, so to speak.
What it really amounts to is the following, that there are different, I'm almost exclusively interested in metaphysics. I frankly couldn't care less about philosophy of art, for example. I think it's an oxymoron. Philosophy of morality, I think Harsany's got some interesting things to say. Rawls, less so. Singer, I think has some interesting things to say, but not all that much more than that.
But so mostly it's metaphysics, in other words, what is reality and so on and so forth. And basically, as you strip and what philosophy house is built upon is the understanding that you can do certain investigations where you make certain assumptions. Then when you weaken those assumptions, a lot of it falls apart and you get to a deeper level of reality where there's less structure.
Weaken assumptions and you keep going down and down and down until the whole thing actually eats itself. So up at one level, I'm more than happy to be a steadfast card carrying member of the school of people who call themselves scientists, scientific realists, whatever you might want to call it. Then at another level, you can start to say, well, what is reality? You can't even define it. And
You start to then start worrying about things like inference devices. You can then say, wait a second, combine what you just said about no reality is more real than any other, that all of them are just different, these subsets of possible universes, that you can find these inference devices structures, that's all that any reality is, that none is a privilege over any other.
Doesn't that also, can't you also then apply that to your own reasoning, your own question answer process, and conclude that that itself is going to be illegitimate in a sort of a false demons kind of sense. So you've just concluded that no reality is privileged above any other. But in doing so, if you want to strip down the very last remaining assumption, you are not allowed to assume that your reasoning process itself is somehow proof against that kind of conclusion.
And so, as was put in, I think, as one of the Hitchhiker's Guides, little episodes, that everything vanishes in a proof of logic. And that's where you would ultimately end up going. So the reason I mention that is when you say the word observation, it can mean very different things depending on where we are in this whole edifice
this philosophy house where we're stripping out assumptions and so on. In the context of inference devices, observation is just like prediction, control, or memory, for that matter. It's just a way for you to be posing questions. You then have an observation device that is providing you answers and what the things like the Laplace's demon theorem and all these kinds of things prove to you is that no matter what the observation device is,
There are always going to be situations where the answer it gives you is wrong. So that's what observation means in the context of inference devices. It means other things in other contexts. But in the structure of inference devices, note that prediction and memory are identical. So this is something that has to do with the block model of the universe that we think we were talking about last time.
the second law and the psychological arrow of time. Memory is just retrodiction. It's a prediction about the past rather than the future that we suppose, for reasons having to do with the second law, has to be much more accurate, at least some of our memory, than our predictions. Why do you say that it's a prediction about the past rather than a statement about the past? Because we have no way of knowing it's true. It's identical.
There is no privileged arrow of time. This is something that physicists have been wrestling with, especially cosmologists. I think I mentioned last time there's a 1973 book by Paul Davies, The Physics of Time Asymmetry, and Sean Carroll's got some nice stuff in this. And there's, I have a paper in, I think it was 1992 in the International Journal of Theoretical Physics on these kinds of things. There is no sense in which the past is more real than the future.
All moments are just as real as one another, and all that we are doing as scientists or as people remembering is making predictions in the statistic sense. Despite that prefix, the word predediction, from a scientific point of view, from the point of view of the laws of physics, there's no difference between a prediction and a retradiction. They are both using
data at the present to make a statistical estimate of the state of a variable at a time other than the present. It's a retradiction. You could be wrong. And in fact, that's, that's, of course, a whole nother body of work and things we could discuss about is how in fact the mind very often is wrong. And there's implanted memories and just generally fallibility of memory and all this business about how consciousness is a
just so story that we concoct to try to give us the illusion that we're actually in control, that we have some memory, accurate memory of the past, and that both are actually erroneous. But so anyway, memory is just retrodiction, and so it is as prone subject to all these impossibility results as any of the inference device. So what that means in particular is the plus was only thinking forward in time.
And so he wouldn't even think there was an issue of having to construct a demon that would actually predict the past. He would just simply say, you remember the past. And so, sure, your memories are 100% accurate. In point of fact, no. According to these impossibility results, for the same reason that Laplace cannot predict the future with 100% accuracy, it's impossible to build a 100% accurate memory device
Is there a relationship between the no free lunch theorem and the limits on inference devices? Very good question. So I have found, somewhat to my bemusement, that a lot of the stuff that I am drawn to investigate,
The things that I chew on have to do with impossibilities. No free lunch is one of them. Inference devices is another one. As I think I mentioned to you, the notion that mathematics itself might be inherently stochastic, which as you offline, we discussed about how that's related to intuitionism. Some of the ideas of Nicholas Gisson and intuitionism, of course, traces back to Brouwer, a great mathematician in the 20th century.
Um, and I think that there are some potentially some impossibility results there as well. But all of these I suspect have to be related to one another. Um, I've That's on my I've got everybody's got to-do lists and that's on one of my longer term greatest size font, so to speak, to-do lists is figure out
how to integrate these impossibility results. There's also ones that I haven't come, the ones that I came up with, there are the impossibility results that are involved with, that have to do with Turing machines and girdle and so on. You think they're all related? Well, for example, I'm doing some work at a very, very low burner. I wish I could do it more intensely on trying to
You exploit the church touring thesis to be able to say that within the context of inference devices, to say that if there is a universal touring machine, it's going to be inference device. And can the impossibility results concerning inference devices, do they, for example, tell us something like you can't have two perfectly accurate universal touring machines?
Do they provide another way to prove the halting theorem? And so that's on my to-do list to see if they might be integrated that way. Another interesting thing is that the inference device's structure that I was describing to you had no probability distributions over it. We were just talking about subsets of the set of all possible world lines. Well, let's slap down a probability distribution.
we can now modify some of our definitions of what it means to infer something accurately to be to infer with a higher than certain probability of being correct. We can then say well over all possible probability distributions, so what that means is that rather than the impossibility results of an inference device that concern for example two deities, instead
What you start talking about the probability distribution of all the worst case probability that both deities could be correct, or what is the best case probability that they could be correct over all possible probability distributions. So rather than saying it's impossible for them to both always be correct. Overall, interesting. Yeah.
And guess what? You actually get some results out of that that look a little bit like the Heisenberg uncertainty principle. Is this published work? I think it's mentioned in one of my papers, but it's both that and the Turing machine stuff. Very tentative. There's a lot more that needs to be done and that I want to get to at some point. I mean, there's this meta-optimization problem for
people who are so very, very lucky, like as me, that society provides us the resources to go off and, well, play is the way it feels like to me. But for us, for people who are so, I wouldn't use the word blessed. I mean, I'm in the, fortunate, but fortunate. Yes. And people who are so fortunate, there's this optimization question of, well, there's all these things that you do want to work on.
You've only got some finite number of years left. You don't even know what they are. Prioritize, dude, or do that, as the case might be. And so I'm not sure. So at some point, yes, I do want to work on these things. But the question is when and how and how to get the time. And if you can do it with collaborators, which I try to do, and so on. And it's one of the challenges of being somebody who's as fortunate as I am, is how best to use that fortune.
Have you heard of Donald Hoffman? No, I don't think so. Okay, Donald Hoffman has an argument that says whatever we see is not reality as it is. Oh, yes, yes, yes. Um, I don't think I drilled into it. I don't think really amounts to that much. I'm to be quite honest. Yes, let's hear why. Um, I have to hear that sound.
That's the sweet sound of success with Shopify. Shopify is the all-encompassing commerce platform that's with you from the first flicker of an idea to the moment you realize you're running a global enterprise. Whether it's handcrafted jewelry or high-tech gadgets, Shopify supports you at every point of sale, both online and in person. They streamline the process with the internet's best converting checkout, making it 36% more effective than other leading platforms.
There's also something called Shopify Magic, your AI powered assistant that's like an all-star team member working tirelessly behind the scenes. What I find fascinating about Shopify is how it scales with your ambition. No matter how big you want to grow, Shopify gives you everything you need to take control and take your business to the next level.
Join the ranks of businesses in 175 countries that have made Shopify the backbone of their commerce. Shopify, by the way, powers 10% of all e-commerce in the United States, including huge names like Allbirds, Rothy's, and Brooklynin. If you ever need help, their award-winning support is like having a mentor that's just a click away. Now, are you ready to start your own success story? Sign up for a $1 per month trial period at Shopify.com
Think Verizon, the best 5G network is expensive? Think again. Bring in your AT&T or T-Mobile bill to a Verizon store today and we'll give you a better deal. Now what to do with your unwanted bills? Ever seen an origami version of the Miami Bull?
Jokes aside, Verizon has the most ways to save on phones and plans where you can get a single line with everything you need. So bring in your bill to your local Miami Verizon store today and we'll give you a better deal. Go back and review my notes. He seemed to be... I'm sorry.
So I've actually got, I'm pulling it up right now, for example, my annotated version of his work. Yeah, and there's just too much. If you want to have a splice in from a third podcast, I could research it, but I don't remember why. That's one of the other challenges of being fortunate
is you have to very quickly come to decisions about whether it makes sense for you to be spending any effort drilling down on things. So yeah, I did look into his stuff and I wasn't very impressed. If there's one particular article of his that you think I should maybe look at rather than whatever I did look at, if you could just email that to me, I'll see if there's anything different about that one. Sure. That's a quick aside.
My background's not computer science, as you know. Whenever I've skimmed the church-turing thesis, it seems more like a definition of what it means to compute rather than a theorem per se. It's very close to that. It's very close to that. There's a question of whether it's ultimately vacuous or not. There's something called the physical church-turing thesis, where instead people say things like you cannot build a physical computer that is
that implements algorithms that can come to conclusions that would not be possible to do on a physical Turing machine. And it's squirrely to even define it there. There are different definitions you can find in the literature. Scott Aronson actually has a nice video, it's like two hours long, where he talks about different challenges to the physical church Turing thesis.
One of the reasons, as I recall, that David Deutsch first started working in quantum computation, his original paper, was to try to say that, well, if the physical Church-Turing thesis can be either fully ironclad formalized and or established as actually a law of nature, that means we have to be able to express computation in purely quantum mechanical systems, because quantum mechanics is the law of nature. So that was a large part, as I understand it.
of his impetus for starting to work on what now has become called quantum computation. But yes, you're correct that the Church-Turing thesis, as it was understood, say in the middle of the 20th century, seemed very dangerously close to simply being a definition rather than actually a statement about the physical world. So before we get to the audience questions, you mentioned in one of your talks... By the way, during the break,
I did quickly remind myself about Professor Hoffman. OK. And so what I understand, my take on his work, so in a nutshell, he's a neuroscientist. And in an extension of what we were talking about earlier, a vast elaboration of the notion that consciousness and so on and so forth, it's all a fable.
He has been doing a lot of work showing that, in particular, our understanding of the external world is completely concocted, that it's whatever our sensory apparatus is going to think is actually most effective. Our sensory apparatus is going to be telling us precisely whatever natural selection finds out by trial and error
is the best thing to tell us to get us to act in an efficient way as far as our reproductive success is concerned. That which it is telling us to do that might have nothing really to do with any kind of an external physical reality. It's kind of an elaboration of all the aspects that what you think you are seeing, for example, you are not. It's all being filled in with what's called predictive coding in its broadest sense in the neuroscience community.
and so my understanding of his work was that he has done some interesting things related to that particular scientific issue but then sort of well frankly puffs it up as to saying something about there is no objective reality and so on and so forth it's no rather our human being fallible perception of reality
It can be very, very different. A natural selection, in fact, you would say, which should make it be very, very different from what we would, from what an actual reality is.
Okay, now before we get to some audience, quite and just so you know, as an aside, I don't believe Donald Hoffman, although maybe you have a different paper, I don't believe he indicates that there is no reality. He does say what you said, which is that our relationship to that reality, because the way that he constructs reality, it looks strange, the way that it's like three agents interacting with one another, etc. But that's the true game. Yeah, yeah. That was just based on a quick skim reminding myself from I have to have to admit is a very pedestrian source, which is quantum magazine.
That's okay. And one of your talks, I think in one of your FQXI talks, you mentioned that some of the most profound results in philosophy come from algorithmic information theory. Yeah. And coincidentally, I'll be speaking to Gregory Chaitin in a few weeks. Oh, okay. Please explain to me why you think that. Okay, I can use one of Gregory Chaitin. I mean, he does insist that it be pronounced Chaitin rather than Chaitin.
He's a very interesting fellow in many ways, and he's done some great work. One of his results, in fact, I think, does have this character to it. He's got an impossibility theorem, which is basically it says that there is a integer l such that we can never, ever prove that any particular
computation that we would want to do. Any particular string has comagural complexity greater than L. Okay, there is some such integer and he actually has an estimate of it as I believe based upon Lisp programs of being 4,000 if you're using Lisp as your Turing machine. But think about what that means. There's going to be this number like an L which is like say 4,000 that's saying that you know anything that's longer than 4,000 bytes
So you cannot prove that the Komogoro complexity of any particular string is going to be greater than 4,000 bytes. We know that there's a very finite number of strings that have less than 4,000 bytes. At most, depending on the grammar and syntax of your language and so on, it would be 2 to the 4,000, which would be a big number, but it's very, very finite.
So that means that there's an infinite number of strings that exist whose comagural complexity we can never actually determine. So we can never know what is the simplest program size for determine what the simplest program for actually doing those calculations. It's an amazing restriction on what we human beings can do.
To give you another example, this is one that I know best from my book by Lee and Vitani. It's kind of like the Bible and these Turing machine things. You can actually prove the existence of a function from the integers to the integers, which is always increasing. Sometimes it'll stay the same, but it never goes down and eventually gets to infinity.
such that every function you can possibly compute, no matter how you do it, that is also always increasing and gets to infinity, will be strictly greater than this. You wouldn't have even thought there is such a limitation that could even make sense. Do whatever you want. Say, okay, here's a function which it has the value one for the first million numbers, from one through a million, it's got the value one.
Then it's got the value 2, but that's for the next million to the millionth. Make it be whatever you want. Any rule that you can put down for how to construct this, it's going to actually be getting to infinity faster than this other function, which is a very strange thing. It's one of the most fundamental philosophical results in the sense that philosophy
should not be biased towards what we human beings consider to be compelling. I find it that impossibility results in general, and of these sorts in particular, they are very deep philosophy, whether or not we even appreciate they had meaning before we came across them. What was the name of that one, that second one? Oh, this is, I don't even think it has a name.
I don't even think it has a name. I can point you to the chapter and leave. Yeah, that would be great. But there's another one that, for example, Scott Aronson has a nice blog post on this. It's called the busy beaver function. And he's actually recently written some papers on it as well. And this is kind of the flip side of what I just said, that let's try to make the fastest increasing function possible. So you, Kurt,
Say that, well, for the value 1, it's got the value 1. For the value 2, it's got the value 10 to the 10 to the 10. For the value 3, it's got what it had for the value 2, but now itself. Make it be whatever you want. There is always going to be something which is called the busy beaver function, which is actually increasing faster than the fastest increasing function you can write down. In other words, there's an upper limit to what
There's an upper limit to the speed of increase of any function from the integers to the integers that you can possibly define. It's in a certain sense, it's mind-boggling that there's that kind of a limitation on what we can do. I'm not understanding it correctly. So at first, the way to understand what you said is that you have a function that's increasing and then you can make it increase faster, but then there's a bound? There's a bound. I can write down the definition of a function
And it exists, I can prove it exists, and it is an increasing function, and it will be increasing faster than any function that you can possibly write down. Now, I have a question about that. Like I mentioned, I'm speaking to someone who's an ultrafinitist, an intuitionist, and they don't particularly like existence proofs. They like construction proofs. Yep, and that is the foundation of intuitionism. I don't think that actually Nicholas Gibson
that goes that far, though I'm not sure he might in some side idea one of his papers. But yeah, that was the foundation. They don't like existence proofs. They want a constructive proof only. Is this constructed or you just showed the existence of this function? Oh, the busy beaver function. It has been constructed for the first some number of values of what of the integer. So, you know, we can write down a busy beaver function for one, for two, for three, for four, and so on.
up to some particular value. But I think that in the Busy Beaver function, its entirety, almost by definition, no, it can't be constructed. Because if it could be constructed, you would construct it. That's the point, that this thing exists. Wait, sorry, I don't get that last point. Wait, if it could be constructed, I could construct it, what do you mean? So the definition, so if by construct a function, we mean you write down a program,
that spits out its values. That's what we mean to construct a function. Then the busy beaver function is something that increases faster than could be the output of any such program you can write down. It's very, very easy to define. It is purely an existence function that you can't construct it, but I can define it very, very easily.
It takes only a couple of sentences to define what the thing is. So there are these kinds of results which I find in many ways flabbergasting because if one adopts the Church-Turing thesis, these are limitations on human thinking. And there are other ones, and this is in addition to all the ones like the halting problem, you know, halting theorem.
which is very closely related to, of course, some girl's incompleteness theorem. And then there's Rice's theorem and all these other things, which really, I think that the only actual results hear that sound.
That's the sweet sound of success with Shopify. Shopify is the all-encompassing commerce platform that's with you from the first flicker of an idea to the moment you realize you're running a global enterprise. Whether it's handcrafted jewelry or high-tech gadgets, Shopify supports you at every point of sale, both online and in person. They streamline the process with the internet's best converting checkout, making it 36% more effective than other leading platforms.
There's also something called Shopify Magic, your AI powered assistant that's like an all-star team member working tirelessly behind the scenes. What I find fascinating about Shopify is how it scales with your ambition. No matter how big you want to grow, Shopify gives you everything you need to take control and take your business to the
Join the ranks of businesses in 175 countries that have made Shopify the backbone of their commerce. Shopify, by the way, powers 10% of all e-commerce in the United States, including huge names like Allbirds, Rothy's, and Brooklynin. If you ever need help, their award-winning support is like having a mentor that's just a click away. Now, are you ready to start your own success story?
