Audio Player
✓ Using synced audio (timestamps accurate)
Starting at:
David Wolpert on Free Will, The Limits of Science, No Free Lunch theorem, and Mathematical Universe
January 28, 2022
•
2:12:04
•
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
293 sentences
18,890 words
Method: api-polled
Transcription time: 129m 8s
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.
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.
This is part one of a conversation with the inimitable David Walpert. You'll want to watch this one all the way until the end to understand the concepts which will be referenced in part two, where we take questions from you, as well as hearing questions from Kevin Knuth, Anil Seth, Karl Friston, and Chris Langan, all posed to David Walpert. Check the description for the link to part two.
David Walpart is a mathematician, a physicist, and a computer scientist. He's also the pioneer of three limiting theorems, including the No Free Lunch Theorem, which, as astounding as the No Free Lunch Theorem is, is actually the least monumental of the three. That statement alone is a testament to how seminal David's work is.
Today we talk about free will and what the limiting theorems have to say about the restrictions on attaining scientific knowledge in general. As usual, click on the timestamps in the description if you'd like to skip this intro. This episode, much like most of the podcasts on the Toe Channel, is extremely technical, but don't be dismayed, stick with it even if you don't understand it, as repeat viewings are what allow you to, in John Wheeler's words, get wet. The whole point is simply to get wet rather than to drink from the fire holes.
This episode will serve as an introduction to David Walpart's ideas and then later we'll have a deep dive into the intricacies of David Walpart's theorems for part two. This means you should write down your questions. Write down all your questions either privately in some documents or publicly, let's say in the YouTube comment section, especially because I'll be culling from there, since when you write it may spur questions in others or it may even answer a question that someone else had.
Limiting theorems and no-go theorems, which can be used interchangeably, you'll hear this terminology plenty, are epochal and particularly relevant to the subject of theories of everything, since many people attempt to formulate their own theory of everything and contravene certain bounds without realizing that they're overstepping certain limitations. For example, if one believes that every well-founded mathematical statement can be either proved or disproved, then one is
There's nothing wrong with violating per se, it's just that one should know when they're violating and what the implications of that violation are. So for example, perhaps it implies that you can create a perpetual motion machine, or that you have a method of trivially predicting the outcome of some chaotic configuration.
In short, understanding no-go theorems allow you to understand what loopholes exist and thus gives you an intricate understanding of the inner workings of reality. My name is Kurt Jaimungal. 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 delineating the possible connection consciousness has to the fundamental laws of nature
if those laws exist at all and are knowable to us. If you enjoy witnessing and engaging with others on the topic of psychology, consciousness, physics, etc., then check the description for a link to the Theories of Everything Discord as well as subreddit. There's also a link to the Patreon, that is patreon.com slash Kurtjmungle, if you'd like to contribute as the patrons and the sponsors are the only reasons that I can do podcasts with this level of quality and depth
As now, this is what I do full-time thanks to their support. With regard to sponsors, there are two. The first sponsor is Brilliant. During the break, I decided to brush up on some fundamentals in physics, so I committed myself to learning one lesson per day on Brilliant. Some point soon, I'd like to speak to Chiara Marletto on constructor theory, which is predicated heavily in information theory, so I took a course on random variable distributions and knowledge and uncertainty. Despite previously knowing the formula for entropy as it's hammered into you as an undergrad,
It was extremely edifying to see a different explanation of the formula for entropy that doesn't seem like one that's been handed down arbitrarily from God. Instead, after taking the course, it's easy to see why the formula for entropy is the way that it is and how it's an extremely natural choice. There are plenty of courses, including ones on group theory, which is what's being referenced when you hear that the standard model has in it u1 cross su2 cross su3.
I recommend that you don't stop before four lessons and I think you'll be greatly surprised at the ease at which you can now comprehend subjects that you previously had a difficult time grokking. At least, I know I did.
The second sponsor is Algo. Algo is an end to end supply chain optimization software company with software that helps business users optimize sales and operations planning to avoid stock outs, reduce returns and inventory write downs while reducing inventory investment. It's a supply chain AI that drives smart ROI headed by Amjad Hussain, who himself has a podcast on artificial intelligence and consciousness and business growth.
I've always been interested in theories of everything, as I think any physicist, any theoretical physicist should be.
Well, that's a controversial statement. But at least if you're naive, you should be interested in that. And then maybe later you realize, okay, there are limitations and so on. Now, those limitations is where our conversation comes in. I get email probably much like yourself, many theories from people. Maybe you don't get email that but either way, I get email plenty of theories. And they violate some limiting theorem or no go theorem. And they don't realize that they're doing this. So I thought, you know what, how about
While I'm on this journey to investigate the different theories of everything, how about we educate people on the different limiting theorems that exist? Now you have three, which is extremely rare. Most people would kill for one. So that's why I'm extremely interested in speaking with you. And we're just going to talk about those limiting theorems and what the implications are. Okay, I'll do my best. I thought I'd break this down into three stages. One for each of the limiting theorems. Now you have
three as we talked about off air, the no free lunch theorem, which we're going to abbreviate as NFL, and then the unreality of Laplace's demon slash limits on science that's under the umbrella of, I'm unsure what you call the theorem, but as far as I call it the limits on inferences. So limits on inferences is second, and the noisy deductive reasoning, which you've told me that it flourishes the foundations of mathematics, but I'm unfamiliar with that theorem. So I'll be I'm extremely excited to hear about that.
Okay, so no free lunch, there's actually several different flavors of the no free lunch theorems. There's the ones that have to do with induction, which can be viewed as a formalization of David Hume's informal idea, and which have lots of implications for machine learning and so on and so forth.
then there's also the versions that have to do with optimization and search. Inference devices, there's actually a whole bunch of impossibility results. The monotheism theorems and the deism theorems and so on, those are two of them and they're up front. The noisy deductive reasoning, that I wouldn't refer to that as
impossibility or no-go theorems. That's very much work in progress. That's probably, it's on the front-most burner of all those separate projects. What is, what me and my collaborator David Kinney are doing there is instead trying to understand what the implications would be if we just
pursued the time-honored tradition now in mathematics of weakening one of its foundation stones, let the edifice then crumble and relax into a somewhat different shape than we had known before, and try to say some things about that particular shape. So it's not so much a no-go theorem. You will not be able to
invoke it when you get one of your email correspondence and say oh this violates the blah blah. It's instead much more of a project that in a certain sense is even undermining the legitimacy of the others and that it's saying it's taking seriously the fact that in mathematical deduction
you are never, it does not seem to be possible, and so let's not take it as an axiom rather than try to prove it or not, that you're ever actually 100% sure that every step in your proof is correct. And that's obviously true when people first come up with, say, a suggested solution to the Poincare conjecture or something like that. Or when Andy Wiles came up with his proofs that, among other things, established Fermat's Last Theorem,
It was very, very clear at the time that there was a non-zero probability that there was a mistake, so we thought it was wrong. Mathematicians tend to treat these results as though they almost, for lack of a better word, undergo a phase transition, that after people have gone over them enough that, yeah, now we're sure of them and we're going to treat them as, what, 100% ironclad, wholly writ
When in point of fact, of course, that's never what happens. It's rather the set of people are always, it's a stochastic process by which you come to greater and greater confidence in the results. May I interject? Please, please, please, do that as much as you possibly can. Otherwise, I have a tendency to open loop. Yeah. I'd like to make this clear for the audience. Generally, when a mathematician proves something, you don't go through
what you may learn in a logic course which is you state the axioms then you have some rules of inference like modus ponens and then you go step one is a step two step three step four step five that's actually extremely tedious even to prove the most simple theorems now this can be done on a computer but well this can be done on a computer so I don't think professor you're talking about those actually even then because this is for example the proofs of the four color map theorem which is the first one that was done on a
computer. It's actually a somewhat controversial issue in practicing mathematics as well as philosophy of mathematics is, first of all, is it a proof if human beings can't understand it? But even more to the point, you now have to validate the actual code, the program that's being used, and that can never be done to a hundred percent. So you can't ever quite, I don't think there's actually anybody has tried to formally have a no-go theorem
that proves this, which is an interesting idea. But as a matter of practice, there's always some place that you need to simply say, well, it's been vetted, it's good enough, let's run with it and not scrutinize this anymore. Your operating system, computer programs, cosmic ray can hit a transistor and it can make a fluctuation. So you can think that's validated by computer, even then you're not completely
This is Marshawn 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 touchdowns 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.
What do mathematicians say in response to what you're saying? Because it's as if you're undermining the entire validity of their existence. Well, they say that very, very quickly. You get to the situation where the probability that it's actually mistaken
The more it gets scrutinized, the lower that probability becomes. Formally though? Because informally you can say that. Can you say, well, we believe with a 99.58% certainty in Fermat's Last Theorem's proof, for example. They never actually do that. This is all a matter of heuristics, the sociology of the field. So one of the ways of viewing the motivation for the noisy deductive reasoning stuff
is that you can imagine trying to actually formalize that. Say there is a stochastic process, you can imagine just where we are, what's called IID, so that's independent identically distributed, doing samples of whether a, so you can imagine a very, very simple model where here is a hypothesis, a proposed theorem, and that you are many times independently testing with a simple bit flip
Is this actually true or not? And if you keep getting those flips that keep coming up saying true, true, true, true, and they're all independent, then according to many things like Bayesian reasoning and so on and so forth, it depends upon what are called your prior probabilities. What's called your posterior probability that actually is true given the data
that it's been tested now a thousand times, now a hundred thousand times, now a million times. The probability that's wrong as the number of tests, the number of different people who have checked it over keeps growing, that probability of error shrinks as the number of checks grows. It's actually, you can make a very, very simple model where it's formally identical to the proposition that the sun will rise tomorrow.
You personally, Kurt, have seen the sun rise, I don't know, a couple of thousand times. If we may be squinting our definition of what it means for you personally to see it rise, we could allow you to pull upon social media or just general media and so on and say it's even tens of thousands. So there's something called Laplace's Law of Succession. This goes back to Simon Laplace.
where he uses very, very simple Bayesian reasoning. He was actually the first person to really use Bayesian reasoning. Bayes didn't really use it, essentially. So Laplace used it to estimate trajectories of planets and whatnot. But in any case, there's something called Laplace's Law of Succession, which says that if you have seen it rise a thousand times and you've never seen it fail to rise,
then under a very reasonable choice for what's called the prior probability, you would expect, you would assign the expected chance of it rising tomorrow is 1001 divided by 1002. But that's not the same thing as one. So you can build a very, very simple model of what mathematicians are doing, which is that every time a new mathematician takes a new look at the proposed proof,
that it's just like having the sun do another test of whether the sun's rising. And just like every time you see the sun rise from one day to the next, you get more and more confident that, yep, the sun will also rise tomorrow. Mathematicians, every time they do a new test of having somebody scrutinize it and coming up and saying, yep, looks legitimate to me, you can say there's a greater probability that the next mathematician will come to that same conclusion.
but in both cases you'll never actually achieve a notion, anything like, whether you can even define it, saying it is true, 100 percent. And so you can build simple mathematical models or the mathematical process along those lines, and doing that you can actually justify some common mathematics heuristics. Mathematical reasoners, for example, if you are
Still in the stage where you're trying to ascertain whether you think a proposition, a proposed theorem is true, if as a mathematician or a theoretician in any field, if there are what you might view as multiple independent lines of reasoning, all of which say that that proposition is true, versus there only being one, when there's multiple independent ones, you are going to assign higher probability to being true than not being true.
That's as a matter of social practice. Humans do that. And by using this model of mathematical reasoning, you can actually say, guess what? That's actually formally justifiable using some of the rules of Bayesian reasoning. So one thing one can do is take this approach to what mathematicians actually do and actually establish that some of the heuristics that mathematicians engage in are actually legitimate.
But where things become much more interesting to me, at least personally, is if you say along the spirit that, well, let's weaken the parallel line postulate. And look, guess what comes out? Non-Euclidean geometry and all of its riches. Let's weaken the postulate, the continuum hypothesis. And guess what? What comes up is all different kinds of mathematical logics and so on and so forth. We can say let's weaken the hypothesis
that mathematics itself is in some sense external to humans, that humans are discovering it and that it, mathematics, is ironclad with every step being following from a previous step in a deterministic manner. Instead, let's weaken that and assume that it's not just mathematicians who are sarcastic prophecies, but in some sense that which they are engaging with, mathematics itself,
the underpinnings of our physical reality, if you are going to adopt a Max Tegmark's view of what physics is, that's just mathematics, that it itself is inherently a stochastic process. That the quadratic equation, it's not just human mathematicians whose confidence in the quadratic equation is never 100%, that actually the quadratic equation itself
If you're to adopt an almost platinous view of mathematics, this external thing that humanity is trying to get its hands around, that in itself also is not 100% true, but is only squishy. What are the consequences? And that's what we're wrestling with in that body of work. So professor, you mentioned that you use the phrase for a reasonable choice of priors. So how does one
Yep. So here's what I'm thinking. The priors, in some sense, when you're doing Bayesian reasoning, it doesn't matter. What matters is that you update it in accordance to a certain model. So then you want your posteriors to be a certain amount. Let's say you want to be 99% certain of a certain fact. Is it not true that you can then change your priors? So like, let me figure out how to phrase this. You're spot on. The priors? Essentially, you can always change the priors to give you any posterior probability that you want. You can reverse the equation. Yep. You're correct.
You're correct. And that's related to the no free lunch theorems. Another way of phrasing, this is why sociologically, Bayesian reasoning fell into very ill repute around the beginning of the 20th century. And that's why Fisher and other fellow travelers, they came up with things like hypothesis testing, confidence intervals, all that kind of stuff.
what's called sampling theory statistics became ascendant and Bayesian reasoning fell into disrepute because people misused it by choosing priors that were leading to results that kept coming up wrong. And Bayesian reasoning is a matter of garbage in, garbage out. To push it even further, there is no way that anybody's even imagined that you can somehow prove what the priors are.
what they should be, how you should reason. There are heuristics. People like to use what's called uninformative priors, what are called Jeffrey's priors. Ed Jaynes in his book, The Logic of Science, and many of his, I use the phrase, fellow travelers, they will try to say that your priors, that it behooves you to choose priors that accurately reflect the precise forms of ignorance.
that you have about a scenario. That's nice, seems reasonable. They've not proven that injunction. It's just they're taking it almost as a feature of good practice. One might dispute it, and even if one does not dispute it, it's only going to get you so far. There are many scenarios where just what you might mean by
quantifying my ignorance as a prior is not at all clear. So yeah, so we are, in a certain sense, the no-fee lunch theorems are saying this by one way of to view one of their contents is that if I have any particular reasoning algorithm, be it in say the machine learning algorithms,
or optimization algorithms, evolution itself, any of which is a search algorithm. And if I find that of my two algorithms, A and B, that for one set of priors, A outperforms B, so it's going to make more accurate predictions about whether the sun will rise tomorrow or what have you, that there's always going to be just as many priors
for which B beats A. So it's a way of taking your intuition and not only quantifying your intuition, but also elaborating that it's more than just one teeny little squirrely. If you want to be a naysaying annoyance, you can find a counter example. It's saying, no, it's not just that there's one prior in which the conclusion that you want to reach
When you say just as many, you mean in terms of the size of the set? The no-free lunch theorems are stated in the following way. They say that if I take a
In the scenarios to which they apply, you can have a prior itself, you can define a probability distribution over priors. And you can say, let's just mathematically, this is a mathematical trick. Say there's a uniform distribution over priors. Then I can say that on average, where I'm averaging over that distribution over priors,
any two algorithms perform the same. So that means that if I find one particular set of priors whose measure under that uniform distribution is such that that algorithm beats A, A beats B under that set of priors, then there's another set which has equal measure under this uniform distribution.
So yes, it's the same number, but number is being defined very precisely in terms of a probability distribution. We're talking about infinite spaces as a thing. So it's not just counting, oh, there's three priors for which a beats b and three priors for which b beats a. It means the size of the set in a volume sense by which a beats b is the same as which b beats a.
I didn't know that your theorem had something to say about Bayesian reasoning or it used Bayesian reasoning in it. Whenever I hear people speak about Bayesian reasoning, they hold it as this pinnacle of rational thinking. And generally people who are who like you mentioned in an email, which we can remove this part where you said there are certain people who try to sound as if they're extremely intellectual. They say it as if it's ironclad. However, the process may be a rational process, but the outcome
You can reverse engineer your priors to give you whichever outcome you like. And I've always found that, so why is it that people keep stating, well, look, I'm a skeptic, I update my priors, I do so and so, and so therefore I've arrived at the correct result. I don't think that's the case. And I haven't heard anyone make a claim otherwise until now, right now, just extemporizing with you. Yep. So the way, so this is, I've even got paper saying like, what do the no free lunch theorems really mean? People,
Even though me and my collaborators, when we produced the theorems, we were very, very explicit about this. People misconstrue their significance. The simplest version of the no-freelance theorem says that if you have a uniform prior, then all algorithms perform the same.
But as we very quickly then follow that in our papers by saying that we can extend this by saying if you have a uniform distribution over priors, then all two algorithms do the same. So we are not, and we do not, we're very careful to say we do not advocate that there is, that the prior is uniform,
We're not advocating that actually you're allowing priors to vary, but that there is a uniform distribution over priors. That's not our point. Our point is rather that a uniform distribution over priors is a mathematical tool by which one can actually formally prove what Kurt just intuited.