The actual philosophical advances that have been made by humanity are these kinds of results. Everything else is not only questionable,
and is being questioned, it is ultimately going to be quicksand and it's just going to be squishy and you're not going to get anywhere. These are the only ones that we have managed to generate so far. These are it folks. By definition, they are the deepest because they are the full set of things that we have actually established incontrovertibly or as incontrovertible as any deductive logic could be for mathematics.
Okay, I'm going to send you a chat right now. Let me know if you can pull it up on your end, or if it automatically gets sent to you. See there? Well, it's a big one. Okay. Yeah. Okay. So the reason I'm sending it to you is because I'm probably going to miss read this. And it's easier if you read along with me. Oh, these are questions from the audience. Yes, this is not live. Obviously, I'm just saying this before. So this question
We're now we're getting on to some audience questions and this comes from Chris Langan. Let's get to question number two, which says he prior to this was was referencing how in your definition of observer, there's a theological aspect because an observer wants to preserve its own self organization properties, self organization required in the definition of observer. It doesn't, there's nothing in there about what your goal is. There's nothing in there about intentionality in any of the inference devices work.
It's just, you could be observing by luck. So it includes the case where your predictions are correct by pure luck. Or maybe even worse, you thought that you were answering a different question from the real one, but you were screwed up in your thinking, you happened to accidentally get the answers right. Well, no, impossibility results. So you can't even do it that way. You can't get there by averted vision, so to speak. So there's nothing about what you mean.
Then let's do number four. Your work straddles the boundary between mind and matter. What about Cartesian mind slash matter dualism, which says that mind and matter are fundamentally different. Do you agree or do you believe they're both aspects of? I don't see there's anything useful to be had with the Cartesian mind matter dualist perspective. Remember what we were talking about before about
Let me just complete the sentence because the audience doesn't have access. Do you agree or do you believe that they are both aspects of a single underlying reality? Okay, I didn't see that second part in the chat. Mind and matter. Okay, so it's got to a degree, I see. It involves a little part, a little bit where we are in that philosophy house.
at what level we've gone to. And I'm sitting there with all the assumptions about the sort of what I could very loosely and inaccurately describe as scientific realism, then there is zero reason to believe that there's anything in mind that does not matter. And in fact, even most, even people that I think are frankly wrong about many things like David Chalmers worrying about consciousness means and qualia and sorry. Penrose as well. Penrose as well, when he talks about the mind.
Even these people will almost uniformly and almost in total accept that mind is ultimately matter. So that this is part of why I find qualia to ultimately be a vacuous concept is that if you do accept that, I don't see where there's really any wiggle room
to worry about, well, what are your subjective impressions then I'm on to, which is essentially what the quote hard question of consciousness is about. So when I'm sitting up there in my scientific realism room of the house, a mind is just a attribute of matter for the reasons that Hoffman elucidates, Libet elucidates, Dan Dennett has some great work, all the zombie within stuff,
what mind thinks it knows about matter actually is in many, many ways completely fallacious. Then when I drill down below that, of course, you start running into things like solipsism, brain in a vat. Brain in a vat is a cutesy little thing that freshmen debate and has kind of been ignored in the academy. Nobody has got an answer to that challenge because you can't have one. Instead, what do you do is you assume that no, let's assume that we're not a brain in a vat.
And that's the basis for most of metaphysics because you can't go beyond that. But if you were to not make that assumption, if you, as I mentioned before, you go through these various rooms in the house of philosophy by throwing out assumptions, if you don't make the a priori assumption that what we are perceiving is somehow coupled to a reality that we are more than a brain in the bat or matrix, if you wish to view it that way or whatever,
then you're stuck right there. And the question of is there a duality between mind and matter? If we're going to allow that I'm a brain in a vat, then that question takes on completely different aspects to it, that mind becomes matter in a different sense. And you can drill it down even further and start saying that, well, in fact, and I think this is where ultimately the quality of people, the heart problem of consciousness people are coming from, that ultimately mind
comes first and whether there is even matter or not is a TBD. Or to put it in another way, I could define a universe in the sense of inference devices, which arguably is only mind. And there is no such thing as matter. So, you know, once we get down to the level of inference devices, where a universe is just a set of world lines, essentially, we're down at what's called sometimes ontic structural realism.
We've actually gone even further than Max Tegmark's multiverses. And all that any reality amounts to is a pattern relating aspects of these world lines. Laws of the universe are just patterns, which could at any time be violated or not. And once you get to that level that's just patterns, just mathematics, that reality to a pawn on a chessboard is the laws of chess,
Because that's what its reality is. It's almost definitional. Then you have no such, there is no distinction between mind and matter or anything else. It's all just patterns at that level. There is no distinction. The distinction between mind and matter, they come to play with you, throw in other assumptions that put you in other rooms in the philosophy house. So the answer depends where you want to situate yourself and show which assumptions you adopt.
This question is from Carl Friston of the Free Energy Principle. It is a compelling notion that systems extract information from their environment to maintain themselves far from equilibrium, but is this extraction cause or consequence? In other words, are non-equilibrium steady states a consequence of observing or making inferences about other systems?
This is touching on a huge bodywork, my primary body of research these days, which we've actually not discussed at all, which is what's called stochastic thermodynamics, non-equilibrium thermodynamics. Carl right there is, I think, without being sure, without this actually being a proper conversation, is he's trying to lead me into saying things which basically justify his notions of having to mark out blankets and so on and so forth.
Instead, to make a plug for myself is a paper that I wrote with a postdoc Artemiy Kolchinsky, with the title have to do with semantic information and non-equilibrium physics, in which we make the case that without necessarily wanting to say that they are living systems, whether they are conscious systems, in a certain sense, interesting systems,
are those that are what are called non-equilibrium steady states, which is a word that Carl is using there, and that maintain themselves as being non-equilibrium steady states because of the information that they are actually getting from the environment, getting from it in a precise sense of conditional information flows. And I think that those are very interesting systems to
investigate, and that's a whole other body of work that I would like to investigate. But in terms of which came first, is this extraction cause or consequence? To me it's actually much more that there are systems which do have that extraction, and I would say it's neither cause nor consequence. Are non-equilibrium steady states a consequence of observing other systems?
And physicists can give a very simple answer. No, I can define non-Euclidean, but a way to make a non-Euclidean steady state is to simply hook up a system to heat baths at different temperatures. And it will go into what is called in physics a non-Euclidean steady state. Is it observing anything? You could say it's in a certain sense, quote, observing the two different temperatures, but that I think does basically expands the definition of the term observe, so it includes everything.
It is interacting with its external world and there are aspects of it that are describable by information theory. And I think to use words like causal or observation in that context can be a very, very fraught thing. It's in danger of being a deepity, as Dan Dennett calls it. Okay, this next question comes from Professor Anil Seth of neuroscience.
And he has a book out on consciousness, which I'll link in the description. He says, ask David whether he thinks consciousness is substrate independent. So a silicone computer can be in principle conscious. And if so, why? And then as a subnote, with regard to free will, I'm always interested if they referring to you, David and anyone else, if you, David realizes that the determinism versus stochastic debate is a total red herring. Well, let me respond to these.
in order. I once asked Julio Tononi, according to his integrated conscious, his integrated information measure, which I am not a fan of, let me say that. But in any case, Scott Aronson is not either. And so he's got a blog post on it. But according to his measure, a computer, a silicon computer,
whether it implements its code in parallel or in serial determines whether it's conscious or not. I asked him that point blank and he said, yes, that determines whether it's conscious or not. I'm first responding to that aspect of my conversation with Julio before getting to Anil. By the way, I do think Anil does some very interesting work. But so my reaction to that is what is gained by defining
You can define words wherever you want to define it. That's one of the liberties of mathematics is to find things wherever you want. If you could define it so that Julio could say, I am by definition correct, that this is consciousness, my measure. OK, you can you can say that. And that's your definition. That's your definition. But getting back to what we were talking about last time, persuasive definitions. This also gets back to Searle with his Chinese rule argument, which I thought is completely vacuous.
If you are to give a definition of consciousness which has any aspects to it which, if it is going to be restricted to being a characteristic of the behavior of a system, of the what rather than the how, then it cannot differ for a serial or parallel computer.
And that's what Searle also missed. If consciousness, if thinking, whatever you want to call it, is a what rather than a how, and I think it's extremely liberating as a scientist to conceive of it that way, then it doesn't matter by definition. You just said no. The how is interesting, it's worthy of study, but it's a different question entirely from the what and how these are, and these should be disentangled. It is a category error to mistake the how for the what.
Am I to understand that another way of saying that is that if it's a state versus a process? Or no, is that is that false? What I just said? Like, that's not the correct analogy. That's not what I mean to say. You right now, if behind those baby brown eyes, if behind your face were a silicon computer, that has no consequence to the behavior of our discussion. That's what touring great insight with the touring test is.
And the importance of the Turing test is not whether you're using a teletype and things like that, it's to say that as far as he, Alan Turing was concerned, and there's a lot to be said for this, that it is useful to carve out from whatever you want to discuss about consciousness, it is very useful to carve out as its own phenomenon, call it consciousness prime, call it pre-conscious, whatever you want it to be, where we're just looking at the behavior of the system.
And you can have things, one's in silicon, one is instead a real human being, one is instead a simulacrum that's being controlled by somebody outside the room looking in, wherever you want, that's behaving that same way. But if it's behaving the same way, you can't tell the difference. And so for this particular aspect of consciousness, it does not make any difference.
So the Tanoni is actually looking at the other side, he's focusing, he's mixing up the two, I would say, the how and the what. So to me, the interesting thing is the what, you first have that, then once we have that what clarified, which we don't have, we don't know how to define it, even at a behavioral level, what a conscious entity would be, or so on and so forth, we can then start to look at the how. And that's where neurobiology comes to play, which is Anil's
particular comparative advantage, his specialty. And that's a completely open issue, frankly, as in one that he's doing work on involving things like predictive coding, blah, blah, blah, blah, is how do neurobiological systems achieve it? Silicon computers can achieve the same what they're doing in a different way. And so it's a different question.
Then with regard to free will, I'm always interested in whether... Whether they think, slash realize, so that's obviously a loaded question, that the determinism versus stochasticity debate is a total red herring. To say there's a red herring, I mean, that's an extremely loaded leading question, and I'm going to assume that it does have that character because of basically word length, so to speak. A total red herring for what?
It's a question that cannot be answered. I would say that free will is commonly used in common discourse. This is what we were talking about yesterday, persuasive definitions, not yesterday, but last time. Because people commonly use free will, they're not even sure what they mean. To some people, it is determinism versus stochasticity. And that's what they mean. That's what they mean. To other people, it's not. Whether it's a total right herring has to do with how you define the concept in the first place.
I don't view there as being a concept worth discussion. I view it as being pointless. Come up with your definition. Your definition is going to then say whatever you want your definition to mean, but it's your definition. That's why I'm saying we should call it what consciousness rather than how consciousness, just so we're clear in our definitions. I'm not going to say either one is consciousness. I'm going to instead just make two different definitions and say that when people use the word consciousness, they sometimes conflate things.
It sounds like this question sets you off. Well, it does because it's such a waste of time. People are debating definitions.
Life's too short, folks. Define things however you want. Just be clear what it is you are defining. And once we have the definition, and we don't have any definitions, have no import by themselves. It's only things we can conclude starting from the definitions. Make your definitions. Feel free. Just be very clear what your definitions are. And then after that, there might be some interesting consequences to investigate.
But to have fights about definitions is preposterous. And that's what a lot of these fights really are about, is just definition pushing. Let me steel man the definite. I, up to just a few months ago, am on board with you 100%. Now, let me steel man the definitions point of view. I was speaking to Professor Jor Bar-Natson of mathematics in Toronto here.
I thought you might be going there, but here's what I think he actually meant. He thought that coming up with definitions that lead to a very rich set
of theorems and proofs that that is the key to being here is my putting words in his mouth to being a successful and important and do important work in mathematics is to first come up with the very fruitful definitions. I fully agree that but that is not what people who are arguing about free will are arguing about. They're not saying for which definition of free will and he wasn't actually either your professor. He wasn't saying which definition of
He was saying, no, what's important is if I put together the words in a certain way, and then I call them, what I call them doesn't matter. But those set of words put together in this way, we'll call it A alpha B beta. We've just defined A alpha B beta. If I can use A alpha B beta to prove a whole bunch of fruitful things, then that's a good definition. These people instead are saying, what should A alpha B beta mean?
which is backwards. So that professor, he was saying that it's not the words, it's not the clothes don't make the man. It's not the words that you're draping this with that matter. It is whatever that this is. And here people are saying, well, I've got my words first and I'm going to fight with you about whether they actually mean this or they mean that. It's useless. It's backwards.
It's not only anti-scientific, I would say it's anti-intellectual almost.
Now, for the sake of playing the devil's advocate, let me know what you think about this. Now, when we're not talking about mathematics, something that's specifically well defined, when we come up with a definition, it works somewhat akin to this. We have an intuitive notion of some concept and it's ambiguous. And then we give it a name and we usually give it a name by pointing to different objects that satisfy what we think is that intuitive definition. So, for example, with life, we'll say the couch behind you is not alive. You're alive. The beetle is alive. The fire is alive.
Let's say minus 3000 years ago. And then there's other people arguing about that. Well, the fire is not alive, but I do agree David is alive. And I also agree that the couch is not alive. So we have these volutinous feelings of amorphous balls and we're trying to see, okay, well, what's the intersection between us and which one captures what we're trying to say most, which one captures the most relevant concepts. So that's why there are disagreements about the definition. We're trying to figure out, we're arguing about what the entries in the OED should be.
There's an entry in the OED for life and what should it be in the Oxford English Dictionary. And so, for example, on that particular topic, people do define life as a bit. And the question is, well, how should I define that bit? They want it to always be it's alive or it's not alive. How do we define what one or the other means? So this is other work of mine early in my career I did with a fellow called Bill McCready and others have picked it up since.
So my reaction to that was that this is silly definition plugging come up with your definition and whatever to say something is or is not alive is silly. At a minimum if you're talking about biological organisms it's not going to be just a single bit it's going to be actually at least a real number and well actually it's going to be a vector well actually no it's going to be some much more higher dimensional thing it's not going to be life is a bit
life is some other mathematical structure. And here's one that we happen to come up with that actually turned out to be somewhat fruitful. Another topic of research I wanted to work with. You know what a fractal is? I assume. Okay. So I take the magnifying glass and magnifying that microscope, but I keep seeing the same thing. Let's take that magnifying glass, that microscope, put it on your skin. As we keep running up the power,
We go down to the level where we're seeing now individual cells. Now we start going within the cells and so on. The images in that microscope keep changing. You are not self-similar. You are self-dissimilar. The patterns in you are very different from one another at each different scale. In fact, that seems to be true of almost all systems that are alive.
It is not true of almost all systems that are not alive. Moreover, this self-disimilarity is a way to give nuance and structure to the form of the life. Because you can actually be looking at how is it that the patterns at each scale are related to one another. You've got an entire matrix. And so this we call it self-disimilarity. And we were very careful not to say it's living or not.
And we simply say, actually, I think we've reversed this effect. Any system that does have self dissimilarity, high self dissimilarity, is almost by definition going to be very interesting. And some people might say it's actually a complex system. But we, again, we don't want to be saddling it with these definitions. We're just going to simply say it's self dissimilarity.
We put it out there as a proposal to the research community that they shouldn't be spending all this effort in trying to figure out what should the definition of this bit, living or not living, be. That that's fruitless and it's just going to be an interminable wars that never are going to end up being rich in the sense of definitions that your mathematics professor was thinking. Let's instead be thinking about the self dissimilarity of different systems.
Because why would you think that an interesting living system might be self-disimilar in the first place? So it's not just that we can notice phenomenologically that living systems are all self-disimilar, whereas non-living systems like mountaintops and so on are self-similar. We can actually ask why might that be the case? And so here's an interesting proposal that I have not yet time to investigate.
What if biological systems, this is somewhat along the lines of the things that Carl Friston is interested in, what if biological systems had to, and also for that matter Don Hoffman, had to, as an evolutionary imperative, extract the maximal computational bang out of a given physical body buck? How might they do that? Well, one thing you can think of doing
is distribute different parts of that computation to different spatial locations in that body. And that's what systems do, biological systems to a large degree. They have the computations that are going on in your brain being a different part of you from the computations that are going on in your kidney, for example. The computations have been distributed. But you might also not only want to distribute the computation across different positions, but also across scales.
There are, if you think about it, different computations being done by the cells in you compared to the computations being done by the organs in you. Moreover, these computations being done at the different scales are actually communicating with one another. So you can imagine that as an evolutionary imperative, biological systems would be driven to want to distribute the computations that they do
across their different scales, as well as their different spatial positions within them, which would then lead us to start to think about what does it even mean to distribute computation across scales? What are the general features of such systems? And it would lead you down a research path, which I've not had time to go, but it would lead you down a fruitful research path to, in fact, refute and dispute and
I vow no to the concept that life can be reduced to a bit. If we start by saying that this exercise is an exercise in futility, that instead we want to come up with not a definition that's reducible to a bit, you are or you are not, but rather look at this much more nuanced structure, it might lead us in much more fruitful research directions. It might be a much more useful definition in the sense of that mathematics professor you were describing before.
Hi, I'm here to pick up my son Milo. There's no Milo here. Who picked up my son from school? Streaming only on Peacock. I'm gonna need the name of everyone that could have a connection. You don't understand. It was just the five of us. So this was all planned? What are you gonna do? I will do whatever it takes to get my son back. I honestly didn't see this coming. These nice people killing each other. All Her Fault, a new series streaming now only on Peacock.
Now, if you can answer this one quickly, because this is an aside about what you said, you mentioned that it may not be a bit, it could be a vector. Now, I see how it could not be a bit, it could be a real number. But I don't see how is it that this dissimilarity measure could be a vector like R2, R3, etc. So here's what I can do. I take that microscope, this is an exercise that we've actually did on various image data sets, for example. And we applied it to things like the logistic dynamics, and so on and so forth.
Take that microscope that's getting distributions of the visual images of the patterns at different scales. These are now going to be distributions over the pixel pattern within the microscope's field of view. There's many measures from machine learning, probability theory, information theory for saying how different two distributions are. So this way I can say
For any one of the following, say 10 scales, how different is the distribution over images at those two scales? So I'm comparing scales magnification two times and magnification seven times. Those scales have associated distributions over images. I look at how different those distributions over images are at the scale two times and seven times. That gives me a real number.
I then look at it for another pair of magnifications, that gives me another real number. So in this, so in essence what you were seeing that this set of differences between distributions at the different scales can be wound up into one, actually it's a matrix, because you're putting in the two scales that I'm comparing into these elements of these matrix. So if you give me two scales, I'll give you a real number,
is actually more rich than a vector as a full matrix. Symmetric matrix? Yes, this particular case is symmetric, exactly so. Okay, so now this question comes from Kevin Knuth and I know you gotta go, don't worry, I gotta go to Kevin Knuth.
Our friend Kevin Knuth, and I'll read it. So he wants to know what is life. It's a question that people have entertained for a long time without obvious answers. Your thoughts on observers, I sent him your video on observers, seem to be relevant. I have always thought of living things as systems that use information to create some kind of model of the world.
around them in order to locate and take advantage of potential energy gradients, which are then used to keep them away from equilibrium. And I can give some examples and so on. I would agree that a hurricane is not alive. A hurricane is a heat engine that he's referencing your talk. However, it's not alive because it doesn't use the information about its surrounding to model these energy gradients. The models living things can employ are quite diverse for humans. These are our mental models, etc, etc, some bacteria, etc, etc. So these are my thoughts about life. I really need to write them down.
Maybe David would be interested in working with me on these. Okay, so these are Kevin Knuth's thoughts. I'm sure you can pull a question out from that or just comment on it as a whole. Yeah, so what Kevin, what he's talking about here, first of all, hi, Kevin. And yes, I think we could have some very interesting discussions about here. So as I mentioned before, and okay, so backing it up.
Kevin is talking about presentation I made. That's where there was observers and hurricanes and so on. The observers are what I was there using as a term to describe these non-equilibrium steady states that if you were to basically, in a Judah-Pearl sense, intervene on the information flow from the outside world into these systems, they would no longer be at non-equilibrium. They would actually collapse to being at equilibrium.
And we were calling that, and this is what we were calling an observer. This was what they were calling this talk of mine that Kevin saw. And we were making the definitions be, we were careful so that we would not, I mean, the words observer and alive are being just mangled around here. But that we were trying to make it so that a hurricane is not in a certain sense observing its environment.
whereas a paramecium going up a gradient of food is observing its environment and is alive. So, let's see. So, well, I and my collaborator Artemiy Kolchinsky were going after in that paper where we define somatic information and observation and so on
as a system that's in a non-equilibrium steady state, such that if you intervened on the information it gets from the outside world, it would collapse to equilibrium. Equilibrium being a syndrome for depth. We were saying that those are systems that are either observers or living or just generally of interest without using loaded terminology. That is all the what.
Kevin is talking about models that the system has internally that basically play a middle step in the map from the information entering from the outside to the actions that the entity does back on the outside that end up allowing the entity to be away from equilibrium. The model is the way that it does that. It's a way of interpreting what's going on in the entity
in between these steps of getting information from the outside and then acting back on the outside. What we were analyzing in our work elided that step. We were not trying to say anything about internal models. We were simply saying the end result is that if you interrupt the information flow from the outside, we were saying nothing about the reasons, whether it goes through models or some other
phenomena, but the ultimate effect was that you collapse back to equilibrium. So it was saying that if you, if I sunder the information you're going from the outside, you're getting right now information from the outside about where a cheeseburger is on a table. If I stop that information flow, eventually you die, you fall back to equilibrium. So that's it. So what you're doing is you are right now in a non-equilibrium steady state
Because you're getting information from the outside world and what we were analyzing there is systems where if you were to chop that information flow from the outside world, the result is that they actually collapse back down into being equilibrium, they die. The means by which that happens
in your particular case are well, if you had that information about where the cheeseburger is, you could then reach out your hand based upon an internal model of the outside world, grab that cheeseburger, bring it to your mouth, digest it, get a whole bunch of glucose, which is what you then use to stay out of equilibrium. So that's everything that happened in between you're getting the information and you're staying out of equilibrium.
In other processes like the paramecium, it's getting information about the gradient of a chemical in its environment. It is then using that information, one could maybe say in implicit genetic model, that's what Kevin is saying, but whatever. The end result is that it uses that to determine how it should reorient itself before it uses this flagella to actually move so that it can actually then ingest these particular chemicals and that's how it stays out of equilibrium.
So in my work with Artemiy Koltchinsky, all that we did was looked at the beginning point and the end point. Notice that you need the information coming in to stay out of equilibrium. Those are interesting systems, we said, from a physics point of view. Kevin is now discussing a special case of such systems, which are physics systems that capture at least a large part of what one might want to say is a living
So that's a very long-winded answer to Kevin's question. Kevin, yes, if you do want to, in particular, the directed information flow of people like Daniel Polanyi and Nihat Ai, I think could be a very, very interesting, fruitful way to try to build more structure into these particular systems, a halfway point to building full-on internal models.
So yeah, Kevin, drop me a line. Now the last question you can answer, hopefully fairly quickly, it's even though it's long, it's one sentence. It's actually just one sentence. So I love this one comes from Professor Edward Lee of Berkeley. I love David, how your work has complicated the faith many scientists hold in an ultimately deterministic world.
I was wondering whether you've looked at John Norton. So we talked about that. So we actually did talk about that in particular. Why is this example controversial? Why are so many people offended by the idea of uncaused action? Yeah, I've also encountered Professor Lee's work and I find also mutual admiration here that I am impressed with a lot of his work. This is actually a subtle thing. I'm very careful in my stuff on inference devices.
There is no laws discussed, mentioned anywhere. There's no rules. There's no mathematicalization of the patterns in the underlying reality. It's just world lines in a set. And an observer is one particular subset of those world lines. Something it's observing is another subset. What it means to observe
has to do with the relationship between those two subsets. What it would mean to make predictions would then be another subset. And basically by means of these subsets and their overlaps, this topology, you would build up the structure of a scientist using theories to make predictions about what might occur in the world. And those theories are going to be in the form of rules and laws. And so, for example, Max Tegmark's multiverse work
I highly plug his Foundations of Physics paper and the earlier Annals of Physics papers on the multiverse. Don't bother with pop books, go for the real me. In that work, he ultimately reduces the notion of physical reality to a set of laws, to a set of mathematical rules, and the presumption that all physical reality is just a set of rules, which is
precisely why John Norton's example is controversial because it seems to wreak havoc with the notion that physical reality can be reduced to a set of rules. That's, I think, ultimately from a sociological point of view, why people are upset about that. I'm very careful in my work that rules don't appear. There's just subsets, there's a topology, there might be patterns. Those patterns can be violated at any point in time. No free lunch.
Newton's laws, even if they are there, they might just suddenly disappear tomorrow before the ball of John Norton has even had a chance to try to think about falling off of the dome. But many people instead want to implicitly, they think of a mathematical universe as being a set of things like differential equations.
rules, laws, and I just think of it instead as being a set, and there can be a topology on that set, and the topology might have patterns which you might be able to embody as such rules and laws, or it might not. It all just depends on the topology, and different topologies are going to either be rule-like or not rule-like, law-like or not law-like.
And it's not even the case that one is real and one is not real. They're just all different realities. It's legitimate to one, as legitimate as one another. There's no privileged reality. I suspect that all of these kinds of answers I'm giving you these questions are not necessarily fully fruitful. As an aside, as an interviewee, to give a answer to a question in a standalone manner rather than in a dialogue,
Professor, thank you so much for spending so much time with me. And the audience, I'm sure, thanks you too. Okay, well, thanks very much. Sorry if I
It was too much open-loop control. I think I was a little bit logarithmic today. So thank you very, very much. And if you could give me the information about the forum for this, so I can then forward that to the SFIP people. And Kevin, contact me. Great. OK, thanks. OK, goodbye, Professor. Thank you. OK, bye. Truly, it was a blast. It's great to be speaking to someone who also is philosophically inclined. Even though you had your excoriations on philosophy, you still had your philosophy of philosophy.
Yeah, in many ways, yes, I think philosophy is almost by definition, the most important intellectual endeavor there is, but and this is being a real snide asshole. I don't think it's really unfortunate, people that actually engage in it have not been trained properly. And that's most of philosophers. It's unfortunate, the obviously exceptions, you know, Dan, Dan, I think,
Training can either be explicitly or implicitly. And I think Dan is an example of somebody who's, quote, trained well to be really speaking in snotty asinine terms. But yeah, I think it's the most important subject there is, almost by definition. All right, take care, sir. Okay, you as well. Thank you. And we'll talk soon. Bye bye. Okay, thank you.
There were two extra questions from Sam Thompson and Chris Langan that we didn't have a chance to address on video due to limited time, but David was kind enough to answer over email. Sam Thompson asks, do you see any sociopolitical implications in your work on the formal limitations of inference devices? David Walpart says, interesting question.
One could consider, for example, the Laplace's demon theorem or the monotheism theorem as concerning limitations of what a government is capable of, but those theorems are based on worst-case scenarios. They show that there is some question-answer pair that an inference device gets wrong, not that it must get wrong a particular such pair. More interesting and future work would be an extension of the inference theorem, would be an extension of the inference device formalism to incorporate probabilities and degrees of failure.
Now on to the question from Chris Langen.
David, you base your definition of observer on principles of physical causation including thermodynamics and statistical mechanics... David interjects. I've used the term observer in different ways depending on whether I'm talking about inference devices or the semantics of information formalism with artemy or noisy deductive reasoning. Which meaning do they have in mind? So then Chris Langan finishes the question.
Do you subscribe to physicalism, the idea that reality is basically physical in nature? Or do you think that a broader, more sophisticated understanding of causality, one that exceeds the classical concept of physical mechanism and quantum theoretic concepts like entanglement, non-locality, and quantum wave function collapse, will eventually be required? David answers. In short, there is right now absolutely zero data slash evidence to believe that will, quote-unquote, eventually be required. So Occam's razor provides the answer.
The podcast is now finished. If you'd like to support conversations like this, then do consider going to patreon.com slash c-u-r-t-j-a-i-m-u-n-g-a-l. That is Kurt Jaimungal. It's support from the patrons and from the sponsors that allow me to do this full time. Every dollar helps tremendously. Thank you.
▶ View Full JSON Data (Word-Level Timestamps)
{
"source": "transcribe.metaboat.io",
"workspace_id": "AXs1igz",
"job_seq": 10377,
"audio_duration_seconds": 6902.47,
"completed_at": "2025-12-01T01:41:17Z",
"segments": [
{
"end_time": 20.896,
"index": 0,
"start_time": 0.009,
"text": " The Economist covers math, physics, philosophy, and AI in a manner that shows how different countries perceive developments and how they impact markets. They recently published a piece on China's new neutrino detector. They cover extending life via mitochondrial transplants, creating an entirely new field of medicine. But it's also not just science they analyze."
},
{
"end_time": 36.067,
"index": 1,
"start_time": 20.896,
"text": " Culture, they analyze finance, economics, business, international affairs across every region. I'm particularly liking their new insider feature. It was just launched this month. It gives you, it gives me, a front row access to The Economist's internal editorial debates."
},
{
"end_time": 64.514,
"index": 2,
"start_time": 36.34,
"text": " Where senior editors argue through the news with world leaders and policy makers in twice weekly long format shows. Basically an extremely high quality podcast. Whether it's scientific innovation or shifting global politics, The Economist provides comprehensive coverage beyond headlines. As a toe listener, you get a special discount. Head over to economist.com slash TOE to subscribe. That's economist.com slash TOE for your discount."
},
{
"end_time": 94.718,
"index": 3,
"start_time": 66.152,
"text": " This is Martian Beast Mode Lynch. Prize pick is making sports season even more fun. On prize picks, whether you're a football fan, a basketball fan, you'll always feel good to be ranked. Right now, new users get $50 instantly in lineups when you play your first $5. The app is simple to use. Pick two or more players. Pick more or less on their stat projections. Anything from touchdown to threes. And if you're right, you can win big. Mix and match players from"
},
{
"end_time": 104.599,
"index": 4,
"start_time": 94.718,
"text": " any sport on PrizePix, America's number one daily fantasy sports app. PrizePix is available in 40 plus states including California, Texas,"
},
{
"end_time": 126.254,
"index": 5,
"start_time": 104.821,
"text": " Florida and Georgia. Most importantly, all the transactions on the app are fast, safe and secure. Download the PricePix app today and use code Spotify to get $50 in lineups after you play your first $5 lineup. That's code Spotify to get $50 in lineups after you play your first $5 lineup. PricePix. It's good to be right. Must be present in certain states. Visit PricePix.com for restrictions and details."
},
{
"end_time": 149.206,
"index": 6,
"start_time": 126.254,
"text": " This is part two of the conversation with the great David Walpart. Part one is in the description, and I recommend that you watch that first in order to gain an overview of the professor's work. In this podcast, we cover what the definition of an observer is, and we also cover many audience questions, including those from the inventive and prolific Chris Langen, Karl Friston, and Neil Seth, and Kevin Knuth, among others."
},
{
"end_time": 166.613,
"index": 7,
"start_time": 149.206,
"text": " We also cover a topic virtually never touched on in the public discussion of the foundations of mathematics and philosophy called algorithmic information theory. David believes that the only unquestionable results in philosophy are those found in algorithmic information theory, so we perform a cursory examination of it."
},
{
"end_time": 191.271,
"index": 8,
"start_time": 167.039,
"text": " Lastly, David makes a distinction between the what consciousness and the how consciousness. Personally, I'm generally a fan of disambiguation and think that many philosophical problems exist because of a befuddlement of our confected language, which later generations disembroil and then they think back and wonder why previous generations were so confused. In fact, John Verbecky also separates consciousness into adjectival consciousness and adverbial consciousness."
},
{
"end_time": 215.179,
"index": 9,
"start_time": 191.271,
"text": " Click on the timestamp in the description if you'd like to skip this intro. My name is Kirchheim Ungel. I'm a Torontonian filmmaker with a background in mathematical physics dedicated to the explication of the variegated terrain of theories of everything from a theoretical physics perspective but as well as analyzing consciousness and seeing its potential connection to fundamental reality whatever that is. Essentially this channel is dedicated"
},
{
"end_time": 234.258,
"index": 10,
"start_time": 215.179,
"text": " to exploring the underived nature of reality, the constitutional laws that govern it, provided those laws exist at all and are even knowable to us. If you enjoy witnessing and engaging with others on the topics of psychology, consciousness, physics, etc., the channel's themes, then do consider going to the Discord and the subreddit, which are linked in the description."
},
{
"end_time": 251.323,
"index": 11,
"start_time": 234.258,
"text": " There's also a link to the Patreon that is patreon.com slash Kurt Jaimungal if you'd like to support this podcast as the patrons and the sponsors are the only reasons that I'm able to have podcasts of this quality and this depth given that I can do this now full time thanks to both the patrons and the sponsors support."
},
{
"end_time": 266.578,
"index": 12,
"start_time": 251.323,
"text": " Speaking of sponsors, there are two. The first sponsor is Brilliant. During the winter break, I decided to brush up on some of the fundamentals of physics, particularly with regard to information theory, as I'd like to interview Chiara Marletto on constructor theory, which is heavily based in information theory."
},
{
"end_time": 285.572,
"index": 13,
"start_time": 266.578,
"text": " Now, information theory is predicated on entropy, at least there's a fundamental formula for entropy. So, I ended up taking the brilliant course, I challenged myself to do one lesson per day, and I took the courses Random Variable Distributions and Knowledge Slash Uncertainty. What I loved is that despite knowing the formula for entropy, which is essentially hammered into you as an undergraduate,"
},
{
"end_time": 310.964,
"index": 14,
"start_time": 285.828,
"text": " It seems like it comes down from the sky arbitrarily, and with Brilliance, for the first time, I was able to see how the formula for entropy, which you're seeing right now, is actually extremely natural, and it'd be strange to define it in any other manner. There are plenty of courses, and you can even learn group theory, which is what's being referenced when you hear that the standard model is predicated on U1 cross SU2 cross SU3. Those are Lie groups, continuous Lie groups. Visit brilliant.org slash tau, t-o-e,"
},
{
"end_time": 338.183,
"index": 15,
"start_time": 310.964,
"text": " to get 20% off an annual subscription, and I recommend that you don't stop before four lessons. I think you'll be greatly surprised at the ease at which you can now comprehend subjects you previously had a difficult time grokking. The second sponsor is Algo. Now, Algo is an end-to-end supply chain optimization software company with software that helps business users optimize sales and operations, planning to avoid stockouts, reduce return and inventory write-downs while reducing inventory investment."
},
{
"end_time": 363.166,
"index": 16,
"start_time": 338.183,
"text": " It's a supply chain AI that drives smart ROI headed by Amjad Hussein, who's been a huge supporter of this podcast since near its inception. In fact, Amjad has his own podcast on AI and consciousness and business growth. And if you'd like to support the Toe podcast, then visit the link in the description to see Amjad's podcast because subscribing to him or at least visiting supports the Toe podcast indirectly. Thank you and enjoy."
},
{
"end_time": 391.118,
"index": 17,
"start_time": 363.456,
"text": " Professor, why don't you explain what Laplace's Demon is and then why there's an impossibility result about the existence of Laplace's Demon? Okay. Not surprisingly, Laplace's Demon is named for Simon Laplace, worked towards the end of the 18th century, mostly the early 19th century. He was, among other things, the first person to really use Bayesian statistics. He applied it to planetary motion."