That is the significance. It's not that we are saying there is a uniform prior over priors, or that there is simply a uniform prior. Uniformity is a tool by which we are using to quantify the fact that if A beats B in some situations, then B beats A in just as many, and we're using this uniformity to be able to quantify what that just as many means.
So many people will say that, oh, well, I don't think that the actual prior when I'm doing my machine learning algorithm is uniform, therefore no free lunch is irrelevant. No, because you've got to justify whatever the implicit prior is that you do use and no free lunch is what it's telling you is that they're going to be just as many other ones for which you're screwed as for which you succeed.
That's what it truly means. Its significance extends far beyond the case. It doesn't suffice for you, the naysayer, to simply say, I don't agree with the uniform prior, therefore no free lunches are relevant. You're missing the point of no free lunch when you say that. It's much more it's a way of formalizing and elaborating Kurt's intuition, which should be, I guess, the title of the new paper. Why should anyone care about the no free lunch theorem?
Because you are correct that people carefully say that according to Savage's axioms, De Finetti's axioms, Cox's axioms, the many axiomatizations of rational reasoning, which will lead you to say you should update your priors according to Bayes' theorem.
which is, it's not even a theorem, it's really just the definition of conditional probability. It shouldn't be glorified to be called a theorem, but you are exactly correct that yes, once you have arrived at your priors, if you then for the rest of your life using those priors want to be rational, you should be Bayesian. If by being rational, you mean you obey these particular axioms of these various people that come up with
But that does not mean anything about what those priors should be in the first place. It just means that you've got a well designed machine that takes garbage in and produces garbage out. It does that very, very cleanly and efficiently. It doesn't change the fact that garbage in becomes garbage out. You're just not adding any more garbage. What's strange is that if we have no assumptions on the underlying input space or the underlying space, then
When we're speaking about strategies, if I'm to think of evolution as employing strategies or organisms as employing strategies, it seems like what your theorem would state is that evolution would have equally produced any strategy. However, when I look around, it doesn't seem like any is equally viable. Yep. But here's the thing. It's just like that. I'm warning at the bottom of a prospectus for a mutual fund that past performance is not indicative of future performance.
It's also true in that I can save, I've got a repeated test, it's like the sun rising. My data set is just every time in the past, let's say, I mean, it's very hard to actually quantify that natural selection, mutation with descent and so on, whether it is actually being efficient or not, because natural selection in the real world has its
Some of its true power is the fact that it is obscenely parallel computation. The number of actual different genomes that are right now out in the environment that are floating around, evolving, even if you don't include things like viral genomes, it's, you know, whatever, it's going to be far, far larger than the number of stars in the universe, you know, that kind of numbers. And when you got that kind of parallelism advantage,
Almost any dumbass algorithm is going to be able to do well. So it's not even clear that in any sense, evolution has, quote, done well in our past. But even if it has, past performance is not indicative of future performance. That's what no freelance theorem is quantifying. All those, even if it were true, if you wanted to say
Every time I ran the experiment, which can have a bit up or down, up means evolution did well in this situation, down mean it did poorly. Even if I've run that experiment huge number of times and each time I kept coming the result, evolution does well, evolution does well, evolution does well. No free lunch is saying that when we now extrapolate off training set to situations we have not seen before, anything can happen.
And that's the crucial thing about no free lunch is that it's not saying anything about situations you've already seen. It's saying things about what's called off training set new situations. It's all about generalization. It's not about if I memorize things, and I know that what I've seen is not is never going to change.
And you are then asked, well, what is it that you saw in that situation? Well, you answer and you're correct. Memorization is fine. The question is always generalization. To ask, what is it that you will see? That's where probabilities come into play. That's where Bayesian reasoning comes into play. And that's where no free lunch comes into play.
We're going to get to cross validation in about maybe five minutes or so. And as I mentioned to you over email, I'm a philosophical person. I'm an introspective person. And I was wondering with regard to the no free lunch theorem, does that imply that there's no optimal strategy for one's life? What I mean is, people, including myself, perhaps, I'm mainly speaking about myself, have regret over the past and squandered opportunities, feeling like
feeling like I didn't implement the optimal strategy, I haven't lived up to my potential. Now, is that a valid concern? Or is your theorem nihilistic, where it doesn't matter what you do? It doesn't seem, it seems like it matters what people do. So it seems like there is something to be said about maximizing on one's potential as ill-defined as that sounds. 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, Rothies, 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
All lowercase. Go to shopify.com.
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. Visit hensonshaving.com slash everything.
If you use that code, you'll get two years worth of blades for free. Just make sure to add them to the cart. Plus 100 free blades when you head to H E N S O N S H A V I N G dot com slash everything and use the code everything. When I was younger, when I had just come up with the no free lunch theorems in my scribblings to myself that concern philosophical matters,
I would have something that I eventually abbreviated as TP. And that came, that was an abbreviation of the problem. And that referred to given no free lunch, what in the world do I do? Um, how old were you? Oh, um, actually my first intimations of the problem.
were as an undergrad, before I formalized the no free lunch, when I was worrying over these things. I actually did not fall into despair. Perhaps I should have. Every time I walk across the street, I am behaving with my life on the line as though no free lunch were wrong. Every time I do anything. So I would say that in a certain sense,
far more deeply than feelings of regret. No free lunch theorem has to do with how I tie my shoes. It's how do I even exist? And how do you act? How do you act? And so I look at this and I say, okay, now what? Um, and here's how to make it even worse. And another thing, I guess,
Through my career, I've not thought of myself this way, but I tend to have these no-go-ish kinds of bent of finding out. To me, it's one of the most glorious... I don't know, this almost sounds Buddhist. To me, it's one of the most glorious aspects of... I wouldn't even use the word existence. The more that I am squashed flat, the more that I am seen to be a teeny little restricted nothing,
That to me is the height of beauty. That to me is the most sublime and rapturing thing to sit on your back in a meadow at night and to just look up and to try to project and realize all those stars are three-dimensionally configured and just go through like a powers of 10 kind of a thing to fully to try to fully appreciate my own insignificance, my own limitations. That to me is the most beautiful liberating thing I can imagine.
and these results all have that character to them that they are in some sense precisely because I am not even a gnat on the back of a fly that produces a sense of joyful glee almost like a boy or a child does oh I am liberated I don't have to go to whatever but there anyway this is a preamble to another
set of results that are along the same lines, really, which are what are sometimes referred to nowadays as the block model of cosmology, which is the following. So I had a paper that was published, I guess I was a grad student at the time, at what was called the International Journal of Theoretical Physics, called
So okay, there's various notions of what's called the arrow of time. There's one of them is the second law of thermodynamics, and another one is what is sometimes called psychological arrow of time, the notion that it feels to us as though time is going forward. Paul Davies had a very nice book on this called the arrow of time in like 73, which is, it's not for popular audiences, it's very much a
for students of physics. But there is no such... When we talk about movement in space, we talk about position against time. To talk about movement in time, you would need to have two dimensions of time. We don't.
The future and the past in, this is the lesson of quantum mechanics really and of general relativity, they are just as real as the present. There is no physical variable that can actually distinguish these things from one another. Okay, but so why do we have the illusion that time seems to be moving forward? And it's actually can be drilled down to being a two-step process. First,
Notice that you as a human being, you have memory of your past, but you have none of your future. A more precise way of phrasing that is that you can retrodict what happened in the past more accurately than you can predict what will happen in the future. You then drill down and try to say, well, why might that be the case?
And eventually you can actually make arguments which are only informal to this day that it's got to do the second law of thermodynamics. But the reason I bring all this up is that in a certain sense, it's only worse because whenever you, Kurt, are making these arguments about what you've observed in your past and your memories and your regret, that's actually only retrodiction. You don't even know for sure that those things even happened.
You only exist in the present or simultaneously the presence in all the futures with no, certainly no kind of a free will that you can somehow supervene into all this, supersede all this and change the way it is. The book has been written and you are simply a single page in the book remembering the pages in the past as best you can and wondering about the pages in the future.
That is all on top of no free lunch, the restrictions of inference devices, and so on and so forth. Even before you get to the sort of traditional concerns of metaphysics about what is reality, there are all these nuances that prevent you from actually in a certain sense even getting to be able to sit up and ask that question in the first place.
It's a very, very strange thing and I'm not sure how to actually deal with it. Okay, so we're going to talk about free will, which it sounds like what you just mentioned is against free will, yet you're one of the largest proponents of the existence of free will that I know of. So we'll talk about that soon because I'm sure you're defining free will differently than the way most people think of it.
Now, you also mentioned that this merriment that you have as an adult and perhaps even as a child is when you think about your insignificance. And just like you see people like Sabine Hassenfelder as intellectually posturing when they dismiss free will,
When I hear people say what you just said, also Lawrence Krauss expressed a similar sentiment that, well, the more I think about my ephemeral nature, the more that I feel like there's meaning because there's a temporary nature to this. I see this as a form of rational posturing to show like, look how much I don't need my importance, but I'll explain why and then please push back. I just want to hear what your thoughts are on this. Because
If one takes that argument seriously, well, look how limited our life is and then that makes it more meaningful. Well, why don't you just kill your child, make your life, make your child's life smaller? Is that not more meaningful? Or maybe it's that we have to say that because the alternative that you're extremely, extremely important and every single thing you do matters beyond
what you can imagine. Maybe that's way too crushing. And so the alternative that, okay, great. I don't have this burden or onus that's freeing, but it's not actually a rational assessment that led you to that freeing feeling. It's more, well, the alternative, I don't, it's a horror show to think about how meaningful and important my life is to other people and to the world. And perhaps, well, let's hear what you have to say to that. I never use the word meaningful.
That those other people, they do then make the argument that though, okay, therefore I must make meaning out of what I have, and that then imbues it with more significance. Uh-uh. No. You are not. This is the rhetorical you. You have no meaning. You have no significance. And when I say that this is liberating and induces a feeling of glee and merriment, that's an aesthetic.
reaction so um it was when i first came to feel that it was shortly after i first learned quantum physics this was when i was a teenager and i think people misconstrue the true significance of quantum physics they talk about the uncertainty principle wave particle duality yes that's all really cool but the
more important aspect of it, I think, lesson of it, is precisely the establishment of the Tegmarkian view that all that a reality is, is a mathematical system, that there is nothing outside of that, that in quantum mechanics what you find when you do it
that you cannot do any of your predictions properly unless both the observer and the system being observed are both in an inclusive way treated quantum mechanically. Wigner and all these other people, this is almost like this is an Everett view which is the one that's really adopted whether they want to admit it or not by everybody working in like quantum computation and so on. There is no external observer. Everything
Sean Carroll also is an effective proponent of this. There is no external observer. You don't need to invoke something beyond just the mathematics itself to be able to give all phenomena that we can experimentally assess. Once you do that and you start to realize that, oh, well, ultimately one mathematics is just as legitimate as another one. So if a reality is just a body of mathematics,
That means other bodies of mathematics are just as real as this one, as physical universes. In any case, it was contemplating that and the associated feelings of insignificance that this was an aesthetic response of mine. And it was only then, up until that point, whenever I thought about Zen Buddhism, haiku, koans, anything like that, spiritualism,
I'd always been very dismissive of it, very much of a standard. Westerners, maybe you're going to be respectful and public, but it's got nothing to say for you personally. I am still very dismissive of some aspects of Buddhism. For example, Buddhism is supposedly all about the irrelevance of the self. But in fact, in practice at least, Buddhist practice is all working on the self.
So that I get to understand how irrelevant and that's a massive contradiction in the foundations. But here's the thing that happened. Once I started to try to internalize this lesson of quantum mechanics and notions along the lines of no free lunch, now for the first time I could actually appreciate aesthetically some of the products of Zen Buddhism.
Zen Buddhism watercolors, Zen Buddhist gardens. I started to be able to appreciate koans not as deep spiritual work, but as aesthetic works, as works of art, as closer to poems, somewhat like haiku are normally perceived. Yeah, I saw you wrote some haikus on your website.
I don't actually characterize myself that way. When I was
younger, trying to figure out what I want to be when I grow up in a very practical, concrete sense, what I want to have as my major and so on. I would say things to myself, like I approach science as a philosopher. I approach philosophy as a mathematician. I approach math as an artist, things like that. That's interesting. Like an orboros kind of a thing.
I have no delusions that I am... Well, okay, another kind of... This is interesting. This conversation is going much to sort of lessons of life kind of thing. One of the tragedies I find of people's lives around me all the time and what I very intensely, viciously am on the guard with that I do not indulge in myself is inflated views
of your own abilities, your own worth, your own capabilities, what you are providing. This has got nothing to do with, oh, my beautiful insignificance. This is just I've seen tragedies where people think that they have the capabilities to engage in something and they change how they're living their lives along those lines and really know that was a misassessment of what they can do. And they began
and therefore there's frustrated expectations and so on and so forth. Yes, your reach must always exceed your grasp, et cetera, but if you're not really aware of what that grasp is in the first place, you're running a real large risk that you're going to be unsuccessful in that reach of yours. The reason I mention all that is I'm very aware of what my own limitations are. I personally do not view the
material on my website as necessarily worthy of comparison to those of a real artist or poet. I don't even necessarily, I don't view myself as actually all that smart as a scientist. There are many, I am aware of what my relative strengths are, but in terms of just raw IQ,
That ain't one of them. There's people out there who are far smarter and who are far better as scientists in that sense. Well, you have a clever saying on your website that you're not intelligent enough to be a mathematician nor careful enough to be an experimentalist. Is that correct? Yeah. Yeah, yeah, yeah. That's from my website. Yep. Okay. Now you were talking about how quantum physics allowed you to appreciate the Zen Collins. And then you're saying aesthetically,
and not spiritually necessarily, and I interrupted you, so do you mind finishing that thought? Yeah, so as I say, I don't view... I don't think... It's sometimes construed as a philosophy, and I think as a philosophy, Buddhism has nothing to say. I think it, like all religions, are just vacuous fables that humans come up with to try to comfort themselves against
The big vast expanse, which is what I actually for whatever weird reasons in my psyche, I actually find so beautiful. It's most people, most reasonable people react with some screaming fear. And I think that's the basis of a lot of religion. You can also actually, there's a lot of very interesting literature in the economics field, economics of religion. It's actually very, very interesting as a topic of study, what makes some religions succeed and other ones not.
And of course, it goes back, many people have worried about, have wondered about, as evolutionary theories of why humans have religion. So construed as a religion, I view Buddhism like the other ones, as at best epiphenomena, side effects of other processes that go on and certainly do not have any legitimacy as philosophies concerning reality, whatever you might want reality to be, as bodies of aesthetics, as art,
That is what I could suddenly appreciate when I started to understand quantum physics and no free lunch and so on. I don't know why I've not really drilled down on this much, but somehow the, this is probably almost like a convergent evolution, that the evisceration of the notion of the worth of the self that I arrived at it, my
processing based upon science, induce the same kind of aesthetic reaction in me as the evisceration of self that Buddhists undergo for their reasons, inducing them. So it's probably got more to do to say about how human psyches react to that cognitive assessment
Even though the ways that those cognitive assessments are arrived at is very, very different, and I would frankly argue long and hard that one is legitimate being grounded in science and the other one is not being grounded in obsession with self, I would say.
anyway at this point there were some technical difficulties with zoom and the audio takes a decrement and improves in a few minutes the view that you outlined about religion reminds me of marks's opiate for the masses where it's just there simply to placate the society and i'm curious i don't want to get on a tangent about religion nor
Well, I don't want to get on a tangent about religion, so if you can briefly comment on this. Why do you think it is that many of the religions have a huge element of hell in them, which is actually terrifying? So if it was just about making people feel better, for me, when I made a conversion to being an atheist, and I know this from virtually every atheist I've spoken to, there's a huge moral weight that's lifted from you when you convert, quote-unquote, to atheism. So it's not as if one is happier necessarily
simply because of the teachings of the religion. I don't think it's as simple as that, and I'd like to hear what you have to say about that. Yes, no, I fully agree with you. For me, there was never a moral weight, personally, for what it's worth. There's never a moral weight being concerned about after a lot during that. There's actually some very interesting literature that those religions which are most successful in the history. 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 Brooklyn. 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 shopify.com slash theories. A religion that this has to do with the economic religion literature are those that make the most demands of their practitioners. Religions that don't make much in the way of demands as opposed to those that do demand things like you go on a missionary
or that you give a large tithe, or that you very, very much change what your daily routine is. Those ones tend to be much more successful. So you've got to, the religions that are successful are those that are particularly demanding. So there are these kinds of counterintuitive aspects to religion. As far as the one about whether, so Marx, of course, with the opiate for the masses, he was instead talking about its large
sociological implications and how he thought it was being exploited as opposed to personally why someone might be taken to be um vested in the notion of an anthropomorphic religion as opposed to spiritualism if you mean by spiritualism something closer to aesthetics there one could say that i personally am very spiritualist it's much more a matter of the supernatural
why and the aspect of things. Now it's not, it is not certainly, the reason that people are inclined to supernaturalism in general is not simply as a comfort. There is that element to it, but certainly many religions do not provide any comfort
Traditional Catholicism, for example, was you're born into sin and you might be screwed up earth and that's the way it is. So it is far more complicated than the kind of snide comments I was making. I think that for many people in modern society, the role it plays is a comforter. Historically and cross-culturally,
Religion has played many, many other roles, wound up with supernaturalism, which is itself wound up with our tendency to overgeneralize the environment. Things like if you are evolving, here's one simple argument in favor of why we are prone to supernaturalism. If I'm evolving human beings in the veldt a million years ago, hominins,
I can either have them be overgeneralizing, finding patterns where there are no patterns or undergeneralizing. If you're overgeneralizing, you might think that a particular rustle in the grass always means there's a tiger there when that's not true, but you're going to be a lot better off than someone who's undergeneralizing so that when there is a pattern that's related to whether there's a tiger in the grass,
They don't draw the proper conclusion and therefore they get eaten. So you can make those kinds of arguments as to why we are prone to supernaturalism in the first place, why we do tend to find patterns where there aren't any there. And the deep psychological compulsions that have resulted in religions, many of which have no
emollient characteristic to them at all, exactly as you're saying. Those ones are not the ones that I see so prevalent in modern American society. So when I made my statements, which were snide and gloob and off the cuff, I was referring much more to modern right-wing religious sentiments.
where I do think it's, you know, there are all these studies that say that those who are very religious in the modern sense of the word, going to, engaging in, in terms of the modern church, that they live eight years longer and things like this, because the stress levels of them are so much reduced. So I'm much more referring to that cohort and their approach to life.