},
{
"end_time": 413.148,
"index": 18,
"start_time": 391.664,
"text": " to try to infer what the actual true or make more accurate estimates of the true parameters of the motion of the planets is a whole bunch of mathematics that we use that undergraduates will be using in the physics and math majors and so on has his name involved. So he's a very, very bright guy."
},
{
"end_time": 440.094,
"index": 19,
"start_time": 413.797,
"text": " Now he was working at a time when people thought that everything was ultimately Newtonian classical physics, well before relativity, well before quantum mechanics. And he realized that in that paradigm, if you give a brilliant scientist or a demon sufficient information about the state of the world at present,"
},
{
"end_time": 467.705,
"index": 20,
"start_time": 440.691,
"text": " they can predict exactly what the state of the world would be in all of its details arbitrarily far into the future. And this was true before not only many notions of quantum mechanics or relativity, but also before people knew about chaos or any of the other kinds of phenomena that have been uncovered in the 20th century that might make one think that the Apostles Demon can't quite hold."
},
{
"end_time": 493.097,
"index": 21,
"start_time": 468.319,
"text": " So that's his demon. One, the work that I've been involved with on inference devices, it's almost kind of a trivial, but one of the kinds of things that it uncovers is that regardless of what actually the laws of physics are, regardless of whether it's got anything to do with what's currently called the standard model, whether it's quantum physics,"
},
{
"end_time": 522.466,
"index": 22,
"start_time": 493.609,
"text": " Whether it's classical physics, it could have been Newtonian. It could have been Newtonian where there was no chaos. It could have been exactly the paradigm that Laplace was thinking in. But even given that kind of a hypothesis, his demon could not exist. There's a logical fallacy embedded in the very notion. And that is actually intuitively, it's quite simple. You can set it up in terms of"
},
{
"end_time": 547.739,
"index": 23,
"start_time": 522.995,
"text": " What's sometimes called a Cretan liars paradox. Some people say Epimenides was the person who first wrote it down. Others say it actually has earlier roots still. But basically you can set it up in such a way that even in a classical universe where it's a finite one even, let's make it be finite, with no chaos,"
},
{
"end_time": 572.841,
"index": 24,
"start_time": 548.148,
"text": " You can very easily construct possible questions concerning the future state of that universe that Laplace could not possibly get correct. By definition, whatever his prediction is, it would be exactly wrong, precisely because Laplace is part of the universe. That is the key. So it's not too hard to convince yourself of why that might be true."
},
{
"end_time": 597.466,
"index": 25,
"start_time": 573.285,
"text": " And we can work through the details if you want. I don't know how you might actually get this out to your audience, but there's some work. I can show a slide later. You can just send me those. Yeah, maybe we can do that later. Yeah, let's get to all that later."
},
{
"end_time": 624.275,
"index": 26,
"start_time": 597.944,
"text": " So in any case, so that is part of a general inference devices really grew out of the desire to quantify and build a formal structure that goes beyond this impossibility argument of the positive demon, which by the way, I have not actually proven it to you, certainly not in the generality that I stated it about how it's impossible where it doesn't depend on the laws of physics."
},
{
"end_time": 652.551,
"index": 27,
"start_time": 624.838,
"text": " So what inference devices is about, that whole formalism, at root what it says, what it first notices is that let's go for, let's adopt the one attribute of reality that most card carrying physicists, certainly cosmologists, would ascribe to independent of the detailed physical theory."
},
{
"end_time": 673.899,
"index": 28,
"start_time": 652.995,
"text": " which is that the human beings themselves are physical systems. So it's really a perspective that turns conventional philosophy on its head in that rather than try to understand what our scientific theories might be in terms of wealth,"
},
{
"end_time": 698.712,
"index": 29,
"start_time": 674.241,
"text": " we just discovering the external reality or like an instrumentalist would say we're discovering useful variables for making predictions about the external reality and so on. Instead to practicing physicists, as I said especially those who deal with for example cosmology, rather physical reality is the starting point and then you derive human beings"
},
{
"end_time": 727.312,
"index": 30,
"start_time": 699.428,
"text": " as an approximation to some of the equations of the actual true world. So to whatever degree those epiphenomena of human beings can themselves make accurate theories about this physical system they're embedded with, that's a secondary question. So rather than starting from the perspective of humans saying what is it that we're talking about when we talk about science,"
},
{
"end_time": 753.729,
"index": 31,
"start_time": 727.466,
"text": " How does it relate to the external reality? The physicists make that be a very, very secondary, minor question saying, let's start with the physical reality. Then human beings are just one of the phenomena in that physical reality that we can describe completely. And if we wanted to, we could drill down into even more minute, semi-trivial questions like, well, some of those human beings, they call themselves philosophers of science."
},
{
"end_time": 782.858,
"index": 32,
"start_time": 754.172,
"text": " And they're looking at these various things that other human beings called scientists that come up with and they're trying to worry about how they might relate to this underlying physical reality. And yet that's all well and good, but it's very, very secondary to the presumption that physical reality is just the set of mathematical questions and we humans and all of our conundrums are just something that comes out from it. So if you adopt that perspective,"
},
{
"end_time": 813.814,
"index": 33,
"start_time": 784.121,
"text": " Then you can go very, very broad scale on some of the issues that are addressed in philosophy of science and associate branch of philosophy called epistemology. You can say, given any agent, doesn't have to be a human being, that is situated in that physical reality. So the states of its brain are simply a subset of, they only, how to phrase it, so the physical reality,"
},
{
"end_time": 843.046,
"index": 34,
"start_time": 814.394,
"text": " to allow it to have any kinds of laws, we will reduce it to a set, which is the set of all possible histories of reality through all time, what is called a world line in relativity. So there's going to be one world line in which you, Kurt, are wearing a reddish jacket. There's another world line that would be the same laws of physics, but it's another element in the set"
},
{
"end_time": 873.336,
"index": 35,
"start_time": 843.422,
"text": " which you're right now wearing instead of a more brownish jacket. These are both world lines that are elements of the set of all possible world line histories. So if I then look at the particular world line in which Kurt is sitting in that, so if I look at the particular event that Kurt is right now sitting in that chair, that actually selects out a subset of all possible world lines. There will be some in which you are sitting in that chair and some of which you are not."
},
{
"end_time": 900.913,
"index": 36,
"start_time": 873.814,
"text": " So that event is really just a subset of the set of all possible world lines of the universe. In addition, the event where you pose to yourself a question like what happens when two bodies are gravitationally attracted to one another and have the following masses and momentum come within a certain distance of one another. The set of realities in which you pose that question"
},
{
"end_time": 926.869,
"index": 37,
"start_time": 901.323,
"text": " is a subset of all possible world lines. The set of realities in which you give a possible particular answer to that question is another subset. And also we can presume that the set of physical realities in which your answer is correct and which your answer is wrong are yet other subsets. So what this sets up, so to speak,"
},
{
"end_time": 947.346,
"index": 38,
"start_time": 927.193,
"text": " is that there's a relationship between these sets where one subset of all realities where you pose a question, another one is where you give a particular answer, both of which are just physical states of your brain, and then there's a third one which has to do with the actual physical system you are asking this question about."
},
{
"end_time": 975.725,
"index": 39,
"start_time": 948.609,
"text": " The answer to your question may in some of those describe that physical system accurately and they may do it inaccurately. They might be wrong in other words. It turns out that just by making that decomposition that everything about you are making a prediction involves questions, answers, and truths. We can also say everything about you're making an observation. You are right now observing something"
},
{
"end_time": 1004.65,
"index": 40,
"start_time": 976.203,
"text": " which you can be then posing a question to yourself, what am I observing? You can then have an answer to that. What am I, what do I think in response to that question? And then it can be the actual truth. And these again are just three subsets of all possible world lines. Whether your observation is correct or is distorted will be determined by the overlap between these world lines. So everything you observe, everything you predict,"
},
{
"end_time": 1033.234,
"index": 41,
"start_time": 1005.282,
"text": " everything that you remember, which is another example of this, everything you can control. You could reach out your hand and try to be having the question, where is this pencil going to be one minute from now, and you are making it so. Anything in this most broad sense imaginable of what in philosophy is called epistemology, of your knowledge about this world, about whether your knowledge could be correct or wrong,"
},
{
"end_time": 1063.097,
"index": 42,
"start_time": 1034.155,
"text": " we can reduce all analysis of it to saying how is it that these three subsets relate to one another. Based upon that, so notice that this level of formalism we're far beyond any specifics like is it a universe where quantum mechanics hold or class mechanics hold. The universe could be finite, the universe could not have chaos on it, it doesn't matter, nor does it matter how you are"
},
{
"end_time": 1091.374,
"index": 43,
"start_time": 1063.456,
"text": " making your questions and answers, whether it be by prediction, predicting whether the sun will come up tomorrow, or what's going to be happening with those two gravitation gravitating bodies I alluded to before, or if it comes via observation, or even memory, or directly control. Any form of these is being covered by the structure of three overlapping subsets. It turns out"
},
{
"end_time": 1118.814,
"index": 44,
"start_time": 1091.476,
"text": " That once you do that, that all that is needed is that formalism of three overlapping subsets that you can get a whole bunch of impossibility results. And one of them, to return to the point that you were raising before, has to do with Laplace's demon. You can say that Laplace is somebody asking a question. We can say that what that question involves is his future answer itself."
},
{
"end_time": 1142.483,
"index": 45,
"start_time": 1120.452,
"text": " Then there can be the answer he gives, and there can be reality. And basically, if Laplace's question is of the form, will I not be thinking that the answer is yes one second from now? Whatever answer he gives now, you can set it up in such a way that his answer to his question will actually be the opposite of the truth."
},
{
"end_time": 1170.964,
"index": 46,
"start_time": 1143.712,
"text": " and it doesn't matter what kind of a system he uses to make that future prediction, it doesn't matter if the paradigm in which he was working of classical non-chaotic finite physics was correct, his demon is in this very formal, very broad sense going to be wrong. He as a demon will be wrong and there will still be instances of such erroneousness even if he builds a"
},
{
"end_time": 1199.616,
"index": 47,
"start_time": 1171.271,
"text": " So that's the way that we can actually formalize the very simple notion of Laplace's demon. It is a flawed concept in its very presumption, in its very structure, independent of the rules of reality, in such a way that actually allows us to then derive things like the monotheism theorem,"
},
{
"end_time": 1225.725,
"index": 48,
"start_time": 1200.265,
"text": " It sounds like what you said is that there exists a question such that he will necessarily be incorrect. So when I was first reading about this, I thought that it was going to go along the lines of there exists configurations of physical space such that he won't necessarily be correct."
},
{
"end_time": 1253.336,
"index": 49,
"start_time": 1226.084,
"text": " But you're saying, no, there's a question such that he will necessarily be wrong. No matter what, no matter what question, answer pair he comes up with, it will be wrong if it is supposed to be related to physical reality. No, there are three things. There's your question, your answer, and then whatever the truth of the matter is. And so there are ways of setting up that triple. I see, I see. So that we can be guaranteed."
},
{
"end_time": 1269.002,
"index": 50,
"start_time": 1253.831,
"text": " that if that question answering reality is supposed to be something that concerns the future state of his brain, then no matter what the details are, we can always set up a question in such a way that the answer"
},
{
"end_time": 1294.582,
"index": 51,
"start_time": 1269.36,
"text": " will not actually be the same thing as reality says, but will be the opposite. Question. When you stated it, it reminded me of the liar's paradox or Gerdelian question. Have you found any impossibility results that are on more trivial questions? And when I say trivial, I mean, recall back in the early 1900s, when Gerdel came up with this, some mathematicians said, sure, in some special cases, there exist statements that you can't prove."
},
{
"end_time": 1319.77,
"index": 52,
"start_time": 1295.043,
"text": " But it's not like those are useful statements or statements that we talk about. They're incredibly contrived statements. That was a critique back at the time. It's no longer a critique. But I'm wondering, so it's your impossibility result. Does it cover those non-trivial statements? Sorry, those trivial statements. This particular one about the plosus demon, I would say is one of the more of the trivial ones. Here's a more interesting one."
},
{
"end_time": 1350.913,
"index": 53,
"start_time": 1321.101,
"text": " And I'm going to state it in very loaded language. We were discussing in the previous podcast about you should not use terms. Afterwards, you and I had this back and forth where it's called the persuasion definition, I guess, where you define something some way in a particular way, so you know that you can define out of it everything that makes interesting in the first place, but be that as it may, this is actually not completely an exercise of that. But so let's define an inference device"
},
{
"end_time": 1380.503,
"index": 54,
"start_time": 1351.544,
"text": " As a question answer pair. So you are an inference device, a computer can be an inference device. My dog, whose name is Chaco, and he's the most amazing dog in the universe. I'm just to make sure that people are aware of that. Is that a true fact? Is that true? You have a dog named Chaco? Yeah, I named him Chaco. He looks somewhat like the Canyon. He's a good mutt somewhere off of the streets of LA. He's got a bunch of long haired chihuahua in him and other things, and he is"
},
{
"end_time": 1405.299,
"index": 55,
"start_time": 1381.459,
"text": " It is really unfortunate that he's been sterilized because he is so gentle. He's so good with people and he is so smart. I never trained him to do anything, but he just can tell by my tone of voice. He took after the gentle and smart nature of his owner. I wish I could say that and rather everything that I took my deepest wisdom about life, all of it I learned by watching him."
},
{
"end_time": 1432.892,
"index": 56,
"start_time": 1405.623,
"text": " Say, no, no, no, be more like him. And then you'll be able to get through this as well. But in any case, these are all inference devices. So you can say that, well, some inference devices and different inference devices are going to be able to make correct inferences about different things. We have this problem of them trying to make these inferences about themselves. Cretan-Liage paradox apply to themselves."
},
{
"end_time": 1461.561,
"index": 57,
"start_time": 1433.336,
"text": " But we can assume that they can make good inferences about almost everything else. And notice that the notion of an inference device, it's not just being able to make predictions, it also had it even in its control. So, for example, a really strong person who's able to move things physically, part of what they can do successfully, them as an inference device, is actually move those objects"
},
{
"end_time": 1487.346,
"index": 58,
"start_time": 1461.903,
"text": " That way, Atlas, the Greek god, would have been a… Titan was he? I forget which. Atlas. Atlas. The one with the world on his shoulders? Yeah, yeah. I just can't remember. I don't think he was a god. I think he may have been a titan, but I don't remember. The titans were the predecessors of the gods. In any case, he would have been an inference device, a very strong one. So basically, you can conjecture"
},
{
"end_time": 1515.077,
"index": 59,
"start_time": 1488.2,
"text": " that the universe might have a god in the sense of an inference device that can control and accurately predict everything except for in the case of Laplace's demon, that it can go after everything else and do it correctly. So as long as you're not asking god, this putative god or goddess, whatever it would be, deity, deity dude, we'll call it deity dude, deity,"
},
{
"end_time": 1536.749,
"index": 60,
"start_time": 1515.333,
"text": " So long as you're not asking, so long as Didi is not asking itself questions along the lines of what will I not be asking a second from now? So long as Didi is asking things like, oh, well, is Adam going to eat this apple that was just proffer to him by Eve? And so long as Didi is doing things like"
},
{
"end_time": 1562.415,
"index": 61,
"start_time": 1537.432,
"text": " reaching down his hand and writing on walls, parting red seas, doing things that are more interesting from the non-Judeo-Christian religions. They, in some ways, had much more interesting deities than just the standard Yahweh from Judeo-Christian Islamic religion, but in any case, Abrahamic. We can suppose that there is a universe that does have such a deity in it."
},
{
"end_time": 1581.323,
"index": 62,
"start_time": 1563.951,
"text": " And so where it would be actually correct in everything that it says about everything in its own universe, except that doesn't have this weird kind of structure related to itself, that it's correct. Now let's consider the question though, could we have two Didis?"
},
{
"end_time": 1606.442,
"index": 63,
"start_time": 1583.319,
"text": " And it turns out that we need to, in addition to these two deities... As an aside, David, I want to make sure these deities, are they embedded in the physical universe? Yes, that's the key. That's the key. But this physical universe itself, you can define it however you want. And whatever the deity is, so long as it itself, it does things at various times."
},
{
"end_time": 1633.524,
"index": 64,
"start_time": 1606.886,
"text": " And we're not even saying requiring it to be lawful in any sense, just that we can write down a set of world lines. And those world lines are by definition, they are the record of what the deity state is across all space and time is in one of these world lines. So this might very well extend to supernatural variables that are not embedded in the laws of physics. It doesn't even matter."
},
{
"end_time": 1661.049,
"index": 65,
"start_time": 1633.933,
"text": " There's no presumption here that the deity is in any sense lawful, bound by laws. It's simply that its state can be recorded. And in a certain sense, if not, then that almost means by definition non-consequential. Wittgenstein's platitude about that which you cannot talk, shut the hell up, certainly would come to play there because we're in any case. But so now let's consider the"
},
{
"end_time": 1690.026,
"index": 66,
"start_time": 1661.561,
"text": " And so this is what I'm going to say. I'm going to say, Kurt, you're an inference device. I'm an inference device. We both have free will if the following property holds."
},
{
"end_time": 1721.271,
"index": 67,
"start_time": 1691.561,
"text": " Whatever state David might be in, it's possible for Kurt to, it's not somehow, there is a element of the states of the set of all possible world lines in which David is in that state and Kurt is in state A. There's also another one in which David is in that state and Kurt is in state B. There's another one in which David is in that state and Kurt is state C. So basically,"
},
{
"end_time": 1749.326,
"index": 68,
"start_time": 1721.681,
"text": " The definition of free will is that the state that David is in does not rule out the possibility of Kurt being in various states. In that sense, it's not a matter of I'm controlling you, it's not even having anything to do with statistical correlation. It's simply saying that I cannot be in a state, like for example, where I have in my head the edict, Kurt is not wearing a jacket."