At this point, there was a minor break, and we started to talk about Sabine Hassenfelder. There was no exact question, otherwise I'd leave it here. It's simply the topic of his thoughts on Sabina's video about super-determinism and free will. I guess, so she's got an axe to grind. She wants there to be hidden variable models.
They want reality to be that way, she and some other people. And so they are searching around for any way that they can to try to find squirrel room where Bell's, in her case, Bell's theorem notwithstanding, you can still have a hidden variable model. And I just don't see the point. It's motivated reasoning.
um to and she says it's super free will as that i view as vacuous because that's just what she is arguing for is just what i would always have considered to be sorry is super i forget what her precise super determinism super determinism that's right super determinism it's kind of like a super version of no free will super determinism is just what i had always understood to be standard physics straight out of the box um
I'm not quite sure why she is trying to make a case, she and others, make a case that's somehow something new. It's much more what they are trying to... It's a strange thing what physicists mean by free will, like in Conway's free will theorem and so on. It's kind of distorted from what the term is normally meant in the larger, in the parlance of philosophers more generally.
And anyway, as I said, I think it's motivated reasoning. How about we define free will or you define free will and then tell the audience what your inference theorems have to say about free will, you'll have to define free will and also define what your inference theorem is and then what the inference theorem has to say about free will. Okay, so um, first a warning, there is a
Unfortunately, it shouldn't have a particular name as one of the rhetorical fallacies. It's, you know, the ad hominem fallacy, the this fallacy, the that fallacy. For people to take a term that's in very common use and is very controversial because of all these reasons, they want to establish that it's true. So they redefine it in a certain way that they can establish that it's true.
that redefinition, all of the original reasons that it was so controversial, they are now defined out. And what you're left with is actually a very different kind of a concept. And they now then herald it and say, oh, look, I've just proved such and such where what they've really done is redefined it in such a way that they could prove it. But all the reason people were interested in the first place is now out of their new definition. So it's a bait and switch.
The fallacy is called Mott and Bailey, I believe. Mott and Bailey, okay. How's that spelled? M-O-T-T-E and then Bailey's as the drink, I believe. Essentially, you take some fact and then you're straw manning it and you use the straw man to disprove the larger one. Yeah, or to prove it as the case might be, Mott and Bailey. That's nice. I've got to look that up later. So, continue. So, yes. So, for example, okay, so certainly if you're familiar with the concept,
Um, the notion of God has been treated that way. Um, you know, many people would say, Oh, Einstein really believe in God because, you know, he said the laws of nature and laws of nature is God. Um, but of course, the reason that God is considered to be a controversial topic is because it's a banal promote, promote. It's this dude in the sky with a big beard who's deciding who's going to win the football game Sunday night.
And that's the one that's controversial and so on and so forth. And Einstein wasn't saying anything about in favor of that particular one. So it's a little bit of a Mott and Bailey's. And frankly, I think that he has done some great work and he's a good friend and a great guy. But I think that Dan Dennett was recently involved in this with free will.
He wrote a book in which he tries to now claim I've proven free will. And in essence, he's defined away its aspects that made it controversial in the first place. So what do I mean by free will? There's one aspect to it, which would be that of the people who write dictionaries. What does it mean in common discourse?
what I think free will means and common discourse has something to do with a means by which it is possible to abrogate the laws of deterministic science in such a way that a human's cognitions are not subject to those laws. So that what is going on in your head right now is not
some very complicated, too complicated to calculate, but nonetheless unviable function of what it was you were thinking 10 seconds ago together with the stimuli, sensory stimuli from the environment, that somehow there's something else
that is not subject to any laws of science in some sense. It's not even subject to sarcastic laws. It's just amorphous and ill-defined. And so certainly all of us have a subjective experience of such a thing. But the proponents of free will in this sense, which is controversial, would say that no, that's actually an objective truth. I don't see that there's any room for that, to be quite honest.
So in that sense, no, I don't, if that's what, if we do adopt that kind of a definition, which I think is in accord with the way it's used in common parlance, no, I don't see it, there's any room for such a thing. You can, you can maybe kind of get close and Scott Aronson, again, he's involved engaged in a Martin Bailey's in this particular paper of his,
But he and even worse Martin Bailey's as Seth Lloyd, they try to make the case that Turing machine notions of computability provide a way out in that they say that, yes, it might be the various flavors of the argument, but one version of it is, yes, the present state of your mind is a deterministic function of what it was and so on, but it's an uncomputable function
in some technical senses. Yeah, Stephen Wolfram says something similar where he uses his term computational irreducibility and says, well, yes, there are outcomes that are computationally irreducible, and that's where free will may live. But ultimately, you don't have control over the laws. It's still a deterministic step by step algorithmic process. Yeah. And in a certain sense, it's kind of interesting because one might ask in terms of the laws of science, of physics, and how the universe actually evolves
Is there any room for anything besides deterministic or stochastic evolution? I can write down the definition, you can go look it up in math textbooks what a stochastic process is, and you can look up in math textbooks what a deterministic invertible process is, and you can actually formulate the one is a special case of the other. Is there any way you can even in theory imagine
science that would not be either random or deterministic, as with anything else it could possibly be. And there are notions in the original motivation for Komogorov, in terms of his work on Komogorov complexity, and many other people who are wondering about intrinsic randomness, what does complexity mean, what does random really mean. They came up with definitions which, something that is random,
but is not random, it has no probability distributions. So it's interesting to think about whether that might provide any kind of a, whether that's a difference without significance or whether there is a significance to that possibility that physical systems might in essence violate the physical church Turing thesis and have
this kind of an uncomputable character to them, which is not random. And it's not deterministic, but it's something different. And whether that has any kinds of consequences, even if it's true. So that's interesting in almost like a philosophy of science kind of a way. But it really is, I would say, irredeemably far from what people mean by free will to really say that it's relevant to the discussions that people have about what free will is. I think for them,
the vast majority for what philosophers have meant by it for millennia, it's no, it's certainly nothing to do with comogoro complexity and turning machines and so on and so forth. It's simply the notion of their this supernatural in the literal sense of above nature, above the laws of nature, which is inherent in me and my soul, and I cannot even define it because it is something that's non-definable.
and that allows for superseding of the laws of physics when it comes to my own brain. That's what is sometimes meant. I think that's most fully what is meant by people who are talking about free will, have been talking about historically, and I almost view it as a non-starter. Okay, so to make a summary for the audience who just wants a clear answer,
do we have free will or not? In that sense, I don't think it's even meaningful to say that we do. I don't see that you can, if you you're almost by that definition, one is almost defining free will to be something that's undefinable. And it are and is supernatural. And I can't even almost react to that. If it's not definable, I can't say anything. And if it's not
And I don't see there as a room for any supernatural things in all notions of reality. Circling back to a previous theme, one of the things that I find most beautiful about our inconsequentiality is that the whole notion of consciousness, it's a post-hoc fable that we cook up to try to give us the illusion that we're in control of our own brains. There's all these fascinating phenomena going back like Ben Libet and so on.
demonstrating that it's at least logically consistent, in fact seems to be the case, that all the things that we think we're in control of, we're not in control in any meaningful sense of any of them. We're making up stories after the fact. Corpus callosum severed people going, you know, that whole literature. It's fascinating that when you ask one half of a brain why the other half is doing what it is,
Let me steelman the opponent. So with regard to the supernatural
Most religious people would say that there is no supernatural... Well, let's forget about most religious people. There is the view that there is no supernatural even in the religious domain.
that god operates by natural law in fact god is natural law and then we may say well god is supposed to be a man in the sky etc however that wasn't the case up until about 200 or 300 years ago at least it wasn't the majority case and mysticism christian mysticism in particular was extremely rampant where you can't define god attempts to do so are actually
And I agree that this is what free will should be defined as. There's somehow
us interacting with nature that is different than just the physical laws we know it. It's a metaphysical statement when we think that we have free will. And then you were saying, yeah, it would have to be supernatural because it's above the laws of physics. Well, Raymond Smullian, I'm sure you've heard of Raymond Smullian, the mathematician, he had many writings on free will, and he would say,
What defines you? What's the difference between you and natural law? You're a product of natural law and we tend to think of free will as not existing because the outside world is more powerful than us and somehow despite our will, it just overcomes us. However, that would mean that there's a distinction between you and natural law. Well, that's illusory. So in some sense, you are natural law. So in some sense, well,
But so what is the meaning of free will? Is it just a psychological phenomenon that people have this sense of agency? Yeah, sure, that's definitely true. That half of a hemisphere thinks it's in control of the other one. And yeah, there's all kinds of... Is there a psychological phenomenon just like people can get depressed, just like people can become obsessive compulsive, just like much more commonly everybody, yours truly as well, feels, has this feeling
Yes, it's interesting maybe to look at some of the neurobiology behind some of that. But if that's what the statement is, then yeah, it's definitely true. Let's go do a survey. Do people feel this way? Yeah, they would do. If instead it's a statement about metaphysics, to say it's natural law, I would say as a Martin Bailey, you're ending up with nothing. So then what do you mean by free will if you don't mean this part of neuropsychology?
If you're going to say that is just part of determinism, what is it? There's nothing left to define it as, I would say. Even the Libet experiments, you mentioned the Libet experiments. Libet says that his, and I would like to interview him. I don't know if he's still alive. I think he's not. And by the way, those are now controversial experiments. Yeah. And Libet himself was a proponent of free will and says that his experiments have been
Interpret it incorrectly. I find it to be controversial whenever someone says free will exists or free will doesn't exist I'm curious like how are you making that without a plethora of assumptions? That's also a metaphysical claim. We both agree It's a metaphysical claim and maybe you don't agree with this, but I don't think science makes metaphysical claims I think science is an instrument and if anything, it's Instrumentalism here's what science is just if then statements it says if you do this this happens doesn't say behind this is
So here's a way to cut the Gordian knot. So I think we're agreed that free will, I view it as either
Total logical in the sense that if it is a psycho neuropsychological phenomenon, then people do have that feeling of free will. Yes, they do. Just like people have a feeling of being happy or sad. And these are things that you that neuro biologists can put people into fMRI scanners and try to drill down. If it is instead use a metaphysical statement, then I don't know what it means.
And that might, in fact, some people might define it to be not as mean, unless we try to go across some Turing computability and so on. Now, in terms of whether science and metaphysics,
What I mean is, does science ever make a claim to ontology? Instrumentalism or not? Yeah, yeah, yeah. I mean, it's a subtle thing. So instrumentalism, one way to interpret it, and like all names of schools and philosophy, it's got various flavors and proponents and so on. But it's the view that there is some ill-defined reality out there which produces effects on our experimental apparatuses.
And what science is, is a way of trying to actually relate the patterns in those effects in our apparatuses and maybe make predictions about other patterns we might see in future apparatuses without saying anything about this underlying physical reality. That an electron is a construct for describing these patterns and there is some other external reality out there
that the word electron is not necessarily relevant to it at all. And certainly, since science in its current instantiation is progressing, it could not make any claim that it is reality. But here's a different perspective, which is implicit in what's called ontic structural realism, also in some of Max Tegmark's more careful work and like his Analysts of Physics paper,
and so on rather than his pop work, and which also underlies some of the ways that my collaborators have been pushing the formalism in noise deterministic reasoning, which also relates to what I was talking about before, the problem. So here is a very simple, not that simple, but here is a hypothesized
Mathematics, which in well and fell swoop explains all explains away all of the conundrums of philosophy together with science. And here it is. It's only mathematics all the way down. That there is no external reality. Everything is just a mathematical system.
and that one mathematical system is in fact just as legitimate, just as real as any other one. It's that the reason a chess piece on a chessboard feels the laws of chess to be real rather than the laws of checkers is simply because it's having to right now be viewed as a token in a game of chess, but it could just as easily be in a game of checkers. From that perspective then, there is a mathematical system
which would have as some of its equations human beings looking out at experimental universes around them and noticing patterns which might be approximations in some sense or not of this underlying mathematics, but underlying it all is just that mathematics. So it's somewhat stepping to the side of saying
It's a single monolithic edifice that explains both philosophies, humans, our instruments, the patterns, and underlying reality. It's all just one single thing, and there is no room for anything else. It is informed by science, but it is not actually somehow something that is proven by science.
You know how you said that you're unsure what it means for free will to have the metaphysical notion of free will? What the heck does that mean? I'm unsure what it means that mathematics is the ultimate reality. I don't understand Max Tegmark's ideas and I'm also unsure if you're claiming that you like those ideas and that's the one that you see as fitting reality most. I view it as in one fell swoop solving
Well, it doesn't completely solve the, the conundrums because for example, laws of mathematics, um, girdles and completeness theorem says there is no ultimate, um, uh, unviable consistency in the laws of mathematics themselves. So 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 slash theories, all lowercase.
Go to Shopify.com slash theories now to grow your business no matter what stage you're in Shopify.com slash theories. It eats its own tail. And if one then says noisy, um, deter, um, deductive reasoning, you then get an even worse this oration of the foundations of that statement that all is mathematics disappears. What do you mean? Cause there's no such thing as a mathematics, but here's the sense in which
So those are extra extensions of the underlying first concept, which are very, very important, but before getting to them, before invoking Gödel and so on. I view ontic structural realism, Tegmark's elaborations, as explaining it all in the sense of quantum physics, that you, Kurt,
are part of a wave function that involves the screen you're looking at and everything else. It's the wave function of the universe and all aspects of your neurobiology and therefore your subjective notion of free will. All aspects of your scrutinizing the results of experiments involving what in your lab notebook is called an electron. All of that is arising
as simply solutions to Schrodinger's equation and that there is nothing else besides that. There is no underlying concrete reality that your mathematics is describing. It suffices to just dispense with it. You don't need that concrete reality. All those mathematics is not the mathematics we currently have in our textbooks. But that's fine because our textbooks themselves are simply
some of the characteristics, the projection operators of the quantum physics of the wave function of the universe. Have you ever heard the phrase, we shouldn't mistake the map for the territory? I don't believe I've heard the phrase. Yes. Or it's actually it's sometimes the map and the territory. Yeah, don't mistake them. Right. That's right. What did I say?
You said the map and the territory. No, I meant the map if I misspoke. Yeah, the map and the territory because in that case there's stuff that's not in the map that is in the territory and you can experimentally ascertain that. This point is that the map that we currently have is not the ultimate map but the ultimate map so to speak. The Ur map has in it, this is almost like a Borgesian kind of a
How do we know that all is math?
We don't. Remember, I was very, very careful the way to phrase this, that one perspective, which is consistent and simple, it doesn't invoke unseen, as since going back to Kant, at least, Western philosophy has been prone to doing, or Plato, for that matter, with his Platonic ideals, it doesn't invoke all those unseen, because it's pointing out
that ultimately it's by flipping the perspective to say that we are phenomena generated by it rather than it is something external to us that we are trying to understand. It's a flip of the perspective. You start with the it and we are manifestations of it, projection operators as they call them in physics, of those pieces of the it. Rather than starting from ourselves in almost a Copernican and pre-Copernican
pre Copernican view of the earth as the center. We are the center trying to infer the outside territory. No, we are ourselves just one small little map in this much bigger map. And from that perspective, you don't need anything outside and everything is explained. All of these conundrums, you know, no free lunch. How does that allow me to even get going?
Yep, you're correct, little boy, that no free lunch doesn't allow you personally to get going. But that doesn't change the fact that here is a way that here is a, I'm not even sure what you would call it. It's not saying it's a reality. It's not saying it's a hypothesis. It's just a edifice of
interrelated explanatory phenomena of mathematics, which we don't even currently have, but which would actually do a jiu-jitsu move of all of our philosophical conundrums by simply saying that they are what come up in the neurobiology of these particular little systems that are being generated as part of the solution to the Schrodinger equation when they try to look outward at things outside of themselves.
Sir, how would you respond to my thought, which is it sounds like we have quantities and then we say, well, all that exists are these quantities and not everything is quantitative. However, it seems to that's what I mean when I say that. Sorry. That's my that's my rendition of all is math. However, it seems like there are large classes of our experience which are qualitative that don't lend themselves, at least not readily.