},
{
"end_time": 1779.787,
"index": 69,
"start_time": 1750.35,
"text": " You could, that either could be true or it could be false. In that sense, you are free. And conversely, I am free in that sense of you. So as long as we have free will in that sense, in that definition, and I, it's very, very much tongue in cheek for me to call that free will. But if we have free will in that sense at at least one moment, then it is impossible to have two deities. You cannot have two deities"
},
{
"end_time": 1807.244,
"index": 70,
"start_time": 1780.367,
"text": " in the sense that they can both have perfectly accurate inferences about everything other than themselves if they both have free will at least once. That's incredibly interesting. So it's an impossibility result against more than one god? Yes. Yep, that's why it's called the monotheism theorem. It might be that we are in a universe in which you could have one god, who knows? I'm not going to go there. I mean, my personal feeling"
},
{
"end_time": 1835.52,
"index": 71,
"start_time": 1807.585,
"text": " There could be universes, I would say, there's no sense in which we can actually rule them out, in fact, in which there are deities. But there cannot be any of them that support two deities."
},
{
"end_time": 1865.452,
"index": 72,
"start_time": 1836.476,
"text": " So we might be in one of those ones in which there is a deity, who knows, but we can't be in one which is more than one. When you say deity, there can't be two omniscient deities? There can't be two that have free will. So you could have one where there's Zeus and Hera, but Zeus can always be in a particular state that restricts Hera from being in some of her particular states. Interesting. In that sense, she does not have free will of him."
},
{
"end_time": 1895.179,
"index": 73,
"start_time": 1865.896,
"text": " At one particular time, one particular state of Zeus, it's just not going to be any possible world line in any of the universe, in our universe, in which Hera is in one of her particular states. There's some limitation. My being in one particular state at one particular time causes a restriction on the possible states that you could be in at that time. And if that's true, then we have no free will. Have you heard of Norton's dome?"
},
{
"end_time": 1919.735,
"index": 74,
"start_time": 1895.981,
"text": " Norton's dome. I think I did a while ago, ringing a bell, but I can't bring it up. Sure, it's a thought experiment about Newtonian mechanics and it's to show that Newtonian mechanics isn't deterministic, even though it's often said it is, and the reason is there are certain configurations you can set up such that there's not a unique answer to the differential equations."
},
{
"end_time": 1948.677,
"index": 75,
"start_time": 1919.735,
"text": " You know, ordinarily in physics, just for people to know, one of the reasons why mathematicians quibble with physicists is that physicists hand wave and gloss over many details. And so one of them is whenever we have an ordinary differential equation, we tend to say there's uniqueness in existence. However, that's contingent on something called the Lipschitz continuity. And if you don't have Lipschitz continuity, you don't necessarily have a unique solution. So you basically set up a certain situation with a ball on a dome, and the equation for the dome is fairly simple. It's almost like a parabola."
},
{
"end_time": 1977.363,
"index": 76,
"start_time": 1949.002,
"text": " And then it turns out one solution is it stays there forever, the zero velocity initially. And then another solution is at some point t and the time t is not specified. It goes down some route and any one of them. So that's extremely interesting. Let's imagine we live in a Newtonian world. Is that related to free will? Would you say, or is that not related to free will? That's something different. I know that free will forget about the sense of intention interacting with the laws of nature to produce that effect. Yeah. The Norton's dome, it's also, I think it was actually Sabine."
},
{
"end_time": 2002.91,
"index": 77,
"start_time": 1978.012,
"text": " who has one of her FQXI essays that she points out that chaos is, in some sense, it can be a much stronger phenomenon than people understand, and that you can set up physical systems in which the chaos is to such a degree that actually passed a certain point in time, it is not defined, but the status system will be after that. So I think that's the context in which I ran across it."
},
{
"end_time": 2026.596,
"index": 78,
"start_time": 2003.387,
"text": " It's also their related things, the work that goes back to two people called Porel and Richards. Those were physicists who is well known that, for example, the three body problem, where even if you do, so there you do have the standalicious continuity and so on, you've got gravitational traction and so on. You can set it up to be a, what's called a universal Turing machine."
},
{
"end_time": 2052.159,
"index": 79,
"start_time": 2027.449,
"text": " you basically what you do is you encode the input tape to that Turing machine into the actual precise initial conditions of these three bodies. And then by reading the appropriate bits of the state of the system at some future time, you can figure out what that universal Turing machine state of its tape would be at that time."
},
{
"end_time": 2078.729,
"index": 80,
"start_time": 2053.387,
"text": " What this means is that you can feed in a configuration that's actually the halting problem, so that that physical system, in fact, it violates the Church-Turing thesis. That physical system, it's a state in the future would not be computable. So it's a very closely related kind of result. The impossibility results that I'm describing are stronger in that they don't presuppose differential equations."
},
{
"end_time": 2105.811,
"index": 81,
"start_time": 2080.316,
"text": " they don't suppose any of that kind of lawful structure whatsoever. So that's the sense in which they would hold no matter what the laws of physics are. And that is for that reason as well, that they encompass any possible supernatural phenomenology, impact on the real world, no matter how you might even think about it,"
},
{
"end_time": 2134.889,
"index": 82,
"start_time": 2106.254,
"text": " because they're just subsets of the set of all possible world lines, and world lines, in this sense, they're just themselves are a set of events in a completely arbitrary sense of what the word event means. It doesn't even presuppose space-time. Really, the mathematics just says, here is a set, I'm going to call the elements of the set world lines, because that's what they happen to correspond to in our kind of world,"
},
{
"end_time": 2164.514,
"index": 83,
"start_time": 2135.23,
"text": " But there are things. And everything that we as humans would call rules and laws concerning them has to do with things like observation operators on them and projections and things like that. But that's all extra structure, just the topology of it. The fact that you have these interlocking subsets, that itself has many consequences. One of which is the monotheism theorem. You could maybe have a god, but you can't have more than one."
},
{
"end_time": 2195.589,
"index": 84,
"start_time": 2165.981,
"text": " I play a bunch of fast and loose games when I describe this result. On the one hand, almost to protect myself in intellectual discourse, I want to say that I'm phrasing that theorem very much tongue-in-cheek, and there's an element of truth to that. But without making a firm"
},
{
"end_time": 2218.78,
"index": 85,
"start_time": 2196.237,
"text": " declaration, I would say quite honestly that this does prove that by any notion of a deity in any of the world's religions and any philosopher might ever have come up with, no, you can't have more than one. This actually does prove monotheism. So if you push me to the wall, I'm going to go tongue-in-cheek and say, well, I won't be using my definitions."
},
{
"end_time": 2245.452,
"index": 86,
"start_time": 2219.275,
"text": " But there's a large part of me that's thinking that no, in fact, this does prove that you can't have more, you can't, polytheism is ruled out. Sorry. This sounds like a great time to talk about what an observer is. So an observer in this sense is basically an inference device. If we're talking, so an observer, you can be thinking about there's many different frameworks. The way that I, how to phrase this. Okay. This is a digression. Um,"
},
{
"end_time": 2269.019,
"index": 87,
"start_time": 2246.152,
"text": " or a diversion. I'll come back. I'm circling around. So when I do my own personal philosophizing, if that's the word, when I construct, configure my own internal... Hear that sound?"
},
{
"end_time": 2296.084,
"index": 88,
"start_time": 2269.94,
"text": " That's the sweet sound of success with Shopify. Shopify is the all-encompassing commerce platform that's with you from the first flicker of an idea to the moment you realize you're running a global enterprise. Whether it's handcrafted jewelry or high-tech gadgets, Shopify supports you at every point of sale, both online and in person. They streamline the process with the Internet's best converting checkout, making it 36% more effective than other leading platforms."
},
{
"end_time": 2324.548,
"index": 89,
"start_time": 2296.084,
"text": " There's also something called Shopify Magic, your AI powered assistant that's like an all-star team member working tirelessly behind the scenes. What I find fascinating about Shopify is how it scales with your ambition. No matter how big you want to grow, Shopify gives you everything you need to take control and take your business to the next level. Join the ranks of businesses in 175 countries that have made Shopify the backbone of their commerce. Shopify, by the way,"
},
{
"end_time": 2353.473,
"index": 90,
"start_time": 2324.548,
"text": " powers 10% of all e-commerce in the United States, including huge names like Allbirds, Rothy's, and Brooklinen. If you ever need help, their award-winning support is like having a mentor that's just a click away. Now, are you ready to start your own success story? Sign up for a $1 per month trial period at shopify.com slash theories, all lowercase. Go to shopify.com slash theories now to grow your business no matter what stage you're in."
},
{
"end_time": 2376.817,
"index": 91,
"start_time": 2353.695,
"text": " Razor blades are like diving boards. The longer the board, the more the wobble, the more the wobble, the more nicks, cuts, scrapes. A bad shave isn't a blade problem, it's an extension problem. Henson is a family-owned aerospace parts manufacturer that's made parts for the International Space Station and the Mars Rover."
},
{
"end_time": 2398.626,
"index": 92,
"start_time": 2376.817,
"text": " Now they're bringing that precision engineering to your shaving experience. By using aerospace-grade CNC machines, Henson makes razors that extend less than the thickness of a human hair. The razor also has built-in channels that evacuates hair and cream, which make clogging virtually impossible. Henson Shaving wants to produce the best razors, not the best razor business,"
},
{
"end_time": 2418.626,
"index": 93,
"start_time": 2398.626,
"text": " So that means no plastics, no subscriptions, no proprietary blades and no planned obsolescence. It's also extremely affordable. The Henson razor works with the standard dual edge blades that give you that old school shave with the benefits of this new school tech. It's time to say no to subscriptions and yes to a razor that'll last you a lifetime."
},
{
"end_time": 2446.664,
"index": 94,
"start_time": 2418.626,
"text": " Dwelling that's got to do with that I call the philosophy house, so to speak."
},
{
"end_time": 2475.265,
"index": 95,
"start_time": 2447.602,
"text": " What it really amounts to is the following, that there are different, I'm almost exclusively interested in metaphysics. I frankly couldn't care less about philosophy of art, for example. I think it's an oxymoron. Philosophy of morality, I think Harsany's got some interesting things to say. Rawls, less so. Singer, I think has some interesting things to say, but not all that much more than that."
},
{
"end_time": 2499.787,
"index": 96,
"start_time": 2475.828,
"text": " But so mostly it's metaphysics, in other words, what is reality and so on and so forth. And basically, as you strip and what philosophy house is built upon is the understanding that you can do certain investigations where you make certain assumptions. Then when you weaken those assumptions, a lot of it falls apart and you get to a deeper level of reality where there's less structure."
},
{
"end_time": 2529.582,
"index": 97,
"start_time": 2500.657,
"text": " Weaken assumptions and you keep going down and down and down until the whole thing actually eats itself. So up at one level, I'm more than happy to be a steadfast card carrying member of the school of people who call themselves scientists, scientific realists, whatever you might want to call it. Then at another level, you can start to say, well, what is reality? You can't even define it. And"
},
{
"end_time": 2554.514,
"index": 98,
"start_time": 2530.111,
"text": " You start to then start worrying about things like inference devices. You can then say, wait a second, combine what you just said about no reality is more real than any other, that all of them are just different, these subsets of possible universes, that you can find these inference devices structures, that's all that any reality is, that none is a privilege over any other."
},
{
"end_time": 2583.951,
"index": 99,
"start_time": 2554.974,
"text": " Doesn't that also, can't you also then apply that to your own reasoning, your own question answer process, and conclude that that itself is going to be illegitimate in a sort of a false demons kind of sense. So you've just concluded that no reality is privileged above any other. But in doing so, if you want to strip down the very last remaining assumption, you are not allowed to assume that your reasoning process itself is somehow proof against that kind of conclusion."
},
{
"end_time": 2611.288,
"index": 100,
"start_time": 2584.582,
"text": " And so, as was put in, I think, as one of the Hitchhiker's Guides, little episodes, that everything vanishes in a proof of logic. And that's where you would ultimately end up going. So the reason I mention that is when you say the word observation, it can mean very different things depending on where we are in this whole edifice"
},
{
"end_time": 2641.493,
"index": 101,
"start_time": 2611.664,
"text": " this philosophy house where we're stripping out assumptions and so on. In the context of inference devices, observation is just like prediction, control, or memory, for that matter. It's just a way for you to be posing questions. You then have an observation device that is providing you answers and what the things like the Laplace's demon theorem and all these kinds of things prove to you is that no matter what the observation device is,"
},
{
"end_time": 2668.763,
"index": 102,
"start_time": 2641.766,
"text": " There are always going to be situations where the answer it gives you is wrong. So that's what observation means in the context of inference devices. It means other things in other contexts. But in the structure of inference devices, note that prediction and memory are identical. So this is something that has to do with the block model of the universe that we think we were talking about last time."
},
{
"end_time": 2696.869,
"index": 103,
"start_time": 2669.326,
"text": " the second law and the psychological arrow of time. Memory is just retrodiction. It's a prediction about the past rather than the future that we suppose, for reasons having to do with the second law, has to be much more accurate, at least some of our memory, than our predictions. Why do you say that it's a prediction about the past rather than a statement about the past? Because we have no way of knowing it's true. It's identical."
},
{
"end_time": 2726.8,
"index": 104,
"start_time": 2697.21,
"text": " There is no privileged arrow of time. This is something that physicists have been wrestling with, especially cosmologists. I think I mentioned last time there's a 1973 book by Paul Davies, The Physics of Time Asymmetry, and Sean Carroll's got some nice stuff in this. And there's, I have a paper in, I think it was 1992 in the International Journal of Theoretical Physics on these kinds of things. There is no sense in which the past is more real than the future."
},
{
"end_time": 2755.589,
"index": 105,
"start_time": 2727.637,
"text": " All moments are just as real as one another, and all that we are doing as scientists or as people remembering is making predictions in the statistic sense. Despite that prefix, the word predediction, from a scientific point of view, from the point of view of the laws of physics, there's no difference between a prediction and a retradiction. They are both using"
},
{
"end_time": 2784.036,
"index": 106,
"start_time": 2756.203,
"text": " data at the present to make a statistical estimate of the state of a variable at a time other than the present. It's a retradiction. You could be wrong. And in fact, that's, that's, of course, a whole nother body of work and things we could discuss about is how in fact the mind very often is wrong. And there's implanted memories and just generally fallibility of memory and all this business about how consciousness is a"
},
{
"end_time": 2812.261,
"index": 107,
"start_time": 2784.582,
"text": " just so story that we concoct to try to give us the illusion that we're actually in control, that we have some memory, accurate memory of the past, and that both are actually erroneous. But so anyway, memory is just retrodiction, and so it is as prone subject to all these impossibility results as any of the inference device. So what that means in particular is the plus was only thinking forward in time."
},
{
"end_time": 2840.879,
"index": 108,
"start_time": 2813.439,
"text": " And so he wouldn't even think there was an issue of having to construct a demon that would actually predict the past. He would just simply say, you remember the past. And so, sure, your memories are 100% accurate. In point of fact, no. According to these impossibility results, for the same reason that Laplace cannot predict the future with 100% accuracy, it's impossible to build a 100% accurate memory device"
},
{
"end_time": 2869.189,
"index": 109,
"start_time": 2841.834,
"text": " Is there a relationship between the no free lunch theorem and the limits on inference devices? Very good question. So I have found, somewhat to my bemusement, that a lot of the stuff that I am drawn to investigate,"
},
{
"end_time": 2898.439,
"index": 110,
"start_time": 2869.599,
"text": " The things that I chew on have to do with impossibilities. No free lunch is one of them. Inference devices is another one. As I think I mentioned to you, the notion that mathematics itself might be inherently stochastic, which as you offline, we discussed about how that's related to intuitionism. Some of the ideas of Nicholas Gisson and intuitionism, of course, traces back to Brouwer, a great mathematician in the 20th century."
},
{
"end_time": 2927.005,
"index": 111,
"start_time": 2899.104,
"text": " Um, and I think that there are some potentially some impossibility results there as well. But all of these I suspect have to be related to one another. Um, I've That's on my I've got everybody's got to-do lists and that's on one of my longer term greatest size font, so to speak, to-do lists is figure out"
},
{
"end_time": 2954.07,
"index": 112,
"start_time": 2927.619,
"text": " how to integrate these impossibility results. There's also ones that I haven't come, the ones that I came up with, there are the impossibility results that are involved with, that have to do with Turing machines and girdle and so on. You think they're all related? Well, for example, I'm doing some work at a very, very low burner. I wish I could do it more intensely on trying to"
},
{
"end_time": 2982.995,
"index": 113,
"start_time": 2955.691,
"text": " You exploit the church touring thesis to be able to say that within the context of inference devices, to say that if there is a universal touring machine, it's going to be inference device. And can the impossibility results concerning inference devices, do they, for example, tell us something like you can't have two perfectly accurate universal touring machines?"
},
{
"end_time": 3011.118,
"index": 114,
"start_time": 2983.729,
"text": " Do they provide another way to prove the halting theorem? And so that's on my to-do list to see if they might be integrated that way. Another interesting thing is that the inference device's structure that I was describing to you had no probability distributions over it. We were just talking about subsets of the set of all possible world lines. Well, let's slap down a probability distribution."
},
{
"end_time": 3036.988,
"index": 115,
"start_time": 3012.039,
"text": " we can now modify some of our definitions of what it means to infer something accurately to be to infer with a higher than certain probability of being correct. We can then say well over all possible probability distributions, so what that means is that rather than the impossibility results of an inference device that concern for example two deities, instead"
},
{
"end_time": 3059.326,
"index": 116,
"start_time": 3037.295,
"text": " What you start talking about the probability distribution of all the worst case probability that both deities could be correct, or what is the best case probability that they could be correct over all possible probability distributions. So rather than saying it's impossible for them to both always be correct. Overall, interesting. Yeah."
},
{
"end_time": 3083.08,
"index": 117,
"start_time": 3059.753,
"text": " And guess what? You actually get some results out of that that look a little bit like the Heisenberg uncertainty principle. Is this published work? I think it's mentioned in one of my papers, but it's both that and the Turing machine stuff. Very tentative. There's a lot more that needs to be done and that I want to get to at some point. I mean, there's this meta-optimization problem for"
},
{
"end_time": 3110.384,
"index": 118,
"start_time": 3083.626,
"text": " people who are so very, very lucky, like as me, that society provides us the resources to go off and, well, play is the way it feels like to me. But for us, for people who are so, I wouldn't use the word blessed. I mean, I'm in the, fortunate, but fortunate. Yes. And people who are so fortunate, there's this optimization question of, well, there's all these things that you do want to work on."
},
{
"end_time": 3135.247,
"index": 119,
"start_time": 3111.067,
"text": " You've only got some finite number of years left. You don't even know what they are. Prioritize, dude, or do that, as the case might be. And so I'm not sure. So at some point, yes, I do want to work on these things. But the question is when and how and how to get the time. And if you can do it with collaborators, which I try to do, and so on. And it's one of the challenges of being somebody who's as fortunate as I am, is how best to use that fortune."
},
{
"end_time": 3158.029,
"index": 120,
"start_time": 3135.503,
"text": " Have you heard of Donald Hoffman? No, I don't think so. Okay, Donald Hoffman has an argument that says whatever we see is not reality as it is. Oh, yes, yes, yes. Um, I don't think I drilled into it. I don't think really amounts to that much. I'm to be quite honest. Yes, let's hear why. Um, I have to hear that sound."
},
{
"end_time": 3185.009,
"index": 121,
"start_time": 3158.916,
"text": " That's the sweet sound of success with Shopify. Shopify is the all-encompassing commerce platform that's with you from the first flicker of an idea to the moment you realize you're running a global enterprise. Whether it's handcrafted jewelry or high-tech gadgets, Shopify supports you at every point of sale, both online and in person. They streamline the process with the internet's best converting checkout, making it 36% more effective than other leading platforms."
},
{
"end_time": 3204.889,
"index": 122,
"start_time": 3185.009,
"text": " There's also something called Shopify Magic, your AI powered assistant that's like an all-star team member working tirelessly behind the scenes. What I find fascinating about Shopify is how it scales with your ambition. No matter how big you want to grow, Shopify gives you everything you need to take control and take your business to the next level."
},
{
"end_time": 3234.497,
"index": 123,
"start_time": 3204.889,
"text": " Join the ranks of businesses in 175 countries that have made Shopify the backbone of their commerce. Shopify, by the way, powers 10% of all e-commerce in the United States, including huge names like Allbirds, Rothy's, and Brooklynin. If you ever need help, their award-winning support is like having a mentor that's just a click away. Now, are you ready to start your own success story? Sign up for a $1 per month trial period at Shopify.com"
},
{
"end_time": 3256.903,
"index": 124,
"start_time": 3234.497,
"text": " Think Verizon, the best 5G network is expensive? Think again. Bring in your AT&T or T-Mobile bill to a Verizon store today and we'll give you a better deal. Now what to do with your unwanted bills? Ever seen an origami version of the Miami Bull?"
},
{
"end_time": 3286.971,
"index": 125,
"start_time": 3257.363,
"text": " Jokes aside, Verizon has the most ways to save on phones and plans where you can get a single line with everything you need. So bring in your bill to your local Miami Verizon store today and we'll give you a better deal. Go back and review my notes. He seemed to be... I'm sorry."
},
{
"end_time": 3315.794,
"index": 126,
"start_time": 3287.944,
"text": " So I've actually got, I'm pulling it up right now, for example, my annotated version of his work. Yeah, and there's just too much. If you want to have a splice in from a third podcast, I could research it, but I don't remember why. That's one of the other challenges of being fortunate"
},
{
"end_time": 3339.531,
"index": 127,
"start_time": 3316.22,
"text": " is you have to very quickly come to decisions about whether it makes sense for you to be spending any effort drilling down on things. So yeah, I did look into his stuff and I wasn't very impressed. If there's one particular article of his that you think I should maybe look at rather than whatever I did look at, if you could just email that to me, I'll see if there's anything different about that one. Sure. That's a quick aside."
},
{
"end_time": 3367.892,
"index": 128,
"start_time": 3339.991,
"text": " My background's not computer science, as you know. Whenever I've skimmed the church-turing thesis, it seems more like a definition of what it means to compute rather than a theorem per se. It's very close to that. It's very close to that. There's a question of whether it's ultimately vacuous or not. There's something called the physical church-turing thesis, where instead people say things like you cannot build a physical computer that is"
},
{
"end_time": 3394.735,
"index": 129,
"start_time": 3368.217,
"text": " that implements algorithms that can come to conclusions that would not be possible to do on a physical Turing machine. And it's squirrely to even define it there. There are different definitions you can find in the literature. Scott Aronson actually has a nice video, it's like two hours long, where he talks about different challenges to the physical church Turing thesis."
},
{
"end_time": 3423.558,
"index": 130,
"start_time": 3395.265,
"text": " One of the reasons, as I recall, that David Deutsch first started working in quantum computation, his original paper, was to try to say that, well, if the physical Church-Turing thesis can be either fully ironclad formalized and or established as actually a law of nature, that means we have to be able to express computation in purely quantum mechanical systems, because quantum mechanics is the law of nature. So that was a large part, as I understand it."
},
{
"end_time": 3450.998,
"index": 131,
"start_time": 3424.087,
"text": " of his impetus for starting to work on what now has become called quantum computation. But yes, you're correct that the Church-Turing thesis, as it was understood, say in the middle of the 20th century, seemed very dangerously close to simply being a definition rather than actually a statement about the physical world. So before we get to the audience questions, you mentioned in one of your talks... By the way, during the break,"
},
{
"end_time": 3480.213,
"index": 132,
"start_time": 3451.374,
"text": " I did quickly remind myself about Professor Hoffman. OK. And so what I understand, my take on his work, so in a nutshell, he's a neuroscientist. And in an extension of what we were talking about earlier, a vast elaboration of the notion that consciousness and so on and so forth, it's all a fable."
},
{
"end_time": 3506.8,
"index": 133,
"start_time": 3480.401,
"text": " He has been doing a lot of work showing that, in particular, our understanding of the external world is completely concocted, that it's whatever our sensory apparatus is going to think is actually most effective. Our sensory apparatus is going to be telling us precisely whatever natural selection finds out by trial and error"
},
{
"end_time": 3535.299,
"index": 134,
"start_time": 3507.363,
"text": " is the best thing to tell us to get us to act in an efficient way as far as our reproductive success is concerned. That which it is telling us to do that might have nothing really to do with any kind of an external physical reality. It's kind of an elaboration of all the aspects that what you think you are seeing, for example, you are not. It's all being filled in with what's called predictive coding in its broadest sense in the neuroscience community."
},
{
"end_time": 3560.282,
"index": 135,
"start_time": 3535.759,
"text": " and so my understanding of his work was that he has done some interesting things related to that particular scientific issue but then sort of well frankly puffs it up as to saying something about there is no objective reality and so on and so forth it's no rather our human being fallible perception of reality"
},
{
"end_time": 3571.476,
"index": 136,
"start_time": 3561.049,
"text": " It can be very, very different. A natural selection, in fact, you would say, which should make it be very, very different from what we would, from what an actual reality is."
},
{
"end_time": 3602.142,
"index": 137,
"start_time": 3572.568,
"text": " Okay, now before we get to some audience, quite and just so you know, as an aside, I don't believe Donald Hoffman, although maybe you have a different paper, I don't believe he indicates that there is no reality. He does say what you said, which is that our relationship to that reality, because the way that he constructs reality, it looks strange, the way that it's like three agents interacting with one another, etc. But that's the true game. Yeah, yeah. That was just based on a quick skim reminding myself from I have to have to admit is a very pedestrian source, which is quantum magazine."
},
{
"end_time": 3629.735,
"index": 138,
"start_time": 3602.517,
"text": " That's okay. And one of your talks, I think in one of your FQXI talks, you mentioned that some of the most profound results in philosophy come from algorithmic information theory. Yeah. And coincidentally, I'll be speaking to Gregory Chaitin in a few weeks. Oh, okay. Please explain to me why you think that. Okay, I can use one of Gregory Chaitin. I mean, he does insist that it be pronounced Chaitin rather than Chaitin."
},
{
"end_time": 3658.78,
"index": 139,
"start_time": 3630.282,
"text": " He's a very interesting fellow in many ways, and he's done some great work. One of his results, in fact, I think, does have this character to it. He's got an impossibility theorem, which is basically it says that there is a integer l such that we can never, ever prove that any particular"
},
{
"end_time": 3689.531,
"index": 140,
"start_time": 3659.735,
"text": " computation that we would want to do. Any particular string has comagural complexity greater than L. Okay, there is some such integer and he actually has an estimate of it as I believe based upon Lisp programs of being 4,000 if you're using Lisp as your Turing machine. But think about what that means. There's going to be this number like an L which is like say 4,000 that's saying that you know anything that's longer than 4,000 bytes"
},
{
"end_time": 3716.476,
"index": 141,
"start_time": 3690.162,
"text": " So you cannot prove that the Komogoro complexity of any particular string is going to be greater than 4,000 bytes. We know that there's a very finite number of strings that have less than 4,000 bytes. At most, depending on the grammar and syntax of your language and so on, it would be 2 to the 4,000, which would be a big number, but it's very, very finite."
},
{
"end_time": 3742.841,
"index": 142,
"start_time": 3716.954,
"text": " So that means that there's an infinite number of strings that exist whose comagural complexity we can never actually determine. So we can never know what is the simplest program size for determine what the simplest program for actually doing those calculations. It's an amazing restriction on what we human beings can do."
},
{
"end_time": 3769.94,
"index": 143,
"start_time": 3744.104,
"text": " To give you another example, this is one that I know best from my book by Lee and Vitani. It's kind of like the Bible and these Turing machine things. You can actually prove the existence of a function from the integers to the integers, which is always increasing. Sometimes it'll stay the same, but it never goes down and eventually gets to infinity."
},
{
"end_time": 3799.002,
"index": 144,
"start_time": 3771.493,
"text": " such that every function you can possibly compute, no matter how you do it, that is also always increasing and gets to infinity, will be strictly greater than this. You wouldn't have even thought there is such a limitation that could even make sense. Do whatever you want. Say, okay, here's a function which it has the value one for the first million numbers, from one through a million, it's got the value one."
},
{
"end_time": 3826.34,
"index": 145,
"start_time": 3799.445,
"text": " Then it's got the value 2, but that's for the next million to the millionth. Make it be whatever you want. Any rule that you can put down for how to construct this, it's going to actually be getting to infinity faster than this other function, which is a very strange thing. It's one of the most fundamental philosophical results in the sense that philosophy"
},
{
"end_time": 3853.951,
"index": 146,
"start_time": 3827.005,
"text": " should not be biased towards what we human beings consider to be compelling. I find it that impossibility results in general, and of these sorts in particular, they are very deep philosophy, whether or not we even appreciate they had meaning before we came across them. What was the name of that one, that second one? Oh, this is, I don't even think it has a name."
},
{
"end_time": 3883.08,
"index": 147,
"start_time": 3854.701,
"text": " I don't even think it has a name. I can point you to the chapter and leave. Yeah, that would be great. But there's another one that, for example, Scott Aronson has a nice blog post on this. It's called the busy beaver function. And he's actually recently written some papers on it as well. And this is kind of the flip side of what I just said, that let's try to make the fastest increasing function possible. So you, Kurt,"
},
{
"end_time": 3913.473,
"index": 148,
"start_time": 3883.643,
"text": " Say that, well, for the value 1, it's got the value 1. For the value 2, it's got the value 10 to the 10 to the 10. For the value 3, it's got what it had for the value 2, but now itself. Make it be whatever you want. There is always going to be something which is called the busy beaver function, which is actually increasing faster than the fastest increasing function you can write down. In other words, there's an upper limit to what"
},
{
"end_time": 3942.022,
"index": 149,
"start_time": 3913.865,
"text": " There's an upper limit to the speed of increase of any function from the integers to the integers that you can possibly define. It's in a certain sense, it's mind-boggling that there's that kind of a limitation on what we can do. I'm not understanding it correctly. So at first, the way to understand what you said is that you have a function that's increasing and then you can make it increase faster, but then there's a bound? There's a bound. I can write down the definition of a function"
},
{
"end_time": 3966.681,
"index": 150,
"start_time": 3942.346,
"text": " And it exists, I can prove it exists, and it is an increasing function, and it will be increasing faster than any function that you can possibly write down. Now, I have a question about that. Like I mentioned, I'm speaking to someone who's an ultrafinitist, an intuitionist, and they don't particularly like existence proofs. They like construction proofs. Yep, and that is the foundation of intuitionism. I don't think that actually Nicholas Gibson"
},
{
"end_time": 3997.193,
"index": 151,
"start_time": 3967.312,
"text": " that goes that far, though I'm not sure he might in some side idea one of his papers. But yeah, that was the foundation. They don't like existence proofs. They want a constructive proof only. Is this constructed or you just showed the existence of this function? Oh, the busy beaver function. It has been constructed for the first some number of values of what of the integer. So, you know, we can write down a busy beaver function for one, for two, for three, for four, and so on."
},
{
"end_time": 4026.8,
"index": 152,
"start_time": 3997.654,
"text": " up to some particular value. But I think that in the Busy Beaver function, its entirety, almost by definition, no, it can't be constructed. Because if it could be constructed, you would construct it. That's the point, that this thing exists. Wait, sorry, I don't get that last point. Wait, if it could be constructed, I could construct it, what do you mean? So the definition, so if by construct a function, we mean you write down a program,"
},
{
"end_time": 4053.507,
"index": 153,
"start_time": 4027.312,
"text": " that spits out its values. That's what we mean to construct a function. Then the busy beaver function is something that increases faster than could be the output of any such program you can write down. It's very, very easy to define. It is purely an existence function that you can't construct it, but I can define it very, very easily."
},
{
"end_time": 4083.422,
"index": 154,
"start_time": 4054.974,
"text": " It takes only a couple of sentences to define what the thing is. So there are these kinds of results which I find in many ways flabbergasting because if one adopts the Church-Turing thesis, these are limitations on human thinking. And there are other ones, and this is in addition to all the ones like the halting problem, you know, halting theorem."
},
{
"end_time": 4104.292,
"index": 155,
"start_time": 4083.814,
"text": " which is very closely related to, of course, some girl's incompleteness theorem. And then there's Rice's theorem and all these other things, which really, I think that the only actual results hear that sound."
},
{
"end_time": 4131.408,
"index": 156,
"start_time": 4105.265,
"text": " That's the sweet sound of success with Shopify. Shopify is the all-encompassing commerce platform that's with you from the first flicker of an idea to the moment you realize you're running a global enterprise. Whether it's handcrafted jewelry or high-tech gadgets, Shopify supports you at every point of sale, both online and in person. They streamline the process with the internet's best converting checkout, making it 36% more effective than other leading platforms."
},
{
"end_time": 4150.333,
"index": 157,
"start_time": 4131.408,
"text": " There's also something called Shopify Magic, your AI powered assistant that's like an all-star team member working tirelessly behind the scenes. What I find fascinating about Shopify is how it scales with your ambition. No matter how big you want to grow, Shopify gives you everything you need to take control and take your business to the"
},
{
"end_time": 4176.152,
"index": 158,
"start_time": 4150.333,
"text": " Join the ranks of businesses in 175 countries that have made Shopify the backbone of their commerce. Shopify, by the way, powers 10% of all e-commerce in the United States, including huge names like Allbirds, Rothy's, and Brooklynin. If you ever need help, their award-winning support is like having a mentor that's just a click away. Now, are you ready to start your own success story?"
},
{
"end_time": 4203.882,
"index": 159,
"start_time": 4176.152,
"text": " The actual philosophical advances that have been made by humanity are these kinds of results. Everything else is not only questionable,"
},
{
"end_time": 4229.445,
"index": 160,
"start_time": 4204.411,
"text": " and is being questioned, it is ultimately going to be quicksand and it's just going to be squishy and you're not going to get anywhere. These are the only ones that we have managed to generate so far. These are it folks. By definition, they are the deepest because they are the full set of things that we have actually established incontrovertibly or as incontrovertible as any deductive logic could be for mathematics."
},
{
"end_time": 4251.51,
"index": 161,
"start_time": 4229.991,
"text": " Okay, I'm going to send you a chat right now. Let me know if you can pull it up on your end, or if it automatically gets sent to you. See there? Well, it's a big one. Okay. Yeah. Okay. So the reason I'm sending it to you is because I'm probably going to miss read this. And it's easier if you read along with me. Oh, these are questions from the audience. Yes, this is not live. Obviously, I'm just saying this before. So this question"
},
{
"end_time": 4280.691,
"index": 162,
"start_time": 4252.312,
"text": " We're now we're getting on to some audience questions and this comes from Chris Langan. Let's get to question number two, which says he prior to this was was referencing how in your definition of observer, there's a theological aspect because an observer wants to preserve its own self organization properties, self organization required in the definition of observer. It doesn't, there's nothing in there about what your goal is. There's nothing in there about intentionality in any of the inference devices work."
},
{
"end_time": 4308.456,
"index": 163,
"start_time": 4281.493,
"text": " It's just, you could be observing by luck. So it includes the case where your predictions are correct by pure luck. Or maybe even worse, you thought that you were answering a different question from the real one, but you were screwed up in your thinking, you happened to accidentally get the answers right. Well, no, impossibility results. So you can't even do it that way. You can't get there by averted vision, so to speak. So there's nothing about what you mean."
},
{
"end_time": 4335.759,
"index": 164,
"start_time": 4308.968,
"text": " Then let's do number four. Your work straddles the boundary between mind and matter. What about Cartesian mind slash matter dualism, which says that mind and matter are fundamentally different. Do you agree or do you believe they're both aspects of? I don't see there's anything useful to be had with the Cartesian mind matter dualist perspective. Remember what we were talking about before about"
},
{
"end_time": 4359.65,
"index": 165,
"start_time": 4336.254,
"text": " Let me just complete the sentence because the audience doesn't have access. Do you agree or do you believe that they are both aspects of a single underlying reality? Okay, I didn't see that second part in the chat. Mind and matter. Okay, so it's got to a degree, I see. It involves a little part, a little bit where we are in that philosophy house."