To me, it sounds like a large statement to say all of what we think of as qualitative, so anger, the sight and experience of red and blue, etc. So you're talking about qualia? Yeah, yeah. Not everything lends itself to a quantitative analysis. Okay, but are you specifically talking about like charmers and these people, their notion of what they call qualia? Sure. So I view that
conundrum is ultimately, ultimately vacuous. We think that we feel something. I mean, famously, Dan Dennett had a great put down of the paper. What does it feel like to be a hummingbird? I think it was that very famous paper. Dennett eviscerates it. But ultimately, the question, what is it
I don't see if there's any room for that question any more than there's any room for our belief, for one hemisphere's belief that's in control of the other one. That belief is wrong. It is demonstrably wrong. Neurobiology can presumably in some future incarnation explain why that hemisphere would come to that conclusion after first neurobiology has to even
figure out what it means to say that a particular piece of brain tissue is coming to a conclusion or not. But it is wrong. It has got nothing to do with reality, that qualitative experience of that hemisphere being in control. No, it's not. But we can now expect this future incarnation of neurobiology would be able to explain all aspects. Whatever question you might ask,
of that subjective experience, it would be able to provide the answer that there's not necessarily any room for anything else. You said this is yet to be demonstrated. We just see you see this as reasonable. Yes, neurobiology has not done this by a long shot. But I'm just saying that what's consistent with everything that we now know with everything that we can think
Let me tell you my thoughts quickly, I'm sorry to interject. When it comes to neurobiology and then we have neural correlates of certain qualitative experiences and then you can say those correlates are quantitative, so we found an association between quantitative and the qualitative, so therefore the qualitative ultimately reduces to the quantitative. Those ultimately rely on many, many, many, many, many reports of subjective experiences. So for example, I'll have to tell you
I flip the question on you.
I wouldn't even say these are claims. My statements do not have to do with current neurobiology, or even necessarily that neurobiology will get there. It's much more the following. It's a challenge to you. If, in fact, there is something inherent about the subjective that
would not be reducible to mathematical properties. Given, especially in light of the Bin-Libet experiments, the split-blades and all the other related zombie within kind of phenomena, how could one even speak about it without reducing it to being a
mathematical thing what is even it how can we even say anything how can we have a discussion with unless it's going to be defined precisely and i don't know how to define something precisely without defining it precisely it's it's a challenge how do you other than sitting around rather than having an aesthetic experience which we can have how do we actually engage in it in a reasoning manner
Let's have fun here. When you're saying that you don't know how we can speak unless we define our words precisely, well,
You and I can speak about viewing art. For example, right now you've got a guitar back there. I presume that you play. I don't wish I could but I could imagine that we could
have an interaction concerning music. And if in a different life where those particular regrets returning to a previous theme did not exist in me, I might be able to have some facility on and we could do it. We could jam together or whatnot. That would be us communicating. It would all be legitimate. It would be valid. It's a different form of communicating from when we are trying to ascertain jointly
using reasoning processes, something concerning, I wouldn't even use the word existence because that by itself is a fraud exercise, but so it's a different realm of discourse. You and I can very definitely talk about what does it feel like to experience beautiful music,
What does it feel like to, I mean, red is a simple minded one, but to experience beautiful art. I have often described how to me, uh, I have been ambushed in my life by paintings, um, turn a corner in a gallery or a museum. And literally the phrase that keeps coming to my mind, I keep repeating is I'm being ambushed. I'm just knocked flat. I'm stunned by it.
It's a physical reaction by that actual painting. To give you a simple example, it's things like Monet's and Van Gogh's. Because it's so beautiful that you're in awe? Cézanne's in particular. I never particularly appreciated Cézanne until when I first saw one in the flesh, then it ambushed me because there's so many details of it that just don't reproduce well. When you read about a Cézanne piece, Simon's Scent, I think it was,
in the art books, it's all intellectual that he had this effect on art history and so on and so forth. Okay, so you can read it like that. But then when I saw it, wow, I was knocked out of my socks. And then more generally, I can talk with you or anybody else about how for me to actually judge a painting, the way that I judge it, the process I go through is what can it teach me?
If I myself were trying to redo that painting and try to change some things, would I actually learn, no, wait, you can't do that because of these aspects that you weren't even sensitive to. Does it have its own three-dimensional reality, that painting, or is it just a flat caricature? I don't mean necessarily sense of perspective. I mean, does it exist? I understand. So we can have that discussion.
Does that have anything to do with Schrodinger's equation? Well, I mean, you know, ultimately, if you want to flip things around, yeah, sure, David, you can have a different set of discussions where you're saying that these aesthetic experiences in your head are just neurobiology and so on and so forth. Sure, you can. And if you want to, go off and do it. But that's a different conversation we're having from talking about that art piece, Ambushing Me. It's irrelevant to talk about it in terms of Schrodinger's equation. Sure, yes, but who cares?
That's got nothing to do with the actual discussion we're having about what art pieces have that they can teach you. Conversely though, if I am instead in a different mood, I've woken up on a different side of the bed and I'm discussing philosophy, formal reasoning, what is the ultimate nature of reality, then
all of the other stuff about being ambushed it just the question gets reduced to is all of that in theory at least explainable or is there something there that is not even in theory explainable as being a meta as an epiphenomenon an effect of mathematical Schrodinger's equation. It's a completely different way of of engaging with the same subject domain.
And what I'm saying is that when I'm over there wearing my metaphysics hat, if I'm just going to push metaphysics to be what I personally, David Walport, I'm interested in, which is all the way down the road reason I see, I see. Okay, I see. So let me put it like this. It sounds like what you're saying is if one is making a claim that
your brain, David, that you have some free will and it supervenes the laws of physics. Sorry, it doesn't super, the physics supervenes on that your free will somehow determines the laws of physics, at least momentarily enough for a neuron to fire differently. And so therefore you have free will. So you have some, your free will has some ontological status. And then you're saying, well, okay, where along the chain from electrons to cells to
Neurons to, well, neuron is this type of stuff, to the whole brain configuration, to your behavior, where on that chain was physical law broken? Because you can look at any given one of those chains and you can say, well, look, there's no violation of physical law. And so if the nebulous person is making the claim about some nebulous concept like free will, you'd better show me where it comes in. So far, the laws of physics are not violated at any level as far as we've examined it. Is that what you're saying? Essentially, yeah.
As opposed to do I have a feeling of free will? Yes, I do.
okay now what I would say to that is I'd still say that we don't know and I don't think you're making the claim that you know either so I don't think that we're speaking differently let me put a bit of a monkey wrench so when we look and we examine at any one of these features so we look at let's say the electron bound with a proton we say there are no laws of physics being violated here this is all explainable with what we have currently as our models or at least we have the idea that some other model may exist in the future and then cells and so on so on it's as if you're looking individually at small parts but still
The counterclaim could be that you show me that the laws of physics are obeyed for someone making a decision across the whole spectrum and not looking at it individually across different people in different circumstances and saying here is now violated. It would be almost as if what you're doing is looking at a small billiard balls bouncing around in a small environment saying there's no time here. There's no time here. There's no time here. But yet you're saying you're saying that there's time at the higher level. There can be. There can be a merger. I mean, Phil Anderson's more is different.
You can have laws, I mean, that's what condensed matter physics is all about. Parisi just won the Nobel Prize for spin glass models, which is thermodynamic limit, an infinite number of phenomena that arise strictly speaking only when you have an infinite number of these interactors, but it's all being governed by mathematics. And so you're correct. I'm not saying that I now, and maybe not humanity ever, would even be able to fill in all those steps.
I'm saying that there is a consistent picture in which those steps are part of a mathematics, which may forevermore be beyond humans. But I'm then going a little bit further and saying, give me provide for me anything other than essentially an elaborated version of
describing the experience that you have when looking at a beautiful painting, something that is not mathematics. I don't even see what one could call free will. I just don't even understand what else there could be besides mathematics and then aesthetics. I don't see that there's
I don't see how we can be doing reasoning, reasoning rigid. Iron cloud, we know that with the things that iron clouds can come up with, given all this stuff about noise deterministic reasoning, deductive reasoning, but how we can actually come to any kinds of conclusions if we're not using logic, let me say that second order logic, that's part of mathematics.
Second order logic is part of mathematics and I don't see how we can, we can't even define the terms in something like what is called in mathematical logic a language where you've got an associated set of axioms and so on and so forth. I don't even see how we can be speaking about it in those terms. I don't see what philosophy can be if it's not that.
I see. When we're talking about how some of these philosophical claims are volutinous, they're bleary, they're opaque, dubious, you don't know what the heck it means because they're speaking ambiguously. And then you're saying, well, I don't know what it means. I understand what mathematics would mean if mathematics was fundamental. For me,
Please help clarify for me. What the heck does that mean? Because to me, that's just equally, I can understand the statement that mathematics describes what we've seen so far. But then to say that mathematics is what we've seen so far, I don't understand what that is is what is the mathematics and then when we say it's a rule, what is a rule we can play that game. So can you help me understand the ontology? Um, this is the
So first of all, assuming that we're not worrying here about things like Girdle's incompleteness theorem or the other things about math eating its own tail. Sure. We're not trying to do, for example, Pressberg arithmetic, which would allow you to avoid Girdle's incompleteness theorem or anything. That's one of the things that I find so deep
flaw or attribute in me is that the more I can be smashed on my ass by the universe realizing that I have been misconstruing my limitations for being those of the universe the more I feel I'm making progress the more I actually that's my aesthetic
I don't think I can answer your question and that to me, and I understand your question and feel your question. And the fact that I cannot answer it to me is a beautiful, stunning illustration of the fact that the deficiency is in me. It's a feature, not a bug. It's a feature, not a bug. Yup. Um, things like, um,
I don't know how much you know about algorithmic information complexities. No, I have a question about that. I would like to get to that later. Some of algorithmic information complexity, I view it as the most, the only true philosophy metaphysics, and by far the deepest that we've ever come across. Gödel was, his result was one of the most important philosophical results, precisely because, to my thinking, aesthetically, because it was a
incontrovertible demonstration of how all that we were so sure, we all thought the same as David Hilbert did, and we were wrong. And whether we can understand it or not is irrelevant. And this is a similar kind of a thing. I do feel that question you're asking and the fact that I don't understand what it means to give it as an answer,
doesn't mean that that conclusion, which is causing that question of yours, that the conclusion is wrong, it's a limitation in me. That I can't grapple with what the conclusion even means, just like Hilbert couldn't grapple with what Gödel meant.
Okay, speaking of limitations, I know you said that you have to, you're keeping an eye on the clock. So how about I give an outline of what we have left to talk about and we can determine how we're going to parse our time to most efficiently get through these. Okay. Okay. And I have, we're not even third way done so that we'll have to pick and choose some of these. Or would it make sense to have a follow up? We can definitely have a follow up. Yeah. Given the time, how about we then do
I'm not sure how you want to do the video splicing on your end. We'll just wear the same clothing and pretend it's all one session. Yeah, we could do that. Much much of my clothing this time of year tends to be the same anyway. Yeah. But yeah, but if we could then just schedule another time. Yeah.
Okay, so I'm assuming you have to go in about 10 minutes. Is that correct? Yeah. Okay, so let me at least go through the questions and then we don't have to answer them. But let me just outline to Laplace's demon I wanted to know.
about how your inference limits on inferences, limits on inference devices, what that has to say about Laplace's demon, but outline that for the audience. And then Joseph Rakoveka apparently has a substantially simplified argument involving Turing machines and free will, which like we just talked about free will, I haven't looked at his, and I want to know if you could comment like what was his argument? How is it simpler? I don't I don't know that particular one. Let's forget about that then.
Cross-validation. How can cross-validation be used to exploit the scientific method? I think that one we should definitely do before you leave. And then we've got to get to the definition of observers. We can save that for later because that's quite lengthy conversation. And then I wanted to talk about Kolmogorov complexity
And you mentioned that there are different bounds, which was extremely interesting on Kalmogorov isn't bounded. And that's intuitively obvious why it's not bounded. But then thermodynamic complexity is bounded. Well, I don't know what thermodynamic complexity is. And I wanted you to explain to me in the audience what it is. And then I also wanted to know if you know about assembly theory. Have you heard about assembly theory?
You mean the stuff of annoyments? No, no. It's recent. It's by Lee Cronin, who's a chemist. Oh, oh, Lee's stuff. Yeah, oh yeah, Lee is quite the... He is quite a character, Lee. Yeah, yeah, yeah. Okay, so he's not calling it assembly theory, is he?
Okay, logical irreversibility versus thermodynamic irreversibility, what's the difference between those? Then I also wanted to know if the way that you can quantify an observer, which is extremely interesting, is that related to a notion of intelligence? So we could talk about that later.
You also mentioned the most profound results in philosophy come out of algorithmic information theory, and that's where my question for you was, what the heck is algorithmic information theory? Can you give some examples? Why do you say that it's the most profound? And then we have several audience questions. So we have Edward Lee of Berkeley. Professor Kevin Knuth has a question for you. He has quite a few, and he also said to say hi to you. Okay, I'll say hi back.
Let me put it this way. This is some of the most
near and dear to my heart would be a glib thing. I view these as like the most profound things that any human minds can be engaging with. And I am sure that just as in so many other things, I'm just full of fecal matter, shall we say, about tons of things. And if I don't, you're helping to have these kinds of considerations and cognitive wrestlings
um uh conveyed out to a broader set of people who can join this mosh pit of wrestling with these things the more the better more power to you independent of me let me fall under a bus tomorrow but these are the well let me not but whatever but these issues i think yeah they should be engaged with and i think that's you know i am
Something I do passionately agree with, and I thank you very, very much for helping to have them be engaged with in a broader.
Well, thank you for providing me some material to engage with. Okay, now with regard to cross validation and exploiting the scientific method, why don't we get to that quickly and then we'll wrap up for today. Okay, so you said that in one of your talks, we can use cross validation to exploit the scientific method. First, what is cross validation, you have to explain that to the audience. And then how does it undermine science? Yeah, it's actually I'm not sure what the exact what was it basically cross validation is the scientific method formalized. So
One aspect of the form of the scientific method, like so many glib phrases, it's glib and leaves a lot out. The scientific method, we put far more posterior probability, we assign far more credence in a theory if it predicts phenomena that we have not already observed and then are found to be true.
So, you know, famously, for example, the Eddington and Einstein, the observation of during the eclipse and so on and so forth. So another way to say that is that if I give you two theories, A and B, and I ask both of them to make a prediction about something that neither of them has yet seen, A predicts it accurately and B does not, then we say let's go with A on making any subsequent predictions.
I'm going over and above. And so that can be viewed as part of the scientific method. Now, let's see what that might mean in the context of machine learning. Well, if I want to choose among a bunch of machine learning algorithms, one way to do it is to give them all a bunch of data and hold out some data, let them train on the data they have, and then just see which one does best on the data it doesn't have.
And this is actually done in contests, bake-offs, they're sometimes called, all the time. You know, we're going to choose between the following character recognition algorithms by giving them all of this data set, pulling out some other examples, and then seeing who does best. You can also, though, do that yourself, so to speak, by saying, I've got myself, I've got a bunch of different algorithms that I can use. I've got a fixed data set. Let me
play like I'm holding a bake-off, train all these algorithms, a part of my data set, on part of my data set, see which one does best on my held out data, take that one, now train on everything and use it to make my subsequent predictions and use that as to choose among different algorithms, just me myself as a solo machine learner, or to set parameters of a single algorithm.
That what I've just described to you is what's called cross validation. It is part and parcel of machine learning algorithms. It is part and parcel of the scientific method. It does not get around the no free lunch theorems. That in point of fact, you can actually show, and this is some really strange stuff that I would be very happy to talk about, but it'll require some care in the discussion, that the
algorithm of anti-cross validation, of choosing the scientific theory that does worst at predicting, there are just as many priors for which that is going to outperform the scientific method as the other way around. So very strange, it almost makes no sense when I just say it that way.
So what would be an example of a statement that we believe is scientifically true, but you believe it was founded by inadvertently exploiting something? Everything in science. Whenever I cross the street, everything, this is what I meant by why the no-freelance theorem is sometimes just my notes to myself, TP. Everything. The anti-reasoning by no-freelance, anti-reasoning rather than reasoning.
in terms of inductive reasoning. Let me state what cross validation is one more time simply to the audience so that some of the people who weren't able to follow can follow. Cross validation, if you've heard on a previous podcast, I believe I talked about Kryptonstein, there's a rule following paradox, so it's similar to this. Imagine you have a kid, a child, you're teaching how do you add digits. So you have a set of a thousand cards and each of them is an arithmetic statement, a statement about arithmetic. So two plus five is seven, three to the power of four is
81 and then you show the child the set of statements that are correct and you show them the set of statements that are incorrect so some of these are true plus five equals minus one and then you say and then you want to find out is the child following a rule but the child could have simply if you give them the whole deck the child could simply memorize all of these cards and you can show them a card two plus three equals five and you can say is this correct and they'll be like that's correct and then you don't know are they following the rules of arithmetic or have they simply
So one way you can get around that is you take, let's say, 70% of your cards and you say, learn some rule, child. I'm not even going to tell you what the rule is. That's the point. I want you to learn, figure out how to add by looking at these examples. And I'm saving 30% of these here so I can show you new cards and say, is this one correct or incorrect? And then judging by that, I can say whether or not you're following the rules correctly.