},
{
"end_time": 4388.968,
"index": 166,
"start_time": 4360.401,
"text": " at what level we've gone to. And I'm sitting there with all the assumptions about the sort of what I could very loosely and inaccurately describe as scientific realism, then there is zero reason to believe that there's anything in mind that does not matter. And in fact, even most, even people that I think are frankly wrong about many things like David Chalmers worrying about consciousness means and qualia and sorry. Penrose as well. Penrose as well, when he talks about the mind."
},
{
"end_time": 4417.756,
"index": 167,
"start_time": 4389.616,
"text": " Even these people will almost uniformly and almost in total accept that mind is ultimately matter. So that this is part of why I find qualia to ultimately be a vacuous concept is that if you do accept that, I don't see where there's really any wiggle room"
},
{
"end_time": 4444.48,
"index": 168,
"start_time": 4418.148,
"text": " to worry about, well, what are your subjective impressions then I'm on to, which is essentially what the quote hard question of consciousness is about. So when I'm sitting up there in my scientific realism room of the house, a mind is just a attribute of matter for the reasons that Hoffman elucidates, Libet elucidates, Dan Dennett has some great work, all the zombie within stuff,"
},
{
"end_time": 4473.933,
"index": 169,
"start_time": 4444.94,
"text": " what mind thinks it knows about matter actually is in many, many ways completely fallacious. Then when I drill down below that, of course, you start running into things like solipsism, brain in a vat. Brain in a vat is a cutesy little thing that freshmen debate and has kind of been ignored in the academy. Nobody has got an answer to that challenge because you can't have one. Instead, what do you do is you assume that no, let's assume that we're not a brain in a vat."
},
{
"end_time": 4500.435,
"index": 170,
"start_time": 4474.428,
"text": " And that's the basis for most of metaphysics because you can't go beyond that. But if you were to not make that assumption, if you, as I mentioned before, you go through these various rooms in the house of philosophy by throwing out assumptions, if you don't make the a priori assumption that what we are perceiving is somehow coupled to a reality that we are more than a brain in the bat or matrix, if you wish to view it that way or whatever,"
},
{
"end_time": 4530.742,
"index": 171,
"start_time": 4500.998,
"text": " then you're stuck right there. And the question of is there a duality between mind and matter? If we're going to allow that I'm a brain in a vat, then that question takes on completely different aspects to it, that mind becomes matter in a different sense. And you can drill it down even further and start saying that, well, in fact, and I think this is where ultimately the quality of people, the heart problem of consciousness people are coming from, that ultimately mind"
},
{
"end_time": 4559.735,
"index": 172,
"start_time": 4530.947,
"text": " comes first and whether there is even matter or not is a TBD. Or to put it in another way, I could define a universe in the sense of inference devices, which arguably is only mind. And there is no such thing as matter. So, you know, once we get down to the level of inference devices, where a universe is just a set of world lines, essentially, we're down at what's called sometimes ontic structural realism."
},
{
"end_time": 4589.48,
"index": 173,
"start_time": 4560.435,
"text": " We've actually gone even further than Max Tegmark's multiverses. And all that any reality amounts to is a pattern relating aspects of these world lines. Laws of the universe are just patterns, which could at any time be violated or not. And once you get to that level that's just patterns, just mathematics, that reality to a pawn on a chessboard is the laws of chess,"
},
{
"end_time": 4619.667,
"index": 174,
"start_time": 4589.889,
"text": " Because that's what its reality is. It's almost definitional. Then you have no such, there is no distinction between mind and matter or anything else. It's all just patterns at that level. There is no distinction. The distinction between mind and matter, they come to play with you, throw in other assumptions that put you in other rooms in the philosophy house. So the answer depends where you want to situate yourself and show which assumptions you adopt."
},
{
"end_time": 4650.486,
"index": 175,
"start_time": 4620.538,
"text": " This question is from Carl Friston of the Free Energy Principle. It is a compelling notion that systems extract information from their environment to maintain themselves far from equilibrium, but is this extraction cause or consequence? In other words, are non-equilibrium steady states a consequence of observing or making inferences about other systems?"
},
{
"end_time": 4679.701,
"index": 176,
"start_time": 4650.879,
"text": " This is touching on a huge bodywork, my primary body of research these days, which we've actually not discussed at all, which is what's called stochastic thermodynamics, non-equilibrium thermodynamics. Carl right there is, I think, without being sure, without this actually being a proper conversation, is he's trying to lead me into saying things which basically justify his notions of having to mark out blankets and so on and so forth."
},
{
"end_time": 4706.51,
"index": 177,
"start_time": 4680.401,
"text": " Instead, to make a plug for myself is a paper that I wrote with a postdoc Artemiy Kolchinsky, with the title have to do with semantic information and non-equilibrium physics, in which we make the case that without necessarily wanting to say that they are living systems, whether they are conscious systems, in a certain sense, interesting systems,"
},
{
"end_time": 4732.176,
"index": 178,
"start_time": 4707.108,
"text": " are those that are what are called non-equilibrium steady states, which is a word that Carl is using there, and that maintain themselves as being non-equilibrium steady states because of the information that they are actually getting from the environment, getting from it in a precise sense of conditional information flows. And I think that those are very interesting systems to"
},
{
"end_time": 4761.63,
"index": 179,
"start_time": 4733.029,
"text": " investigate, and that's a whole other body of work that I would like to investigate. But in terms of which came first, is this extraction cause or consequence? To me it's actually much more that there are systems which do have that extraction, and I would say it's neither cause nor consequence. Are non-equilibrium steady states a consequence of observing other systems?"
},
{
"end_time": 4789.872,
"index": 180,
"start_time": 4762.841,
"text": " And physicists can give a very simple answer. No, I can define non-Euclidean, but a way to make a non-Euclidean steady state is to simply hook up a system to heat baths at different temperatures. And it will go into what is called in physics a non-Euclidean steady state. Is it observing anything? You could say it's in a certain sense, quote, observing the two different temperatures, but that I think does basically expands the definition of the term observe, so it includes everything."
},
{
"end_time": 4820.145,
"index": 181,
"start_time": 4790.913,
"text": " It is interacting with its external world and there are aspects of it that are describable by information theory. And I think to use words like causal or observation in that context can be a very, very fraught thing. It's in danger of being a deepity, as Dan Dennett calls it. Okay, this next question comes from Professor Anil Seth of neuroscience."
},
{
"end_time": 4850.077,
"index": 182,
"start_time": 4820.572,
"text": " And he has a book out on consciousness, which I'll link in the description. He says, ask David whether he thinks consciousness is substrate independent. So a silicone computer can be in principle conscious. And if so, why? And then as a subnote, with regard to free will, I'm always interested if they referring to you, David and anyone else, if you, David realizes that the determinism versus stochastic debate is a total red herring. Well, let me respond to these."
},
{
"end_time": 4877.159,
"index": 183,
"start_time": 4850.52,
"text": " in order. I once asked Julio Tononi, according to his integrated conscious, his integrated information measure, which I am not a fan of, let me say that. But in any case, Scott Aronson is not either. And so he's got a blog post on it. But according to his measure, a computer, a silicon computer,"
},
{
"end_time": 4907.568,
"index": 184,
"start_time": 4877.91,
"text": " whether it implements its code in parallel or in serial determines whether it's conscious or not. I asked him that point blank and he said, yes, that determines whether it's conscious or not. I'm first responding to that aspect of my conversation with Julio before getting to Anil. By the way, I do think Anil does some very interesting work. But so my reaction to that is what is gained by defining"
},
{
"end_time": 4936.92,
"index": 185,
"start_time": 4908.251,
"text": " You can define words wherever you want to define it. That's one of the liberties of mathematics is to find things wherever you want. If you could define it so that Julio could say, I am by definition correct, that this is consciousness, my measure. OK, you can you can say that. And that's your definition. That's your definition. But getting back to what we were talking about last time, persuasive definitions. This also gets back to Searle with his Chinese rule argument, which I thought is completely vacuous."
},
{
"end_time": 4965.538,
"index": 186,
"start_time": 4937.432,
"text": " If you are to give a definition of consciousness which has any aspects to it which, if it is going to be restricted to being a characteristic of the behavior of a system, of the what rather than the how, then it cannot differ for a serial or parallel computer."
},
{
"end_time": 4994.599,
"index": 187,
"start_time": 4965.811,
"text": " And that's what Searle also missed. If consciousness, if thinking, whatever you want to call it, is a what rather than a how, and I think it's extremely liberating as a scientist to conceive of it that way, then it doesn't matter by definition. You just said no. The how is interesting, it's worthy of study, but it's a different question entirely from the what and how these are, and these should be disentangled. It is a category error to mistake the how for the what."
},
{
"end_time": 5024.753,
"index": 188,
"start_time": 4995.06,
"text": " Am I to understand that another way of saying that is that if it's a state versus a process? Or no, is that is that false? What I just said? Like, that's not the correct analogy. That's not what I mean to say. You right now, if behind those baby brown eyes, if behind your face were a silicon computer, that has no consequence to the behavior of our discussion. That's what touring great insight with the touring test is."
},
{
"end_time": 5054.275,
"index": 189,
"start_time": 5025.418,
"text": " And the importance of the Turing test is not whether you're using a teletype and things like that, it's to say that as far as he, Alan Turing was concerned, and there's a lot to be said for this, that it is useful to carve out from whatever you want to discuss about consciousness, it is very useful to carve out as its own phenomenon, call it consciousness prime, call it pre-conscious, whatever you want it to be, where we're just looking at the behavior of the system."
},
{
"end_time": 5080.708,
"index": 190,
"start_time": 5055.043,
"text": " And you can have things, one's in silicon, one is instead a real human being, one is instead a simulacrum that's being controlled by somebody outside the room looking in, wherever you want, that's behaving that same way. But if it's behaving the same way, you can't tell the difference. And so for this particular aspect of consciousness, it does not make any difference."
},
{
"end_time": 5110.026,
"index": 191,
"start_time": 5081.186,
"text": " So the Tanoni is actually looking at the other side, he's focusing, he's mixing up the two, I would say, the how and the what. So to me, the interesting thing is the what, you first have that, then once we have that what clarified, which we don't have, we don't know how to define it, even at a behavioral level, what a conscious entity would be, or so on and so forth, we can then start to look at the how. And that's where neurobiology comes to play, which is Anil's"
},
{
"end_time": 5132.073,
"index": 192,
"start_time": 5110.486,
"text": " particular comparative advantage, his specialty. And that's a completely open issue, frankly, as in one that he's doing work on involving things like predictive coding, blah, blah, blah, blah, is how do neurobiological systems achieve it? Silicon computers can achieve the same what they're doing in a different way. And so it's a different question."
},
{
"end_time": 5161.92,
"index": 193,
"start_time": 5132.688,
"text": " Then with regard to free will, I'm always interested in whether... Whether they think, slash realize, so that's obviously a loaded question, that the determinism versus stochasticity debate is a total red herring. To say there's a red herring, I mean, that's an extremely loaded leading question, and I'm going to assume that it does have that character because of basically word length, so to speak. A total red herring for what?"
},
{
"end_time": 5191.169,
"index": 194,
"start_time": 5162.073,
"text": " It's a question that cannot be answered. I would say that free will is commonly used in common discourse. This is what we were talking about yesterday, persuasive definitions, not yesterday, but last time. Because people commonly use free will, they're not even sure what they mean. To some people, it is determinism versus stochasticity. And that's what they mean. That's what they mean. To other people, it's not. Whether it's a total right herring has to do with how you define the concept in the first place."
},
{
"end_time": 5219.36,
"index": 195,
"start_time": 5191.988,
"text": " I don't view there as being a concept worth discussion. I view it as being pointless. Come up with your definition. Your definition is going to then say whatever you want your definition to mean, but it's your definition. That's why I'm saying we should call it what consciousness rather than how consciousness, just so we're clear in our definitions. I'm not going to say either one is consciousness. I'm going to instead just make two different definitions and say that when people use the word consciousness, they sometimes conflate things."
},
{
"end_time": 5249.497,
"index": 196,
"start_time": 5219.872,
"text": " It sounds like this question sets you off. Well, it does because it's such a waste of time. People are debating definitions."
},
{
"end_time": 5278.37,
"index": 197,
"start_time": 5250.589,
"text": " Life's too short, folks. Define things however you want. Just be clear what it is you are defining. And once we have the definition, and we don't have any definitions, have no import by themselves. It's only things we can conclude starting from the definitions. Make your definitions. Feel free. Just be very clear what your definitions are. And then after that, there might be some interesting consequences to investigate."
},
{
"end_time": 5299.172,
"index": 198,
"start_time": 5278.933,
"text": " But to have fights about definitions is preposterous. And that's what a lot of these fights really are about, is just definition pushing. Let me steel man the definite. I, up to just a few months ago, am on board with you 100%. Now, let me steel man the definitions point of view. I was speaking to Professor Jor Bar-Natson of mathematics in Toronto here."
},
{
"end_time": 5323.609,
"index": 199,
"start_time": 5299.445,
"text": " I thought you might be going there, but here's what I think he actually meant. He thought that coming up with definitions that lead to a very rich set"
},
{
"end_time": 5354.002,
"index": 200,
"start_time": 5324.548,
"text": " of theorems and proofs that that is the key to being here is my putting words in his mouth to being a successful and important and do important work in mathematics is to first come up with the very fruitful definitions. I fully agree that but that is not what people who are arguing about free will are arguing about. They're not saying for which definition of free will and he wasn't actually either your professor. He wasn't saying which definition of"
},
{
"end_time": 5384.394,
"index": 201,
"start_time": 5354.599,
"text": " He was saying, no, what's important is if I put together the words in a certain way, and then I call them, what I call them doesn't matter. But those set of words put together in this way, we'll call it A alpha B beta. We've just defined A alpha B beta. If I can use A alpha B beta to prove a whole bunch of fruitful things, then that's a good definition. These people instead are saying, what should A alpha B beta mean?"
},
{
"end_time": 5412.619,
"index": 202,
"start_time": 5385.316,
"text": " which is backwards. So that professor, he was saying that it's not the words, it's not the clothes don't make the man. It's not the words that you're draping this with that matter. It is whatever that this is. And here people are saying, well, I've got my words first and I'm going to fight with you about whether they actually mean this or they mean that. It's useless. It's backwards."
},
{
"end_time": 5417.073,
"index": 203,
"start_time": 5413.217,
"text": " It's not only anti-scientific, I would say it's anti-intellectual almost."
},
{
"end_time": 5446.954,
"index": 204,
"start_time": 5417.961,
"text": " Now, for the sake of playing the devil's advocate, let me know what you think about this. Now, when we're not talking about mathematics, something that's specifically well defined, when we come up with a definition, it works somewhat akin to this. We have an intuitive notion of some concept and it's ambiguous. And then we give it a name and we usually give it a name by pointing to different objects that satisfy what we think is that intuitive definition. So, for example, with life, we'll say the couch behind you is not alive. You're alive. The beetle is alive. The fire is alive."
},
{
"end_time": 5476.271,
"index": 205,
"start_time": 5447.654,
"text": " Let's say minus 3000 years ago. And then there's other people arguing about that. Well, the fire is not alive, but I do agree David is alive. And I also agree that the couch is not alive. So we have these volutinous feelings of amorphous balls and we're trying to see, okay, well, what's the intersection between us and which one captures what we're trying to say most, which one captures the most relevant concepts. So that's why there are disagreements about the definition. We're trying to figure out, we're arguing about what the entries in the OED should be."
},
{
"end_time": 5505.469,
"index": 206,
"start_time": 5476.749,
"text": " There's an entry in the OED for life and what should it be in the Oxford English Dictionary. And so, for example, on that particular topic, people do define life as a bit. And the question is, well, how should I define that bit? They want it to always be it's alive or it's not alive. How do we define what one or the other means? So this is other work of mine early in my career I did with a fellow called Bill McCready and others have picked it up since."
},
{
"end_time": 5532.654,
"index": 207,
"start_time": 5506.374,
"text": " So my reaction to that was that this is silly definition plugging come up with your definition and whatever to say something is or is not alive is silly. At a minimum if you're talking about biological organisms it's not going to be just a single bit it's going to be actually at least a real number and well actually it's going to be a vector well actually no it's going to be some much more higher dimensional thing it's not going to be life is a bit"
},
{
"end_time": 5561.834,
"index": 208,
"start_time": 5533.029,
"text": " life is some other mathematical structure. And here's one that we happen to come up with that actually turned out to be somewhat fruitful. Another topic of research I wanted to work with. You know what a fractal is? I assume. Okay. So I take the magnifying glass and magnifying that microscope, but I keep seeing the same thing. Let's take that magnifying glass, that microscope, put it on your skin. As we keep running up the power,"
},
{
"end_time": 5591.186,
"index": 209,
"start_time": 5562.602,
"text": " We go down to the level where we're seeing now individual cells. Now we start going within the cells and so on. The images in that microscope keep changing. You are not self-similar. You are self-dissimilar. The patterns in you are very different from one another at each different scale. In fact, that seems to be true of almost all systems that are alive."
},
{
"end_time": 5621.067,
"index": 210,
"start_time": 5592.261,
"text": " It is not true of almost all systems that are not alive. Moreover, this self-disimilarity is a way to give nuance and structure to the form of the life. Because you can actually be looking at how is it that the patterns at each scale are related to one another. You've got an entire matrix. And so this we call it self-disimilarity. And we were very careful not to say it's living or not."
},
{
"end_time": 5644.224,
"index": 211,
"start_time": 5621.698,
"text": " And we simply say, actually, I think we've reversed this effect. Any system that does have self dissimilarity, high self dissimilarity, is almost by definition going to be very interesting. And some people might say it's actually a complex system. But we, again, we don't want to be saddling it with these definitions. We're just going to simply say it's self dissimilarity."
},
{
"end_time": 5670.879,
"index": 212,
"start_time": 5644.633,
"text": " We put it out there as a proposal to the research community that they shouldn't be spending all this effort in trying to figure out what should the definition of this bit, living or not living, be. That that's fruitless and it's just going to be an interminable wars that never are going to end up being rich in the sense of definitions that your mathematics professor was thinking. Let's instead be thinking about the self dissimilarity of different systems."
},
{
"end_time": 5696.903,
"index": 213,
"start_time": 5671.834,
"text": " Because why would you think that an interesting living system might be self-disimilar in the first place? So it's not just that we can notice phenomenologically that living systems are all self-disimilar, whereas non-living systems like mountaintops and so on are self-similar. We can actually ask why might that be the case? And so here's an interesting proposal that I have not yet time to investigate."
},
{
"end_time": 5724.206,
"index": 214,
"start_time": 5697.517,
"text": " What if biological systems, this is somewhat along the lines of the things that Carl Friston is interested in, what if biological systems had to, and also for that matter Don Hoffman, had to, as an evolutionary imperative, extract the maximal computational bang out of a given physical body buck? How might they do that? Well, one thing you can think of doing"
},
{
"end_time": 5752.21,
"index": 215,
"start_time": 5724.599,
"text": " is distribute different parts of that computation to different spatial locations in that body. And that's what systems do, biological systems to a large degree. They have the computations that are going on in your brain being a different part of you from the computations that are going on in your kidney, for example. The computations have been distributed. But you might also not only want to distribute the computation across different positions, but also across scales."
},
{
"end_time": 5778.507,
"index": 216,
"start_time": 5753.268,
"text": " There are, if you think about it, different computations being done by the cells in you compared to the computations being done by the organs in you. Moreover, these computations being done at the different scales are actually communicating with one another. So you can imagine that as an evolutionary imperative, biological systems would be driven to want to distribute the computations that they do"
},
{
"end_time": 5808.933,
"index": 217,
"start_time": 5779.206,
"text": " across their different scales, as well as their different spatial positions within them, which would then lead us to start to think about what does it even mean to distribute computation across scales? What are the general features of such systems? And it would lead you down a research path, which I've not had time to go, but it would lead you down a fruitful research path to, in fact, refute and dispute and"
},
{
"end_time": 5839.172,
"index": 218,
"start_time": 5809.855,
"text": " I vow no to the concept that life can be reduced to a bit. If we start by saying that this exercise is an exercise in futility, that instead we want to come up with not a definition that's reducible to a bit, you are or you are not, but rather look at this much more nuanced structure, it might lead us in much more fruitful research directions. It might be a much more useful definition in the sense of that mathematics professor you were describing before."
},
{
"end_time": 5868.985,
"index": 219,
"start_time": 5839.889,
"text": " Hi, I'm here to pick up my son Milo. There's no Milo here. Who picked up my son from school? Streaming only on Peacock. I'm gonna need the name of everyone that could have a connection. You don't understand. It was just the five of us. So this was all planned? What are you gonna do? I will do whatever it takes to get my son back. I honestly didn't see this coming. These nice people killing each other. All Her Fault, a new series streaming now only on Peacock."
},
{
"end_time": 5900.52,
"index": 220,
"start_time": 5870.964,
"text": " Now, if you can answer this one quickly, because this is an aside about what you said, you mentioned that it may not be a bit, it could be a vector. Now, I see how it could not be a bit, it could be a real number. But I don't see how is it that this dissimilarity measure could be a vector like R2, R3, etc. So here's what I can do. I take that microscope, this is an exercise that we've actually did on various image data sets, for example. And we applied it to things like the logistic dynamics, and so on and so forth."
},
{
"end_time": 5930.623,
"index": 221,
"start_time": 5901.084,
"text": " Take that microscope that's getting distributions of the visual images of the patterns at different scales. These are now going to be distributions over the pixel pattern within the microscope's field of view. There's many measures from machine learning, probability theory, information theory for saying how different two distributions are. So this way I can say"
},
{
"end_time": 5957.841,
"index": 222,
"start_time": 5931.015,
"text": " For any one of the following, say 10 scales, how different is the distribution over images at those two scales? So I'm comparing scales magnification two times and magnification seven times. Those scales have associated distributions over images. I look at how different those distributions over images are at the scale two times and seven times. That gives me a real number."
},
{
"end_time": 5985.776,
"index": 223,
"start_time": 5959.172,
"text": " I then look at it for another pair of magnifications, that gives me another real number. So in this, so in essence what you were seeing that this set of differences between distributions at the different scales can be wound up into one, actually it's a matrix, because you're putting in the two scales that I'm comparing into these elements of these matrix. So if you give me two scales, I'll give you a real number,"
},
{
"end_time": 6001.049,
"index": 224,
"start_time": 5986.152,
"text": " is actually more rich than a vector as a full matrix. Symmetric matrix? Yes, this particular case is symmetric, exactly so. Okay, so now this question comes from Kevin Knuth and I know you gotta go, don't worry, I gotta go to Kevin Knuth."
},
{
"end_time": 6019.019,
"index": 225,
"start_time": 6001.578,
"text": " Our friend Kevin Knuth, and I'll read it. So he wants to know what is life. It's a question that people have entertained for a long time without obvious answers. Your thoughts on observers, I sent him your video on observers, seem to be relevant. I have always thought of living things as systems that use information to create some kind of model of the world."
},
{
"end_time": 6047.295,
"index": 226,
"start_time": 6019.019,
"text": " around them in order to locate and take advantage of potential energy gradients, which are then used to keep them away from equilibrium. And I can give some examples and so on. I would agree that a hurricane is not alive. A hurricane is a heat engine that he's referencing your talk. However, it's not alive because it doesn't use the information about its surrounding to model these energy gradients. The models living things can employ are quite diverse for humans. These are our mental models, etc, etc, some bacteria, etc, etc. So these are my thoughts about life. I really need to write them down."
},
{
"end_time": 6073.575,
"index": 227,
"start_time": 6047.295,
"text": " Maybe David would be interested in working with me on these. Okay, so these are Kevin Knuth's thoughts. I'm sure you can pull a question out from that or just comment on it as a whole. Yeah, so what Kevin, what he's talking about here, first of all, hi, Kevin. And yes, I think we could have some very interesting discussions about here. So as I mentioned before, and okay, so backing it up."
},
{
"end_time": 6103.097,
"index": 228,
"start_time": 6074.189,
"text": " Kevin is talking about presentation I made. That's where there was observers and hurricanes and so on. The observers are what I was there using as a term to describe these non-equilibrium steady states that if you were to basically, in a Judah-Pearl sense, intervene on the information flow from the outside world into these systems, they would no longer be at non-equilibrium. They would actually collapse to being at equilibrium."
},
{
"end_time": 6130.503,
"index": 229,
"start_time": 6104.104,
"text": " And we were calling that, and this is what we were calling an observer. This was what they were calling this talk of mine that Kevin saw. And we were making the definitions be, we were careful so that we would not, I mean, the words observer and alive are being just mangled around here. But that we were trying to make it so that a hurricane is not in a certain sense observing its environment."
},
{
"end_time": 6159.343,
"index": 230,
"start_time": 6131.101,
"text": " whereas a paramecium going up a gradient of food is observing its environment and is alive. So, let's see. So, well, I and my collaborator Artemiy Kolchinsky were going after in that paper where we define somatic information and observation and so on"
},
{
"end_time": 6185.64,
"index": 231,
"start_time": 6159.701,
"text": " as a system that's in a non-equilibrium steady state, such that if you intervened on the information it gets from the outside world, it would collapse to equilibrium. Equilibrium being a syndrome for depth. We were saying that those are systems that are either observers or living or just generally of interest without using loaded terminology. That is all the what."
},
{
"end_time": 6216.442,
"index": 232,
"start_time": 6186.988,
"text": " Kevin is talking about models that the system has internally that basically play a middle step in the map from the information entering from the outside to the actions that the entity does back on the outside that end up allowing the entity to be away from equilibrium. The model is the way that it does that. It's a way of interpreting what's going on in the entity"
},
{
"end_time": 6244.599,
"index": 233,
"start_time": 6216.732,
"text": " in between these steps of getting information from the outside and then acting back on the outside. What we were analyzing in our work elided that step. We were not trying to say anything about internal models. We were simply saying the end result is that if you interrupt the information flow from the outside, we were saying nothing about the reasons, whether it goes through models or some other"
},
{
"end_time": 6270.742,
"index": 234,
"start_time": 6245.247,
"text": " phenomena, but the ultimate effect was that you collapse back to equilibrium. So it was saying that if you, if I sunder the information you're going from the outside, you're getting right now information from the outside about where a cheeseburger is on a table. If I stop that information flow, eventually you die, you fall back to equilibrium. So that's it. So what you're doing is you are right now in a non-equilibrium steady state"
},
{
"end_time": 6288.985,
"index": 235,
"start_time": 6271.169,
"text": " Because you're getting information from the outside world and what we were analyzing there is systems where if you were to chop that information flow from the outside world, the result is that they actually collapse back down into being equilibrium, they die. The means by which that happens"
},
{
"end_time": 6315.64,
"index": 236,
"start_time": 6289.838,
"text": " in your particular case are well, if you had that information about where the cheeseburger is, you could then reach out your hand based upon an internal model of the outside world, grab that cheeseburger, bring it to your mouth, digest it, get a whole bunch of glucose, which is what you then use to stay out of equilibrium. So that's everything that happened in between you're getting the information and you're staying out of equilibrium."
},
{
"end_time": 6345.077,
"index": 237,
"start_time": 6316.493,
"text": " In other processes like the paramecium, it's getting information about the gradient of a chemical in its environment. It is then using that information, one could maybe say in implicit genetic model, that's what Kevin is saying, but whatever. The end result is that it uses that to determine how it should reorient itself before it uses this flagella to actually move so that it can actually then ingest these particular chemicals and that's how it stays out of equilibrium."
},
{
"end_time": 6374.377,
"index": 238,
"start_time": 6345.845,
"text": " So in my work with Artemiy Koltchinsky, all that we did was looked at the beginning point and the end point. Notice that you need the information coming in to stay out of equilibrium. Those are interesting systems, we said, from a physics point of view. Kevin is now discussing a special case of such systems, which are physics systems that capture at least a large part of what one might want to say is a living"
},
{
"end_time": 6403.643,
"index": 239,
"start_time": 6374.94,
"text": " So that's a very long-winded answer to Kevin's question. Kevin, yes, if you do want to, in particular, the directed information flow of people like Daniel Polanyi and Nihat Ai, I think could be a very, very interesting, fruitful way to try to build more structure into these particular systems, a halfway point to building full-on internal models."
},
{
"end_time": 6424.326,
"index": 240,
"start_time": 6404.36,
"text": " So yeah, Kevin, drop me a line. Now the last question you can answer, hopefully fairly quickly, it's even though it's long, it's one sentence. It's actually just one sentence. So I love this one comes from Professor Edward Lee of Berkeley. I love David, how your work has complicated the faith many scientists hold in an ultimately deterministic world."
},
{
"end_time": 6453.575,
"index": 241,
"start_time": 6424.326,
"text": " I was wondering whether you've looked at John Norton. So we talked about that. So we actually did talk about that in particular. Why is this example controversial? Why are so many people offended by the idea of uncaused action? Yeah, I've also encountered Professor Lee's work and I find also mutual admiration here that I am impressed with a lot of his work. This is actually a subtle thing. I'm very careful in my stuff on inference devices."
},
{
"end_time": 6481.817,
"index": 242,
"start_time": 6454.053,
"text": " There is no laws discussed, mentioned anywhere. There's no rules. There's no mathematicalization of the patterns in the underlying reality. It's just world lines in a set. And an observer is one particular subset of those world lines. Something it's observing is another subset. What it means to observe"
},
{
"end_time": 6510.469,
"index": 243,
"start_time": 6482.227,
"text": " has to do with the relationship between those two subsets. What it would mean to make predictions would then be another subset. And basically by means of these subsets and their overlaps, this topology, you would build up the structure of a scientist using theories to make predictions about what might occur in the world. And those theories are going to be in the form of rules and laws. And so, for example, Max Tegmark's multiverse work"
},
{
"end_time": 6539.087,
"index": 244,
"start_time": 6510.862,
"text": " I highly plug his Foundations of Physics paper and the earlier Annals of Physics papers on the multiverse. Don't bother with pop books, go for the real me. In that work, he ultimately reduces the notion of physical reality to a set of laws, to a set of mathematical rules, and the presumption that all physical reality is just a set of rules, which is"
},
{
"end_time": 6567.807,
"index": 245,
"start_time": 6539.548,
"text": " precisely why John Norton's example is controversial because it seems to wreak havoc with the notion that physical reality can be reduced to a set of rules. That's, I think, ultimately from a sociological point of view, why people are upset about that. I'm very careful in my work that rules don't appear. There's just subsets, there's a topology, there might be patterns. Those patterns can be violated at any point in time. No free lunch."
},
{
"end_time": 6593.422,
"index": 246,
"start_time": 6568.285,
"text": " Newton's laws, even if they are there, they might just suddenly disappear tomorrow before the ball of John Norton has even had a chance to try to think about falling off of the dome. But many people instead want to implicitly, they think of a mathematical universe as being a set of things like differential equations."
},
{
"end_time": 6619.514,
"index": 247,
"start_time": 6594.07,
"text": " rules, laws, and I just think of it instead as being a set, and there can be a topology on that set, and the topology might have patterns which you might be able to embody as such rules and laws, or it might not. It all just depends on the topology, and different topologies are going to either be rule-like or not rule-like, law-like or not law-like."
},
{
"end_time": 6647.398,
"index": 248,
"start_time": 6620.026,
"text": " And it's not even the case that one is real and one is not real. They're just all different realities. It's legitimate to one, as legitimate as one another. There's no privileged reality. I suspect that all of these kinds of answers I'm giving you these questions are not necessarily fully fruitful. As an aside, as an interviewee, to give a answer to a question in a standalone manner rather than in a dialogue,"
},
{
"end_time": 6675.145,
"index": 249,
"start_time": 6648.217,
"text": " Professor, thank you so much for spending so much time with me. And the audience, I'm sure, thanks you too. Okay, well, thanks very much. Sorry if I"
},
{
"end_time": 6705.384,
"index": 250,
"start_time": 6675.794,
"text": " It was too much open-loop control. I think I was a little bit logarithmic today. So thank you very, very much. And if you could give me the information about the forum for this, so I can then forward that to the SFIP people. And Kevin, contact me. Great. OK, thanks. OK, goodbye, Professor. Thank you. OK, bye. Truly, it was a blast. It's great to be speaking to someone who also is philosophically inclined. Even though you had your excoriations on philosophy, you still had your philosophy of philosophy."
},
{
"end_time": 6734.394,
"index": 251,
"start_time": 6706.8,
"text": " Yeah, in many ways, yes, I think philosophy is almost by definition, the most important intellectual endeavor there is, but and this is being a real snide asshole. I don't think it's really unfortunate, people that actually engage in it have not been trained properly. And that's most of philosophers. It's unfortunate, the obviously exceptions, you know, Dan, Dan, I think,"
},
{
"end_time": 6757.073,
"index": 252,
"start_time": 6735.418,
"text": " Training can either be explicitly or implicitly. And I think Dan is an example of somebody who's, quote, trained well to be really speaking in snotty asinine terms. But yeah, I think it's the most important subject there is, almost by definition. All right, take care, sir. Okay, you as well. Thank you. And we'll talk soon. Bye bye. Okay, thank you."
},
{
"end_time": 6777.022,
"index": 253,
"start_time": 6757.892,
"text": " There were two extra questions from Sam Thompson and Chris Langan that we didn't have a chance to address on video due to limited time, but David was kind enough to answer over email. Sam Thompson asks, do you see any sociopolitical implications in your work on the formal limitations of inference devices? David Walpart says, interesting question."
},
{
"end_time": 6804.309,
"index": 254,
"start_time": 6777.363,
"text": " One could consider, for example, the Laplace's demon theorem or the monotheism theorem as concerning limitations of what a government is capable of, but those theorems are based on worst-case scenarios. They show that there is some question-answer pair that an inference device gets wrong, not that it must get wrong a particular such pair. More interesting and future work would be an extension of the inference theorem, would be an extension of the inference device formalism to incorporate probabilities and degrees of failure."
},
{
"end_time": 6827.892,
"index": 255,
"start_time": 6804.309,
"text": " Now on to the question from Chris Langen."
},
{
"end_time": 6851.186,
"index": 256,
"start_time": 6828.285,
"text": " David, you base your definition of observer on principles of physical causation including thermodynamics and statistical mechanics... David interjects. I've used the term observer in different ways depending on whether I'm talking about inference devices or the semantics of information formalism with artemy or noisy deductive reasoning. Which meaning do they have in mind? So then Chris Langan finishes the question."
},
{
"end_time": 6880.725,
"index": 257,
"start_time": 6851.186,
"text": " Do you subscribe to physicalism, the idea that reality is basically physical in nature? Or do you think that a broader, more sophisticated understanding of causality, one that exceeds the classical concept of physical mechanism and quantum theoretic concepts like entanglement, non-locality, and quantum wave function collapse, will eventually be required? David answers. In short, there is right now absolutely zero data slash evidence to believe that will, quote-unquote, eventually be required. So Occam's razor provides the answer."
},
{
"end_time": 6902.466,
"index": 258,
"start_time": 6883.302,
"text": " The podcast is now finished. If you'd like to support conversations like this, then do consider going to patreon.com slash c-u-r-t-j-a-i-m-u-n-g-a-l. That is Kurt Jaimungal. It's support from the patrons and from the sponsors that allow me to do this full time. Every dollar helps tremendously. Thank you."
}
]
}No transcript available.