Okay, so that's what cross validation is. Now, how does that undermine science? So that's essentially the scientific enterprises you've outlined it. So how does the no free lunch theorem relate to that? Say it one more time, please. It sounds like what you're saying is, well, I could come up, the child can come up with any rule and it would still be, I could still find some situations in which that rule is valid. Just as many. And in fact, you could have whatever the child
says, I could define a new algorithm which is take whatever the child says and go with the opposite, so to speak, and there are going to be as many priors, as many universes, we're talking about the scientific method, in which that anti-algorithm is actually going to end up doing best on yet unseen situations rather than the original ones.
and it becomes very, very pointed when it's things like... It means that, for example, there are going to be as many situations in which your algorithm does worse than random guessing as in which it does better, no matter what your algorithm is.
Now, this implies to me that scientific methods should not work, yet it seems to. So why? The problem. Yep. The problem. TP. Because your very statement, it seems to, is based upon the data set that has worked so far. But then when we go to off-train, the no-free lunch says doesn't mean the fact that it's worked so far well in the past has anything to say about how well it will work in the future. It's just like evolution has worked so well in the past, has no implications about how well it will work
in the future. So it eats its own tail. The fact that anything that we have used go up to a meta level, a meta algorithm of some sort, where I'm choosing between strategies for choosing between strategies doesn't matter. In fact, there's a philosopher, Carl Schurz, who just came out with a book on meta induction, it's making a big
um splash or waves you know whatever the um aquatic metaphor of choice and um he's got a whole chapter on no free lunch and i actually just wrote a commentary um on that there were some that were solicited and it's going to be appearing in um the journal general philosophy of science or something like that i forget the precise name where he claims that his results um get around no free lunch but no they don't
He's talking about meta induction at this level. So you're doing induction about induction about induction to try to essentially break the Gordian, you know, break free your binders, but they don't break free the binders, you're still stuck. And we have no reason to believe the scientific method. But I sure as hell believe it. Listen to the earlier part of this podcast, I was going a whole hog, I was the most dyed in the wool, conservative dogmatic scientific reasoner you could come up with.
And then the next breath, I'll tell you that it's got no legitimacy. By the way, would you consider math to be scientific or do you consider science to be separate from math? Oh, this is like you asked me before about being an artist versus and so on and so forth. I don't I don't make those kinds of distinctions. I will. Yeah. The reason I was asking is if you did consider it, then it would be as if what you're saying is by using science, I've undermined science.
Girdle, by using math, undermined math. Yep. Well, Girdle undermined a part of math, whereas you're undermining the entire scientific enterprise. Yep, pushing it further, pushing it further. And then things like the inference devices work, go even further. They, in some senses, undermine the notion of epistemology, full stop.
of even knowing something, no matter how that information is arrived at, even by luck. And this gets to another very, very much tongue-in-cheek definition of free will, which was very, very much tongue-in-cheek. And then the stuff on noisy deductive reasoning would say that, oh, but if math itself, including that reasoning that you're using to come to these other theorems, is itself a noisy process inherently,
Well, you got to get going. And so for the people who are watching, if you have any questions for David, he'll be coming back on the theories of everything podcast at some point. So leave your questions below. We'll be talking about the same. Well, forget about that. I can add I can change that later. Professor, thank you so much.
Well, thanks. This is a lot of fun. All right, we'll do. Well, thanks very, very much. Yeah, well, thank you, man. Interesting. And the more actually from my sitting in my chair, the more I actually feel that you, Kurt, are engaging in actually revealing what it is that you're actually thinking. That makes it more interesting, more what? More, more interesting, but also more stimulating.
Again, recall that this was part one of a two-part conversation.
In this second part, we take audience questions, as well as hearing from Kevin Knuth, Annel Seth, Karl Fursten, and Chris Langan, who all have questions posed to the great David Wohlpart.
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": 10568,
"audio_duration_seconds": 7748.27,
"completed_at": "2025-12-01T01:45:35Z",
"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": 78.114,
"index": 3,
"start_time": 66.203,
"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": 96.357,
"index": 4,
"start_time": 78.626,
"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."
},
{
"end_time": 120.043,
"index": 5,
"start_time": 96.357,
"text": " This is part one of a conversation with the inimitable David Walpert. You'll want to watch this one all the way until the end to understand the concepts which will be referenced in part two, where we take questions from you, as well as hearing questions from Kevin Knuth, Anil Seth, Karl Friston, and Chris Langan, all posed to David Walpert. Check the description for the link to part two."
},
{
"end_time": 138.729,
"index": 6,
"start_time": 120.52,
"text": " David Walpart is a mathematician, a physicist, and a computer scientist. He's also the pioneer of three limiting theorems, including the No Free Lunch Theorem, which, as astounding as the No Free Lunch Theorem is, is actually the least monumental of the three. That statement alone is a testament to how seminal David's work is."
},
{
"end_time": 166.254,
"index": 7,
"start_time": 138.729,
"text": " Today we talk about free will and what the limiting theorems have to say about the restrictions on attaining scientific knowledge in general. As usual, click on the timestamps in the description if you'd like to skip this intro. This episode, much like most of the podcasts on the Toe Channel, is extremely technical, but don't be dismayed, stick with it even if you don't understand it, as repeat viewings are what allow you to, in John Wheeler's words, get wet. The whole point is simply to get wet rather than to drink from the fire holes."
},
{
"end_time": 191.834,
"index": 8,
"start_time": 166.254,
"text": " This episode will serve as an introduction to David Walpart's ideas and then later we'll have a deep dive into the intricacies of David Walpart's theorems for part two. This means you should write down your questions. Write down all your questions either privately in some documents or publicly, let's say in the YouTube comment section, especially because I'll be culling from there, since when you write it may spur questions in others or it may even answer a question that someone else had."
},
{
"end_time": 216.049,
"index": 9,
"start_time": 191.834,
"text": " Limiting theorems and no-go theorems, which can be used interchangeably, you'll hear this terminology plenty, are epochal and particularly relevant to the subject of theories of everything, since many people attempt to formulate their own theory of everything and contravene certain bounds without realizing that they're overstepping certain limitations. For example, if one believes that every well-founded mathematical statement can be either proved or disproved, then one is"
},
{
"end_time": 235.282,
"index": 10,
"start_time": 216.408,
"text": " There's nothing wrong with violating per se, it's just that one should know when they're violating and what the implications of that violation are. So for example, perhaps it implies that you can create a perpetual motion machine, or that you have a method of trivially predicting the outcome of some chaotic configuration."
},
{
"end_time": 262.176,
"index": 11,
"start_time": 235.282,
"text": " In short, understanding no-go theorems allow you to understand what loopholes exist and thus gives you an intricate understanding of the inner workings of reality. My name is Kurt Jaimungal. 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 delineating the possible connection consciousness has to the fundamental laws of nature"
},
{
"end_time": 287.824,
"index": 12,
"start_time": 263.012,
"text": " if those laws exist at all and are knowable to us. If you enjoy witnessing and engaging with others on the topic of psychology, consciousness, physics, etc., then check the description for a link to the Theories of Everything Discord as well as subreddit. There's also a link to the Patreon, that is patreon.com slash Kurtjmungle, if you'd like to contribute as the patrons and the sponsors are the only reasons that I can do podcasts with this level of quality and depth"
},
{
"end_time": 318.114,
"index": 13,
"start_time": 288.319,
"text": " As now, this is what I do full-time thanks to their support. With regard to sponsors, there are two. The first sponsor is Brilliant. During the break, I decided to brush up on some fundamentals in physics, so I committed myself to learning one lesson per day on Brilliant. Some point soon, I'd like to speak to Chiara Marletto on constructor theory, which is predicated heavily in information theory, so I took a course on random variable distributions and knowledge and uncertainty. Despite previously knowing the formula for entropy as it's hammered into you as an undergrad,"
},
{
"end_time": 342.329,
"index": 14,
"start_time": 318.114,
"text": " It was extremely edifying to see a different explanation of the formula for entropy that doesn't seem like one that's been handed down arbitrarily from God. Instead, after taking the course, it's easy to see why the formula for entropy is the way that it is and how it's an extremely natural choice. There are plenty of courses, including ones on group theory, which is what's being referenced when you hear that the standard model has in it u1 cross su2 cross su3."
},
{
"end_time": 361.578,
"index": 15,
"start_time": 342.773,
"text": " I recommend that you don't stop before four lessons and I think you'll be greatly surprised at the ease at which you can now comprehend subjects that you previously had a difficult time grokking. At least, I know I did."
},
{
"end_time": 387.961,
"index": 16,
"start_time": 361.578,
"text": " The second sponsor is Algo. Algo is an end to end supply chain optimization software company with software that helps business users optimize sales and operations planning to avoid stock outs, reduce returns and inventory write downs while reducing inventory investment. It's a supply chain AI that drives smart ROI headed by Amjad Hussain, who himself has a podcast on artificial intelligence and consciousness and business growth."
},
{
"end_time": 408.66,
"index": 17,
"start_time": 388.302,
"text": " I've always been interested in theories of everything, as I think any physicist, any theoretical physicist should be."
},
{
"end_time": 436.664,
"index": 18,
"start_time": 409.377,
"text": " Well, that's a controversial statement. But at least if you're naive, you should be interested in that. And then maybe later you realize, okay, there are limitations and so on. Now, those limitations is where our conversation comes in. I get email probably much like yourself, many theories from people. Maybe you don't get email that but either way, I get email plenty of theories. And they violate some limiting theorem or no go theorem. And they don't realize that they're doing this. So I thought, you know what, how about"
},
{
"end_time": 461.903,
"index": 19,
"start_time": 436.954,
"text": " While I'm on this journey to investigate the different theories of everything, how about we educate people on the different limiting theorems that exist? Now you have three, which is extremely rare. Most people would kill for one. So that's why I'm extremely interested in speaking with you. And we're just going to talk about those limiting theorems and what the implications are. Okay, I'll do my best. I thought I'd break this down into three stages. One for each of the limiting theorems. Now you have"
},
{
"end_time": 491.476,
"index": 20,
"start_time": 462.432,
"text": " three as we talked about off air, the no free lunch theorem, which we're going to abbreviate as NFL, and then the unreality of Laplace's demon slash limits on science that's under the umbrella of, I'm unsure what you call the theorem, but as far as I call it the limits on inferences. So limits on inferences is second, and the noisy deductive reasoning, which you've told me that it flourishes the foundations of mathematics, but I'm unfamiliar with that theorem. So I'll be I'm extremely excited to hear about that."
},
{
"end_time": 517.637,
"index": 21,
"start_time": 492.483,
"text": " Okay, so no free lunch, there's actually several different flavors of the no free lunch theorems. There's the ones that have to do with induction, which can be viewed as a formalization of David Hume's informal idea, and which have lots of implications for machine learning and so on and so forth."
},
{
"end_time": 545.094,
"index": 22,
"start_time": 518.114,
"text": " then there's also the versions that have to do with optimization and search. Inference devices, there's actually a whole bunch of impossibility results. The monotheism theorems and the deism theorems and so on, those are two of them and they're up front. The noisy deductive reasoning, that I wouldn't refer to that as"
},
{
"end_time": 574.957,
"index": 23,
"start_time": 546.203,
"text": " impossibility or no-go theorems. That's very much work in progress. That's probably, it's on the front-most burner of all those separate projects. What is, what me and my collaborator David Kinney are doing there is instead trying to understand what the implications would be if we just"
},
{
"end_time": 603.166,
"index": 24,
"start_time": 576.101,
"text": " pursued the time-honored tradition now in mathematics of weakening one of its foundation stones, let the edifice then crumble and relax into a somewhat different shape than we had known before, and try to say some things about that particular shape. So it's not so much a no-go theorem. You will not be able to"
},
{
"end_time": 629.991,
"index": 25,
"start_time": 603.746,
"text": " invoke it when you get one of your email correspondence and say oh this violates the blah blah. It's instead much more of a project that in a certain sense is even undermining the legitimacy of the others and that it's saying it's taking seriously the fact that in mathematical deduction"
},
{
"end_time": 660.486,
"index": 26,
"start_time": 631.305,
"text": " you are never, it does not seem to be possible, and so let's not take it as an axiom rather than try to prove it or not, that you're ever actually 100% sure that every step in your proof is correct. And that's obviously true when people first come up with, say, a suggested solution to the Poincare conjecture or something like that. Or when Andy Wiles came up with his proofs that, among other things, established Fermat's Last Theorem,"
},
{
"end_time": 690.299,
"index": 27,
"start_time": 661.032,
"text": " It was very, very clear at the time that there was a non-zero probability that there was a mistake, so we thought it was wrong. Mathematicians tend to treat these results as though they almost, for lack of a better word, undergo a phase transition, that after people have gone over them enough that, yeah, now we're sure of them and we're going to treat them as, what, 100% ironclad, wholly writ"
},
{
"end_time": 719.462,
"index": 28,
"start_time": 691.084,
"text": " When in point of fact, of course, that's never what happens. It's rather the set of people are always, it's a stochastic process by which you come to greater and greater confidence in the results. May I interject? Please, please, please, do that as much as you possibly can. Otherwise, I have a tendency to open loop. Yeah. I'd like to make this clear for the audience. Generally, when a mathematician proves something, you don't go through"
},
{
"end_time": 747.79,
"index": 29,
"start_time": 719.906,
"text": " what you may learn in a logic course which is you state the axioms then you have some rules of inference like modus ponens and then you go step one is a step two step three step four step five that's actually extremely tedious even to prove the most simple theorems now this can be done on a computer but well this can be done on a computer so I don't think professor you're talking about those actually even then because this is for example the proofs of the four color map theorem which is the first one that was done on a"
},
{
"end_time": 777.312,
"index": 30,
"start_time": 748.302,
"text": " computer. It's actually a somewhat controversial issue in practicing mathematics as well as philosophy of mathematics is, first of all, is it a proof if human beings can't understand it? But even more to the point, you now have to validate the actual code, the program that's being used, and that can never be done to a hundred percent. So you can't ever quite, I don't think there's actually anybody has tried to formally have a no-go theorem"
},
{
"end_time": 805.708,
"index": 31,
"start_time": 777.739,
"text": " that proves this, which is an interesting idea. But as a matter of practice, there's always some place that you need to simply say, well, it's been vetted, it's good enough, let's run with it and not scrutinize this anymore. Your operating system, computer programs, cosmic ray can hit a transistor and it can make a fluctuation. So you can think that's validated by computer, even then you're not completely"
},
{
"end_time": 835.811,
"index": 32,
"start_time": 806.425,
"text": " This is Marshawn 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 touchdowns to threes. And if you're right, you can win big. Mix and match players from"
},
{
"end_time": 845.674,
"index": 33,
"start_time": 835.811,
"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": 867.312,
"index": 34,
"start_time": 845.913,
"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": 883.899,
"index": 35,
"start_time": 867.312,
"text": " What do mathematicians say in response to what you're saying? Because it's as if you're undermining the entire validity of their existence. Well, they say that very, very quickly. You get to the situation where the probability that it's actually mistaken"
},
{
"end_time": 910.623,
"index": 36,
"start_time": 884.309,
"text": " The more it gets scrutinized, the lower that probability becomes. Formally though? Because informally you can say that. Can you say, well, we believe with a 99.58% certainty in Fermat's Last Theorem's proof, for example. They never actually do that. This is all a matter of heuristics, the sociology of the field. So one of the ways of viewing the motivation for the noisy deductive reasoning stuff"
},
{
"end_time": 940.384,
"index": 37,
"start_time": 911.374,
"text": " is that you can imagine trying to actually formalize that. Say there is a stochastic process, you can imagine just where we are, what's called IID, so that's independent identically distributed, doing samples of whether a, so you can imagine a very, very simple model where here is a hypothesis, a proposed theorem, and that you are many times independently testing with a simple bit flip"
},
{
"end_time": 964.974,
"index": 38,
"start_time": 940.811,
"text": " Is this actually true or not? And if you keep getting those flips that keep coming up saying true, true, true, true, and they're all independent, then according to many things like Bayesian reasoning and so on and so forth, it depends upon what are called your prior probabilities. What's called your posterior probability that actually is true given the data"
},
{
"end_time": 991.561,
"index": 39,
"start_time": 965.589,
"text": " that it's been tested now a thousand times, now a hundred thousand times, now a million times. The probability that's wrong as the number of tests, the number of different people who have checked it over keeps growing, that probability of error shrinks as the number of checks grows. It's actually, you can make a very, very simple model where it's formally identical to the proposition that the sun will rise tomorrow."
},
{
"end_time": 1016.834,
"index": 40,
"start_time": 992.022,
"text": " You personally, Kurt, have seen the sun rise, I don't know, a couple of thousand times. If we may be squinting our definition of what it means for you personally to see it rise, we could allow you to pull upon social media or just general media and so on and say it's even tens of thousands. So there's something called Laplace's Law of Succession. This goes back to Simon Laplace."
},
{
"end_time": 1044.241,
"index": 41,
"start_time": 1017.5,
"text": " where he uses very, very simple Bayesian reasoning. He was actually the first person to really use Bayesian reasoning. Bayes didn't really use it, essentially. So Laplace used it to estimate trajectories of planets and whatnot. But in any case, there's something called Laplace's Law of Succession, which says that if you have seen it rise a thousand times and you've never seen it fail to rise,"
},
{
"end_time": 1073.558,
"index": 42,
"start_time": 1045.367,
"text": " then under a very reasonable choice for what's called the prior probability, you would expect, you would assign the expected chance of it rising tomorrow is 1001 divided by 1002. But that's not the same thing as one. So you can build a very, very simple model of what mathematicians are doing, which is that every time a new mathematician takes a new look at the proposed proof,"
},
{
"end_time": 1102.756,
"index": 43,
"start_time": 1074.121,
"text": " that it's just like having the sun do another test of whether the sun's rising. And just like every time you see the sun rise from one day to the next, you get more and more confident that, yep, the sun will also rise tomorrow. Mathematicians, every time they do a new test of having somebody scrutinize it and coming up and saying, yep, looks legitimate to me, you can say there's a greater probability that the next mathematician will come to that same conclusion."
},
{
"end_time": 1130.026,
"index": 44,
"start_time": 1103.08,
"text": " but in both cases you'll never actually achieve a notion, anything like, whether you can even define it, saying it is true, 100 percent. And so you can build simple mathematical models or the mathematical process along those lines, and doing that you can actually justify some common mathematics heuristics. Mathematical reasoners, for example, if you are"
},
{
"end_time": 1159.821,
"index": 45,
"start_time": 1130.623,
"text": " Still in the stage where you're trying to ascertain whether you think a proposition, a proposed theorem is true, if as a mathematician or a theoretician in any field, if there are what you might view as multiple independent lines of reasoning, all of which say that that proposition is true, versus there only being one, when there's multiple independent ones, you are going to assign higher probability to being true than not being true."
},
{
"end_time": 1188.558,
"index": 46,
"start_time": 1161.032,
"text": " That's as a matter of social practice. Humans do that. And by using this model of mathematical reasoning, you can actually say, guess what? That's actually formally justifiable using some of the rules of Bayesian reasoning. So one thing one can do is take this approach to what mathematicians actually do and actually establish that some of the heuristics that mathematicians engage in are actually legitimate."
},
{
"end_time": 1218.933,
"index": 47,
"start_time": 1189.138,
"text": " But where things become much more interesting to me, at least personally, is if you say along the spirit that, well, let's weaken the parallel line postulate. And look, guess what comes out? Non-Euclidean geometry and all of its riches. Let's weaken the postulate, the continuum hypothesis. And guess what? What comes up is all different kinds of mathematical logics and so on and so forth. We can say let's weaken the hypothesis"
},
{
"end_time": 1249.394,
"index": 48,
"start_time": 1219.599,
"text": " that mathematics itself is in some sense external to humans, that humans are discovering it and that it, mathematics, is ironclad with every step being following from a previous step in a deterministic manner. Instead, let's weaken that and assume that it's not just mathematicians who are sarcastic prophecies, but in some sense that which they are engaging with, mathematics itself,"
},
{
"end_time": 1278.012,
"index": 49,
"start_time": 1250.077,
"text": " the underpinnings of our physical reality, if you are going to adopt a Max Tegmark's view of what physics is, that's just mathematics, that it itself is inherently a stochastic process. That the quadratic equation, it's not just human mathematicians whose confidence in the quadratic equation is never 100%, that actually the quadratic equation itself"
},
{
"end_time": 1307.91,
"index": 50,
"start_time": 1278.746,
"text": " If you're to adopt an almost platinous view of mathematics, this external thing that humanity is trying to get its hands around, that in itself also is not 100% true, but is only squishy. What are the consequences? And that's what we're wrestling with in that body of work. So professor, you mentioned that you use the phrase for a reasonable choice of priors. So how does one"
},
{
"end_time": 1338.831,
"index": 51,
"start_time": 1309.241,
"text": " Yep. So here's what I'm thinking. The priors, in some sense, when you're doing Bayesian reasoning, it doesn't matter. What matters is that you update it in accordance to a certain model. So then you want your posteriors to be a certain amount. Let's say you want to be 99% certain of a certain fact. Is it not true that you can then change your priors? So like, let me figure out how to phrase this. You're spot on. The priors? Essentially, you can always change the priors to give you any posterior probability that you want. You can reverse the equation. Yep. You're correct."
},
{
"end_time": 1367.671,
"index": 52,
"start_time": 1339.394,
"text": " You're correct. And that's related to the no free lunch theorems. Another way of phrasing, this is why sociologically, Bayesian reasoning fell into very ill repute around the beginning of the 20th century. And that's why Fisher and other fellow travelers, they came up with things like hypothesis testing, confidence intervals, all that kind of stuff."
},
{
"end_time": 1398.2,
"index": 53,
"start_time": 1368.268,
"text": " what's called sampling theory statistics became ascendant and Bayesian reasoning fell into disrepute because people misused it by choosing priors that were leading to results that kept coming up wrong. And Bayesian reasoning is a matter of garbage in, garbage out. To push it even further, there is no way that anybody's even imagined that you can somehow prove what the priors are."
},
{
"end_time": 1426.732,
"index": 54,
"start_time": 1398.524,
"text": " what they should be, how you should reason. There are heuristics. People like to use what's called uninformative priors, what are called Jeffrey's priors. Ed Jaynes in his book, The Logic of Science, and many of his, I use the phrase, fellow travelers, they will try to say that your priors, that it behooves you to choose priors that accurately reflect the precise forms of ignorance."
},
{
"end_time": 1453.899,
"index": 55,
"start_time": 1427.21,
"text": " that you have about a scenario. That's nice, seems reasonable. They've not proven that injunction. It's just they're taking it almost as a feature of good practice. One might dispute it, and even if one does not dispute it, it's only going to get you so far. There are many scenarios where just what you might mean by"
},
{
"end_time": 1480.794,
"index": 56,
"start_time": 1454.753,
"text": " quantifying my ignorance as a prior is not at all clear. So yeah, so we are, in a certain sense, the no-fee lunch theorems are saying this by one way of to view one of their contents is that if I have any particular reasoning algorithm, be it in say the machine learning algorithms,"
},
{
"end_time": 1507.5,
"index": 57,
"start_time": 1481.152,
"text": " or optimization algorithms, evolution itself, any of which is a search algorithm. And if I find that of my two algorithms, A and B, that for one set of priors, A outperforms B, so it's going to make more accurate predictions about whether the sun will rise tomorrow or what have you, that there's always going to be just as many priors"
},
{
"end_time": 1537.09,
"index": 58,
"start_time": 1508.404,
"text": " for which B beats A. So it's a way of taking your intuition and not only quantifying your intuition, but also elaborating that it's more than just one teeny little squirrely. If you want to be a naysaying annoyance, you can find a counter example. It's saying, no, it's not just that there's one prior in which the conclusion that you want to reach"
},
{
"end_time": 1563.729,
"index": 59,
"start_time": 1537.619,
"text": " When you say just as many, you mean in terms of the size of the set? The no-free lunch theorems are stated in the following way. They say that if I take a"
},
{
"end_time": 1592.039,
"index": 60,
"start_time": 1565.026,
"text": " In the scenarios to which they apply, you can have a prior itself, you can define a probability distribution over priors. And you can say, let's just mathematically, this is a mathematical trick. Say there's a uniform distribution over priors. Then I can say that on average, where I'm averaging over that distribution over priors,"
},
{
"end_time": 1618.899,
"index": 61,
"start_time": 1593.285,
"text": " any two algorithms perform the same. So that means that if I find one particular set of priors whose measure under that uniform distribution is such that that algorithm beats A, A beats B under that set of priors, then there's another set which has equal measure under this uniform distribution."
},
{
"end_time": 1646.8,
"index": 62,
"start_time": 1619.531,
"text": " So yes, it's the same number, but number is being defined very precisely in terms of a probability distribution. We're talking about infinite spaces as a thing. So it's not just counting, oh, there's three priors for which a beats b and three priors for which b beats a. It means the size of the set in a volume sense by which a beats b is the same as which b beats a."
},
{
"end_time": 1675.691,
"index": 63,
"start_time": 1647.278,
"text": " I didn't know that your theorem had something to say about Bayesian reasoning or it used Bayesian reasoning in it. Whenever I hear people speak about Bayesian reasoning, they hold it as this pinnacle of rational thinking. And generally people who are who like you mentioned in an email, which we can remove this part where you said there are certain people who try to sound as if they're extremely intellectual. They say it as if it's ironclad. However, the process may be a rational process, but the outcome"
},
{
"end_time": 1705.009,
"index": 64,
"start_time": 1676.613,
"text": " You can reverse engineer your priors to give you whichever outcome you like. And I've always found that, so why is it that people keep stating, well, look, I'm a skeptic, I update my priors, I do so and so, and so therefore I've arrived at the correct result. I don't think that's the case. And I haven't heard anyone make a claim otherwise until now, right now, just extemporizing with you. Yep. So the way, so this is, I've even got paper saying like, what do the no free lunch theorems really mean? People,"
},
{
"end_time": 1728.439,
"index": 65,
"start_time": 1706.067,
"text": " Even though me and my collaborators, when we produced the theorems, we were very, very explicit about this. People misconstrue their significance. The simplest version of the no-freelance theorem says that if you have a uniform prior, then all algorithms perform the same."
},
{
"end_time": 1754.701,
"index": 66,
"start_time": 1729.275,
"text": " But as we very quickly then follow that in our papers by saying that we can extend this by saying if you have a uniform distribution over priors, then all two algorithms do the same. So we are not, and we do not, we're very careful to say we do not advocate that there is, that the prior is uniform,"
},
{
"end_time": 1777.363,
"index": 67,
"start_time": 1755.435,
"text": " We're not advocating that actually you're allowing priors to vary, but that there is a uniform distribution over priors. That's not our point. Our point is rather that a uniform distribution over priors is a mathematical tool by which one can actually formally prove what Kurt just intuited."
},
{
"end_time": 1805.418,
"index": 68,
"start_time": 1778.592,
"text": " That is the significance. It's not that we are saying there is a uniform prior over priors, or that there is simply a uniform prior. Uniformity is a tool by which we are using to quantify the fact that if A beats B in some situations, then B beats A in just as many, and we're using this uniformity to be able to quantify what that just as many means."
},
{
"end_time": 1827.449,
"index": 69,
"start_time": 1806.51,
"text": " So many people will say that, oh, well, I don't think that the actual prior when I'm doing my machine learning algorithm is uniform, therefore no free lunch is irrelevant. No, because you've got to justify whatever the implicit prior is that you do use and no free lunch is what it's telling you is that they're going to be just as many other ones for which you're screwed as for which you succeed."
},
{
"end_time": 1856.169,
"index": 70,
"start_time": 1828.387,
"text": " That's what it truly means. Its significance extends far beyond the case. It doesn't suffice for you, the naysayer, to simply say, I don't agree with the uniform prior, therefore no free lunches are relevant. You're missing the point of no free lunch when you say that. It's much more it's a way of formalizing and elaborating Kurt's intuition, which should be, I guess, the title of the new paper. Why should anyone care about the no free lunch theorem?"
},
{
"end_time": 1885.776,
"index": 71,
"start_time": 1857.073,
"text": " Because you are correct that people carefully say that according to Savage's axioms, De Finetti's axioms, Cox's axioms, the many axiomatizations of rational reasoning, which will lead you to say you should update your priors according to Bayes' theorem."
},
{
"end_time": 1912.688,
"index": 72,
"start_time": 1886.254,
"text": " which is, it's not even a theorem, it's really just the definition of conditional probability. It shouldn't be glorified to be called a theorem, but you are exactly correct that yes, once you have arrived at your priors, if you then for the rest of your life using those priors want to be rational, you should be Bayesian. If by being rational, you mean you obey these particular axioms of these various people that come up with"
},
{
"end_time": 1940.828,
"index": 73,
"start_time": 1913.439,
"text": " But that does not mean anything about what those priors should be in the first place. It just means that you've got a well designed machine that takes garbage in and produces garbage out. It does that very, very cleanly and efficiently. It doesn't change the fact that garbage in becomes garbage out. You're just not adding any more garbage. What's strange is that if we have no assumptions on the underlying input space or the underlying space, then"
},
{
"end_time": 1969.735,
"index": 74,
"start_time": 1941.749,
"text": " When we're speaking about strategies, if I'm to think of evolution as employing strategies or organisms as employing strategies, it seems like what your theorem would state is that evolution would have equally produced any strategy. However, when I look around, it doesn't seem like any is equally viable. Yep. But here's the thing. It's just like that. I'm warning at the bottom of a prospectus for a mutual fund that past performance is not indicative of future performance."
},
{
"end_time": 1996.237,
"index": 75,
"start_time": 1970.418,
"text": " It's also true in that I can save, I've got a repeated test, it's like the sun rising. My data set is just every time in the past, let's say, I mean, it's very hard to actually quantify that natural selection, mutation with descent and so on, whether it is actually being efficient or not, because natural selection in the real world has its"
},
{
"end_time": 2023.473,
"index": 76,
"start_time": 1996.698,
"text": " Some of its true power is the fact that it is obscenely parallel computation. The number of actual different genomes that are right now out in the environment that are floating around, evolving, even if you don't include things like viral genomes, it's, you know, whatever, it's going to be far, far larger than the number of stars in the universe, you know, that kind of numbers. And when you got that kind of parallelism advantage,"
},
{
"end_time": 2044.957,
"index": 77,
"start_time": 2023.985,
"text": " Almost any dumbass algorithm is going to be able to do well. So it's not even clear that in any sense, evolution has, quote, done well in our past. But even if it has, past performance is not indicative of future performance. That's what no freelance theorem is quantifying. All those, even if it were true, if you wanted to say"
},
{
"end_time": 2073.319,
"index": 78,
"start_time": 2045.418,
"text": " Every time I ran the experiment, which can have a bit up or down, up means evolution did well in this situation, down mean it did poorly. Even if I've run that experiment huge number of times and each time I kept coming the result, evolution does well, evolution does well, evolution does well. No free lunch is saying that when we now extrapolate off training set to situations we have not seen before, anything can happen."
},
{
"end_time": 2093.985,
"index": 79,
"start_time": 2074.292,
"text": " And that's the crucial thing about no free lunch is that it's not saying anything about situations you've already seen. It's saying things about what's called off training set new situations. It's all about generalization. It's not about if I memorize things, and I know that what I've seen is not is never going to change."
},
{
"end_time": 2117.858,
"index": 80,
"start_time": 2094.497,
"text": " And you are then asked, well, what is it that you saw in that situation? Well, you answer and you're correct. Memorization is fine. The question is always generalization. To ask, what is it that you will see? That's where probabilities come into play. That's where Bayesian reasoning comes into play. And that's where no free lunch comes into play."
},
{
"end_time": 2145.845,
"index": 81,
"start_time": 2118.677,
"text": " We're going to get to cross validation in about maybe five minutes or so. And as I mentioned to you over email, I'm a philosophical person. I'm an introspective person. And I was wondering with regard to the no free lunch theorem, does that imply that there's no optimal strategy for one's life? What I mean is, people, including myself, perhaps, I'm mainly speaking about myself, have regret over the past and squandered opportunities, feeling like"
},
{
"end_time": 2171.084,
"index": 82,
"start_time": 2146.51,
"text": " feeling like I didn't implement the optimal strategy, I haven't lived up to my potential. Now, is that a valid concern? Or is your theorem nihilistic, where it doesn't matter what you do? It doesn't seem, it seems like it matters what people do. So it seems like there is something to be said about maximizing on one's potential as ill-defined as that sounds. Hear that sound."
},
{
"end_time": 2198.114,
"index": 83,
"start_time": 2172.005,
"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": 2224.224,
"index": 84,
"start_time": 2198.114,
"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."
},
{
"end_time": 2247.585,
"index": 85,
"start_time": 2224.224,
"text": " of their commerce. Shopify, by the way, powers 10% of all e-commerce in the United States, including huge names like Allbirds, Rothies, 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": 2257.858,
"index": 86,
"start_time": 2247.585,
"text": " All lowercase. Go to shopify.com."
},
{
"end_time": 2278.865,
"index": 87,
"start_time": 2261.084,
"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": 2307.346,
"index": 88,
"start_time": 2278.865,
"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, so that means no plastics, no subscriptions, no proprietary blades, and no planned obsolescence."
},
{
"end_time": 2323.712,
"index": 89,
"start_time": 2307.346,
"text": " 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. Visit hensonshaving.com slash everything."
},
{
"end_time": 2353.2,
"index": 90,
"start_time": 2323.712,
"text": " If you use that code, you'll get two years worth of blades for free. Just make sure to add them to the cart. Plus 100 free blades when you head to H E N S O N S H A V I N G dot com slash everything and use the code everything. When I was younger, when I had just come up with the no free lunch theorems in my scribblings to myself that concern philosophical matters,"
},
{
"end_time": 2383.012,
"index": 91,
"start_time": 2354.582,
"text": " I would have something that I eventually abbreviated as TP. And that came, that was an abbreviation of the problem. And that referred to given no free lunch, what in the world do I do? Um, how old were you? Oh, um, actually my first intimations of the problem."
},
{
"end_time": 2413.609,
"index": 92,
"start_time": 2383.729,
"text": " were as an undergrad, before I formalized the no free lunch, when I was worrying over these things. I actually did not fall into despair. Perhaps I should have. Every time I walk across the street, I am behaving with my life on the line as though no free lunch were wrong. Every time I do anything. So I would say that in a certain sense,"
},
{
"end_time": 2441.152,
"index": 93,
"start_time": 2415.247,
"text": " far more deeply than feelings of regret. No free lunch theorem has to do with how I tie my shoes. It's how do I even exist? And how do you act? How do you act? And so I look at this and I say, okay, now what? Um, and here's how to make it even worse. And another thing, I guess,"
},
{
"end_time": 2470.811,
"index": 94,
"start_time": 2441.886,
"text": " Through my career, I've not thought of myself this way, but I tend to have these no-go-ish kinds of bent of finding out. To me, it's one of the most glorious... I don't know, this almost sounds Buddhist. To me, it's one of the most glorious aspects of... I wouldn't even use the word existence. The more that I am squashed flat, the more that I am seen to be a teeny little restricted nothing,"
},
{
"end_time": 2502.005,
"index": 95,
"start_time": 2472.005,
"text": " That to me is the height of beauty. That to me is the most sublime and rapturing thing to sit on your back in a meadow at night and to just look up and to try to project and realize all those stars are three-dimensionally configured and just go through like a powers of 10 kind of a thing to fully to try to fully appreciate my own insignificance, my own limitations. That to me is the most beautiful liberating thing I can imagine."
},
{
"end_time": 2528.251,
"index": 96,
"start_time": 2502.619,
"text": " and these results all have that character to them that they are in some sense precisely because I am not even a gnat on the back of a fly that produces a sense of joyful glee almost like a boy or a child does oh I am liberated I don't have to go to whatever but there anyway this is a preamble to another"
},
{
"end_time": 2557.193,
"index": 97,
"start_time": 2529.036,
"text": " set of results that are along the same lines, really, which are what are sometimes referred to nowadays as the block model of cosmology, which is the following. So I had a paper that was published, I guess I was a grad student at the time, at what was called the International Journal of Theoretical Physics, called"
},
{
"end_time": 2585.23,
"index": 98,
"start_time": 2557.79,
"text": " So okay, there's various notions of what's called the arrow of time. There's one of them is the second law of thermodynamics, and another one is what is sometimes called psychological arrow of time, the notion that it feels to us as though time is going forward. Paul Davies had a very nice book on this called the arrow of time in like 73, which is, it's not for popular audiences, it's very much a"
},
{
"end_time": 2609.07,
"index": 99,
"start_time": 2586.135,
"text": " for students of physics. But there is no such... When we talk about movement in space, we talk about position against time. To talk about movement in time, you would need to have two dimensions of time. We don't."
},
{
"end_time": 2639.548,
"index": 100,
"start_time": 2609.957,
"text": " The future and the past in, this is the lesson of quantum mechanics really and of general relativity, they are just as real as the present. There is no physical variable that can actually distinguish these things from one another. Okay, but so why do we have the illusion that time seems to be moving forward? And it's actually can be drilled down to being a two-step process. First,"
},
{
"end_time": 2665.691,
"index": 101,
"start_time": 2640.026,
"text": " Notice that you as a human being, you have memory of your past, but you have none of your future. A more precise way of phrasing that is that you can retrodict what happened in the past more accurately than you can predict what will happen in the future. You then drill down and try to say, well, why might that be the case?"
},
{
"end_time": 2694.889,
"index": 102,
"start_time": 2666.34,
"text": " And eventually you can actually make arguments which are only informal to this day that it's got to do the second law of thermodynamics. But the reason I bring all this up is that in a certain sense, it's only worse because whenever you, Kurt, are making these arguments about what you've observed in your past and your memories and your regret, that's actually only retrodiction. You don't even know for sure that those things even happened."
},
{
"end_time": 2725.316,
"index": 103,
"start_time": 2695.862,
"text": " You only exist in the present or simultaneously the presence in all the futures with no, certainly no kind of a free will that you can somehow supervene into all this, supersede all this and change the way it is. The book has been written and you are simply a single page in the book remembering the pages in the past as best you can and wondering about the pages in the future."
},
{
"end_time": 2752.005,
"index": 104,
"start_time": 2726.561,
"text": " That is all on top of no free lunch, the restrictions of inference devices, and so on and so forth. Even before you get to the sort of traditional concerns of metaphysics about what is reality, there are all these nuances that prevent you from actually in a certain sense even getting to be able to sit up and ask that question in the first place."
},
{
"end_time": 2776.852,
"index": 105,
"start_time": 2753.78,
"text": " It's a very, very strange thing and I'm not sure how to actually deal with it. Okay, so we're going to talk about free will, which it sounds like what you just mentioned is against free will, yet you're one of the largest proponents of the existence of free will that I know of. So we'll talk about that soon because I'm sure you're defining free will differently than the way most people think of it."
},
{
"end_time": 2794.326,
"index": 106,
"start_time": 2777.381,
"text": " Now, you also mentioned that this merriment that you have as an adult and perhaps even as a child is when you think about your insignificance. And just like you see people like Sabine Hassenfelder as intellectually posturing when they dismiss free will,"
},
{
"end_time": 2818.37,
"index": 107,
"start_time": 2795.247,
"text": " When I hear people say what you just said, also Lawrence Krauss expressed a similar sentiment that, well, the more I think about my ephemeral nature, the more that I feel like there's meaning because there's a temporary nature to this. I see this as a form of rational posturing to show like, look how much I don't need my importance, but I'll explain why and then please push back. I just want to hear what your thoughts are on this. Because"
},
{
"end_time": 2837.688,
"index": 108,
"start_time": 2819.445,
"text": " If one takes that argument seriously, well, look how limited our life is and then that makes it more meaningful. Well, why don't you just kill your child, make your life, make your child's life smaller? Is that not more meaningful? Or maybe it's that we have to say that because the alternative that you're extremely, extremely important and every single thing you do matters beyond"
},
{
"end_time": 2864.053,
"index": 109,
"start_time": 2838.131,
"text": " what you can imagine. Maybe that's way too crushing. And so the alternative that, okay, great. I don't have this burden or onus that's freeing, but it's not actually a rational assessment that led you to that freeing feeling. It's more, well, the alternative, I don't, it's a horror show to think about how meaningful and important my life is to other people and to the world. And perhaps, well, let's hear what you have to say to that. I never use the word meaningful."
},
{
"end_time": 2893.882,
"index": 110,
"start_time": 2864.531,
"text": " That those other people, they do then make the argument that though, okay, therefore I must make meaning out of what I have, and that then imbues it with more significance. Uh-uh. No. You are not. This is the rhetorical you. You have no meaning. You have no significance. And when I say that this is liberating and induces a feeling of glee and merriment, that's an aesthetic."
},
{
"end_time": 2924.053,
"index": 111,
"start_time": 2894.753,
"text": " reaction so um it was when i first came to feel that it was shortly after i first learned quantum physics this was when i was a teenager and i think people misconstrue the true significance of quantum physics they talk about the uncertainty principle wave particle duality yes that's all really cool but the"
},
{
"end_time": 2950.401,
"index": 112,
"start_time": 2924.411,
"text": " more important aspect of it, I think, lesson of it, is precisely the establishment of the Tegmarkian view that all that a reality is, is a mathematical system, that there is nothing outside of that, that in quantum mechanics what you find when you do it"
},
{
"end_time": 2975.503,
"index": 113,
"start_time": 2950.862,
"text": " that you cannot do any of your predictions properly unless both the observer and the system being observed are both in an inclusive way treated quantum mechanically. Wigner and all these other people, this is almost like this is an Everett view which is the one that's really adopted whether they want to admit it or not by everybody working in like quantum computation and so on. There is no external observer. Everything"
},
{
"end_time": 3002.551,
"index": 114,
"start_time": 2976.015,
"text": " Sean Carroll also is an effective proponent of this. There is no external observer. You don't need to invoke something beyond just the mathematics itself to be able to give all phenomena that we can experimentally assess. Once you do that and you start to realize that, oh, well, ultimately one mathematics is just as legitimate as another one. So if a reality is just a body of mathematics,"
},
{
"end_time": 3031.135,
"index": 115,
"start_time": 3003.029,
"text": " That means other bodies of mathematics are just as real as this one, as physical universes. In any case, it was contemplating that and the associated feelings of insignificance that this was an aesthetic response of mine. And it was only then, up until that point, whenever I thought about Zen Buddhism, haiku, koans, anything like that, spiritualism,"
},
{
"end_time": 3060.333,
"index": 116,
"start_time": 3031.971,
"text": " I'd always been very dismissive of it, very much of a standard. Westerners, maybe you're going to be respectful and public, but it's got nothing to say for you personally. I am still very dismissive of some aspects of Buddhism. For example, Buddhism is supposedly all about the irrelevance of the self. But in fact, in practice at least, Buddhist practice is all working on the self."
},
{
"end_time": 3085.623,
"index": 117,
"start_time": 3060.981,
"text": " So that I get to understand how irrelevant and that's a massive contradiction in the foundations. But here's the thing that happened. Once I started to try to internalize this lesson of quantum mechanics and notions along the lines of no free lunch, now for the first time I could actually appreciate aesthetically some of the products of Zen Buddhism."
},
{
"end_time": 3108.234,
"index": 118,
"start_time": 3086.63,
"text": " Zen Buddhism watercolors, Zen Buddhist gardens. I started to be able to appreciate koans not as deep spiritual work, but as aesthetic works, as works of art, as closer to poems, somewhat like haiku are normally perceived. Yeah, I saw you wrote some haikus on your website."
},
{
"end_time": 3131.323,
"index": 119,
"start_time": 3108.729,
"text": " I don't actually characterize myself that way. When I was"
},
{
"end_time": 3159.172,
"index": 120,
"start_time": 3132.261,
"text": " younger, trying to figure out what I want to be when I grow up in a very practical, concrete sense, what I want to have as my major and so on. I would say things to myself, like I approach science as a philosopher. I approach philosophy as a mathematician. I approach math as an artist, things like that. That's interesting. Like an orboros kind of a thing."
},
{
"end_time": 3187.841,
"index": 121,
"start_time": 3159.753,
"text": " I have no delusions that I am... Well, okay, another kind of... This is interesting. This conversation is going much to sort of lessons of life kind of thing. One of the tragedies I find of people's lives around me all the time and what I very intensely, viciously am on the guard with that I do not indulge in myself is inflated views"
},
{
"end_time": 3211.869,
"index": 122,
"start_time": 3188.08,
"text": " of your own abilities, your own worth, your own capabilities, what you are providing. This has got nothing to do with, oh, my beautiful insignificance. This is just I've seen tragedies where people think that they have the capabilities to engage in something and they change how they're living their lives along those lines and really know that was a misassessment of what they can do. And they began"
},
{
"end_time": 3240.708,
"index": 123,
"start_time": 3212.688,
"text": " and therefore there's frustrated expectations and so on and so forth. Yes, your reach must always exceed your grasp, et cetera, but if you're not really aware of what that grasp is in the first place, you're running a real large risk that you're going to be unsuccessful in that reach of yours. The reason I mention all that is I'm very aware of what my own limitations are. I personally do not view the"
},
{
"end_time": 3270.179,
"index": 124,
"start_time": 3241.869,
"text": " material on my website as necessarily worthy of comparison to those of a real artist or poet. I don't even necessarily, I don't view myself as actually all that smart as a scientist. There are many, I am aware of what my relative strengths are, but in terms of just raw IQ,"
},
{
"end_time": 3298.029,
"index": 125,
"start_time": 3270.674,
"text": " That ain't one of them. There's people out there who are far smarter and who are far better as scientists in that sense. Well, you have a clever saying on your website that you're not intelligent enough to be a mathematician nor careful enough to be an experimentalist. Is that correct? Yeah. Yeah, yeah, yeah. That's from my website. Yep. Okay. Now you were talking about how quantum physics allowed you to appreciate the Zen Collins. And then you're saying aesthetically,"
},
{
"end_time": 3323.626,
"index": 126,
"start_time": 3298.626,
"text": " and not spiritually necessarily, and I interrupted you, so do you mind finishing that thought? Yeah, so as I say, I don't view... I don't think... It's sometimes construed as a philosophy, and I think as a philosophy, Buddhism has nothing to say. I think it, like all religions, are just vacuous fables that humans come up with to try to comfort themselves against"
},
{
"end_time": 3350.555,
"index": 127,
"start_time": 3324.138,
"text": " The big vast expanse, which is what I actually for whatever weird reasons in my psyche, I actually find so beautiful. It's most people, most reasonable people react with some screaming fear. And I think that's the basis of a lot of religion. You can also actually, there's a lot of very interesting literature in the economics field, economics of religion. It's actually very, very interesting as a topic of study, what makes some religions succeed and other ones not."
},
{
"end_time": 3381.135,
"index": 128,
"start_time": 3351.152,
"text": " And of course, it goes back, many people have worried about, have wondered about, as evolutionary theories of why humans have religion. So construed as a religion, I view Buddhism like the other ones, as at best epiphenomena, side effects of other processes that go on and certainly do not have any legitimacy as philosophies concerning reality, whatever you might want reality to be, as bodies of aesthetics, as art,"
},
{
"end_time": 3410.009,
"index": 129,
"start_time": 3382.022,
"text": " That is what I could suddenly appreciate when I started to understand quantum physics and no free lunch and so on. I don't know why I've not really drilled down on this much, but somehow the, this is probably almost like a convergent evolution, that the evisceration of the notion of the worth of the self that I arrived at it, my"
},
{
"end_time": 3436.8,
"index": 130,
"start_time": 3411.476,
"text": " processing based upon science, induce the same kind of aesthetic reaction in me as the evisceration of self that Buddhists undergo for their reasons, inducing them. So it's probably got more to do to say about how human psyches react to that cognitive assessment"
},
{
"end_time": 3453.831,
"index": 131,
"start_time": 3437.312,
"text": " Even though the ways that those cognitive assessments are arrived at is very, very different, and I would frankly argue long and hard that one is legitimate being grounded in science and the other one is not being grounded in obsession with self, I would say."
},
{
"end_time": 3472.773,
"index": 132,
"start_time": 3454.718,
"text": " anyway at this point there were some technical difficulties with zoom and the audio takes a decrement and improves in a few minutes the view that you outlined about religion reminds me of marks's opiate for the masses where it's just there simply to placate the society and i'm curious i don't want to get on a tangent about religion nor"
},
{
"end_time": 3502.449,
"index": 133,
"start_time": 3473.473,
"text": " Well, I don't want to get on a tangent about religion, so if you can briefly comment on this. Why do you think it is that many of the religions have a huge element of hell in them, which is actually terrifying? So if it was just about making people feel better, for me, when I made a conversion to being an atheist, and I know this from virtually every atheist I've spoken to, there's a huge moral weight that's lifted from you when you convert, quote-unquote, to atheism. So it's not as if one is happier necessarily"
},
{
"end_time": 3530.128,
"index": 134,
"start_time": 3502.449,
"text": " simply because of the teachings of the religion. I don't think it's as simple as that, and I'd like to hear what you have to say about that. Yes, no, I fully agree with you. For me, there was never a moral weight, personally, for what it's worth. There's never a moral weight being concerned about after a lot during that. There's actually some very interesting literature that those religions which are most successful in the history. Hear that sound?"
},
{
"end_time": 3557.21,
"index": 135,
"start_time": 3531.152,
"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": 3583.387,
"index": 136,
"start_time": 3557.21,
"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."
},
{
"end_time": 3609.121,
"index": 137,
"start_time": 3583.387,
"text": " 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 Brooklyn. 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."
},
{
"end_time": 3634.633,
"index": 138,
"start_time": 3609.121,
"text": " Go to shopify.com slash theories now to grow your business no matter what stage you're in shopify.com slash theories. A religion that this has to do with the economic religion literature are those that make the most demands of their practitioners. Religions that don't make much in the way of demands as opposed to those that do demand things like you go on a missionary"
},
{
"end_time": 3665.469,
"index": 139,
"start_time": 3635.52,
"text": " or that you give a large tithe, or that you very, very much change what your daily routine is. Those ones tend to be much more successful. So you've got to, the religions that are successful are those that are particularly demanding. So there are these kinds of counterintuitive aspects to religion. As far as the one about whether, so Marx, of course, with the opiate for the masses, he was instead talking about its large"
},
{
"end_time": 3694.292,
"index": 140,
"start_time": 3665.981,
"text": " sociological implications and how he thought it was being exploited as opposed to personally why someone might be taken to be um vested in the notion of an anthropomorphic religion as opposed to spiritualism if you mean by spiritualism something closer to aesthetics there one could say that i personally am very spiritualist it's much more a matter of the supernatural"
},
{
"end_time": 3722.483,
"index": 141,
"start_time": 3695.128,
"text": " why and the aspect of things. Now it's not, it is not certainly, the reason that people are inclined to supernaturalism in general is not simply as a comfort. There is that element to it, but certainly many religions do not provide any comfort"
},
{
"end_time": 3752.483,
"index": 142,
"start_time": 3723.336,
"text": " Traditional Catholicism, for example, was you're born into sin and you might be screwed up earth and that's the way it is. So it is far more complicated than the kind of snide comments I was making. I think that for many people in modern society, the role it plays is a comforter. Historically and cross-culturally,"
},
{
"end_time": 3779.565,
"index": 143,
"start_time": 3753.012,
"text": " Religion has played many, many other roles, wound up with supernaturalism, which is itself wound up with our tendency to overgeneralize the environment. Things like if you are evolving, here's one simple argument in favor of why we are prone to supernaturalism. If I'm evolving human beings in the veldt a million years ago, hominins,"
},
{
"end_time": 3806.101,
"index": 144,
"start_time": 3779.872,
"text": " I can either have them be overgeneralizing, finding patterns where there are no patterns or undergeneralizing. If you're overgeneralizing, you might think that a particular rustle in the grass always means there's a tiger there when that's not true, but you're going to be a lot better off than someone who's undergeneralizing so that when there is a pattern that's related to whether there's a tiger in the grass,"
},
{
"end_time": 3833.183,
"index": 145,
"start_time": 3806.732,
"text": " They don't draw the proper conclusion and therefore they get eaten. So you can make those kinds of arguments as to why we are prone to supernaturalism in the first place, why we do tend to find patterns where there aren't any there. And the deep psychological compulsions that have resulted in religions, many of which have no"
},
{
"end_time": 3865.026,
"index": 146,
"start_time": 3836.988,
"text": " emollient characteristic to them at all, exactly as you're saying. Those ones are not the ones that I see so prevalent in modern American society. So when I made my statements, which were snide and gloob and off the cuff, I was referring much more to modern right-wing religious sentiments."
},
{
"end_time": 3893.217,
"index": 147,
"start_time": 3866.084,
"text": " where I do think it's, you know, there are all these studies that say that those who are very religious in the modern sense of the word, going to, engaging in, in terms of the modern church, that they live eight years longer and things like this, because the stress levels of them are so much reduced. So I'm much more referring to that cohort and their approach to life."
},
{
"end_time": 3919.155,
"index": 148,
"start_time": 3893.746,
"text": " At this point, there was a minor break, and we started to talk about Sabine Hassenfelder. There was no exact question, otherwise I'd leave it here. It's simply the topic of his thoughts on Sabina's video about super-determinism and free will. I guess, so she's got an axe to grind. She wants there to be hidden variable models."
},
{
"end_time": 3945.503,
"index": 149,
"start_time": 3920.384,
"text": " They want reality to be that way, she and some other people. And so they are searching around for any way that they can to try to find squirrel room where Bell's, in her case, Bell's theorem notwithstanding, you can still have a hidden variable model. And I just don't see the point. It's motivated reasoning."
},
{
"end_time": 3975.947,
"index": 150,
"start_time": 3946.63,
"text": " um to and she says it's super free will as that i view as vacuous because that's just what she is arguing for is just what i would always have considered to be sorry is super i forget what her precise super determinism super determinism that's right super determinism it's kind of like a super version of no free will super determinism is just what i had always understood to be standard physics straight out of the box um"
},
{
"end_time": 4006.391,
"index": 151,
"start_time": 3976.817,
"text": " I'm not quite sure why she is trying to make a case, she and others, make a case that's somehow something new. It's much more what they are trying to... It's a strange thing what physicists mean by free will, like in Conway's free will theorem and so on. It's kind of distorted from what the term is normally meant in the larger, in the parlance of philosophers more generally."
},
{
"end_time": 4033.729,
"index": 152,
"start_time": 4008.148,
"text": " And anyway, as I said, I think it's motivated reasoning. How about we define free will or you define free will and then tell the audience what your inference theorems have to say about free will, you'll have to define free will and also define what your inference theorem is and then what the inference theorem has to say about free will. Okay, so um, first a warning, there is a"
},
{
"end_time": 4064.957,
"index": 153,
"start_time": 4035.469,
"text": " Unfortunately, it shouldn't have a particular name as one of the rhetorical fallacies. It's, you know, the ad hominem fallacy, the this fallacy, the that fallacy. For people to take a term that's in very common use and is very controversial because of all these reasons, they want to establish that it's true. So they redefine it in a certain way that they can establish that it's true."
},
{
"end_time": 4092.739,
"index": 154,
"start_time": 4065.742,
"text": " that redefinition, all of the original reasons that it was so controversial, they are now defined out. And what you're left with is actually a very different kind of a concept. And they now then herald it and say, oh, look, I've just proved such and such where what they've really done is redefined it in such a way that they could prove it. But all the reason people were interested in the first place is now out of their new definition. So it's a bait and switch."
},
{
"end_time": 4122.176,
"index": 155,
"start_time": 4093.712,
"text": " The fallacy is called Mott and Bailey, I believe. Mott and Bailey, okay. How's that spelled? M-O-T-T-E and then Bailey's as the drink, I believe. Essentially, you take some fact and then you're straw manning it and you use the straw man to disprove the larger one. Yeah, or to prove it as the case might be, Mott and Bailey. That's nice. I've got to look that up later. So, continue. So, yes. So, for example, okay, so certainly if you're familiar with the concept,"
},
{
"end_time": 4149.462,
"index": 156,
"start_time": 4122.637,
"text": " Um, the notion of God has been treated that way. Um, you know, many people would say, Oh, Einstein really believe in God because, you know, he said the laws of nature and laws of nature is God. Um, but of course, the reason that God is considered to be a controversial topic is because it's a banal promote, promote. It's this dude in the sky with a big beard who's deciding who's going to win the football game Sunday night."
},
{
"end_time": 4177.108,
"index": 157,
"start_time": 4150.265,
"text": " And that's the one that's controversial and so on and so forth. And Einstein wasn't saying anything about in favor of that particular one. So it's a little bit of a Mott and Bailey's. And frankly, I think that he has done some great work and he's a good friend and a great guy. But I think that Dan Dennett was recently involved in this with free will."
},
{
"end_time": 4204.48,
"index": 158,
"start_time": 4178.063,
"text": " He wrote a book in which he tries to now claim I've proven free will. And in essence, he's defined away its aspects that made it controversial in the first place. So what do I mean by free will? There's one aspect to it, which would be that of the people who write dictionaries. What does it mean in common discourse?"
},
{
"end_time": 4235.401,
"index": 159,
"start_time": 4205.657,
"text": " what I think free will means and common discourse has something to do with a means by which it is possible to abrogate the laws of deterministic science in such a way that a human's cognitions are not subject to those laws. So that what is going on in your head right now is not"
},
{
"end_time": 4254.087,
"index": 160,
"start_time": 4235.947,
"text": " some very complicated, too complicated to calculate, but nonetheless unviable function of what it was you were thinking 10 seconds ago together with the stimuli, sensory stimuli from the environment, that somehow there's something else"
},
{
"end_time": 4284.565,
"index": 161,
"start_time": 4254.889,
"text": " that is not subject to any laws of science in some sense. It's not even subject to sarcastic laws. It's just amorphous and ill-defined. And so certainly all of us have a subjective experience of such a thing. But the proponents of free will in this sense, which is controversial, would say that no, that's actually an objective truth. I don't see that there's any room for that, to be quite honest."
},
{
"end_time": 4308.183,
"index": 162,
"start_time": 4284.753,
"text": " So in that sense, no, I don't, if that's what, if we do adopt that kind of a definition, which I think is in accord with the way it's used in common parlance, no, I don't see it, there's any room for such a thing. You can, you can maybe kind of get close and Scott Aronson, again, he's involved engaged in a Martin Bailey's in this particular paper of his,"
},
{
"end_time": 4337.398,
"index": 163,
"start_time": 4308.763,
"text": " But he and even worse Martin Bailey's as Seth Lloyd, they try to make the case that Turing machine notions of computability provide a way out in that they say that, yes, it might be the various flavors of the argument, but one version of it is, yes, the present state of your mind is a deterministic function of what it was and so on, but it's an uncomputable function"
},
{
"end_time": 4367.415,
"index": 164,
"start_time": 4338.012,
"text": " in some technical senses. Yeah, Stephen Wolfram says something similar where he uses his term computational irreducibility and says, well, yes, there are outcomes that are computationally irreducible, and that's where free will may live. But ultimately, you don't have control over the laws. It's still a deterministic step by step algorithmic process. Yeah. And in a certain sense, it's kind of interesting because one might ask in terms of the laws of science, of physics, and how the universe actually evolves"
},
{
"end_time": 4393.217,
"index": 165,
"start_time": 4368.114,
"text": " Is there any room for anything besides deterministic or stochastic evolution? I can write down the definition, you can go look it up in math textbooks what a stochastic process is, and you can look up in math textbooks what a deterministic invertible process is, and you can actually formulate the one is a special case of the other. Is there any way you can even in theory imagine"
},
{
"end_time": 4422.875,
"index": 166,
"start_time": 4394.07,
"text": " science that would not be either random or deterministic, as with anything else it could possibly be. And there are notions in the original motivation for Komogorov, in terms of his work on Komogorov complexity, and many other people who are wondering about intrinsic randomness, what does complexity mean, what does random really mean. They came up with definitions which, something that is random,"
},
{
"end_time": 4447.244,
"index": 167,
"start_time": 4423.677,
"text": " but is not random, it has no probability distributions. So it's interesting to think about whether that might provide any kind of a, whether that's a difference without significance or whether there is a significance to that possibility that physical systems might in essence violate the physical church Turing thesis and have"
},
{
"end_time": 4477.398,
"index": 168,
"start_time": 4447.944,
"text": " this kind of an uncomputable character to them, which is not random. And it's not deterministic, but it's something different. And whether that has any kinds of consequences, even if it's true. So that's interesting in almost like a philosophy of science kind of a way. But it really is, I would say, irredeemably far from what people mean by free will to really say that it's relevant to the discussions that people have about what free will is. I think for them,"
},
{
"end_time": 4503.2,
"index": 169,
"start_time": 4477.978,
"text": " the vast majority for what philosophers have meant by it for millennia, it's no, it's certainly nothing to do with comogoro complexity and turning machines and so on and so forth. It's simply the notion of their this supernatural in the literal sense of above nature, above the laws of nature, which is inherent in me and my soul, and I cannot even define it because it is something that's non-definable."
},
{
"end_time": 4533.746,
"index": 170,
"start_time": 4504.036,
"text": " and that allows for superseding of the laws of physics when it comes to my own brain. That's what is sometimes meant. I think that's most fully what is meant by people who are talking about free will, have been talking about historically, and I almost view it as a non-starter. Okay, so to make a summary for the audience who just wants a clear answer,"
},
{
"end_time": 4562.534,
"index": 171,
"start_time": 4534.275,
"text": " do we have free will or not? In that sense, I don't think it's even meaningful to say that we do. I don't see that you can, if you you're almost by that definition, one is almost defining free will to be something that's undefinable. And it are and is supernatural. And I can't even almost react to that. If it's not definable, I can't say anything. And if it's not"
},
{
"end_time": 4592.398,
"index": 172,
"start_time": 4563.114,
"text": " And I don't see there as a room for any supernatural things in all notions of reality. Circling back to a previous theme, one of the things that I find most beautiful about our inconsequentiality is that the whole notion of consciousness, it's a post-hoc fable that we cook up to try to give us the illusion that we're in control of our own brains. There's all these fascinating phenomena going back like Ben Libet and so on."
},
{
"end_time": 4618.643,
"index": 173,
"start_time": 4592.892,
"text": " demonstrating that it's at least logically consistent, in fact seems to be the case, that all the things that we think we're in control of, we're not in control in any meaningful sense of any of them. We're making up stories after the fact. Corpus callosum severed people going, you know, that whole literature. It's fascinating that when you ask one half of a brain why the other half is doing what it is,"
},
{
"end_time": 4644.684,
"index": 174,
"start_time": 4619.172,
"text": " Let me steelman the opponent. So with regard to the supernatural"
},
{
"end_time": 4654.428,
"index": 175,
"start_time": 4645.179,
"text": " Most religious people would say that there is no supernatural... Well, let's forget about most religious people. There is the view that there is no supernatural even in the religious domain."
},
{
"end_time": 4675.196,
"index": 176,
"start_time": 4654.991,
"text": " that god operates by natural law in fact god is natural law and then we may say well god is supposed to be a man in the sky etc however that wasn't the case up until about 200 or 300 years ago at least it wasn't the majority case and mysticism christian mysticism in particular was extremely rampant where you can't define god attempts to do so are actually"
},
{
"end_time": 4704.701,
"index": 177,
"start_time": 4675.538,
"text": " And I agree that this is what free will should be defined as. There's somehow"
},
{
"end_time": 4724.377,
"index": 178,
"start_time": 4705.282,
"text": " us interacting with nature that is different than just the physical laws we know it. It's a metaphysical statement when we think that we have free will. And then you were saying, yeah, it would have to be supernatural because it's above the laws of physics. Well, Raymond Smullian, I'm sure you've heard of Raymond Smullian, the mathematician, he had many writings on free will, and he would say,"
},
{
"end_time": 4745.947,
"index": 179,
"start_time": 4725.384,
"text": " What defines you? What's the difference between you and natural law? You're a product of natural law and we tend to think of free will as not existing because the outside world is more powerful than us and somehow despite our will, it just overcomes us. However, that would mean that there's a distinction between you and natural law. Well, that's illusory. So in some sense, you are natural law. So in some sense, well,"
},
{
"end_time": 4776.732,
"index": 180,
"start_time": 4746.988,
"text": " But so what is the meaning of free will? Is it just a psychological phenomenon that people have this sense of agency? Yeah, sure, that's definitely true. That half of a hemisphere thinks it's in control of the other one. And yeah, there's all kinds of... Is there a psychological phenomenon just like people can get depressed, just like people can become obsessive compulsive, just like much more commonly everybody, yours truly as well, feels, has this feeling"
},
{
"end_time": 4806.596,
"index": 181,
"start_time": 4777.227,
"text": " Yes, it's interesting maybe to look at some of the neurobiology behind some of that. But if that's what the statement is, then yeah, it's definitely true. Let's go do a survey. Do people feel this way? Yeah, they would do. If instead it's a statement about metaphysics, to say it's natural law, I would say as a Martin Bailey, you're ending up with nothing. So then what do you mean by free will if you don't mean this part of neuropsychology?"
},
{
"end_time": 4832.688,
"index": 182,
"start_time": 4807.278,
"text": " If you're going to say that is just part of determinism, what is it? There's nothing left to define it as, I would say. Even the Libet experiments, you mentioned the Libet experiments. Libet says that his, and I would like to interview him. I don't know if he's still alive. I think he's not. And by the way, those are now controversial experiments. Yeah. And Libet himself was a proponent of free will and says that his experiments have been"
},
{
"end_time": 4859.326,
"index": 183,
"start_time": 4832.807,
"text": " Interpret it incorrectly. I find it to be controversial whenever someone says free will exists or free will doesn't exist I'm curious like how are you making that without a plethora of assumptions? That's also a metaphysical claim. We both agree It's a metaphysical claim and maybe you don't agree with this, but I don't think science makes metaphysical claims I think science is an instrument and if anything, it's Instrumentalism here's what science is just if then statements it says if you do this this happens doesn't say behind this is"
},
{
"end_time": 4880.742,
"index": 184,
"start_time": 4859.872,
"text": " So here's a way to cut the Gordian knot. So I think we're agreed that free will, I view it as either"
},
{
"end_time": 4906.459,
"index": 185,
"start_time": 4882.09,
"text": " Total logical in the sense that if it is a psycho neuropsychological phenomenon, then people do have that feeling of free will. Yes, they do. Just like people have a feeling of being happy or sad. And these are things that you that neuro biologists can put people into fMRI scanners and try to drill down. If it is instead use a metaphysical statement, then I don't know what it means."
},
{
"end_time": 4926.476,
"index": 186,
"start_time": 4907.278,
"text": " And that might, in fact, some people might define it to be not as mean, unless we try to go across some Turing computability and so on. Now, in terms of whether science and metaphysics,"
},
{
"end_time": 4958.302,
"index": 187,
"start_time": 4929.616,
"text": " What I mean is, does science ever make a claim to ontology? Instrumentalism or not? Yeah, yeah, yeah. I mean, it's a subtle thing. So instrumentalism, one way to interpret it, and like all names of schools and philosophy, it's got various flavors and proponents and so on. But it's the view that there is some ill-defined reality out there which produces effects on our experimental apparatuses."
},
{
"end_time": 4987.381,
"index": 188,
"start_time": 4959.258,
"text": " And what science is, is a way of trying to actually relate the patterns in those effects in our apparatuses and maybe make predictions about other patterns we might see in future apparatuses without saying anything about this underlying physical reality. That an electron is a construct for describing these patterns and there is some other external reality out there"
},
{
"end_time": 5016.596,
"index": 189,
"start_time": 4987.978,
"text": " that the word electron is not necessarily relevant to it at all. And certainly, since science in its current instantiation is progressing, it could not make any claim that it is reality. But here's a different perspective, which is implicit in what's called ontic structural realism, also in some of Max Tegmark's more careful work and like his Analysts of Physics paper,"
},
{
"end_time": 5043.933,
"index": 190,
"start_time": 5017.363,
"text": " and so on rather than his pop work, and which also underlies some of the ways that my collaborators have been pushing the formalism in noise deterministic reasoning, which also relates to what I was talking about before, the problem. So here is a very simple, not that simple, but here is a hypothesized"
},
{
"end_time": 5074.275,
"index": 191,
"start_time": 5047.773,
"text": " Mathematics, which in well and fell swoop explains all explains away all of the conundrums of philosophy together with science. And here it is. It's only mathematics all the way down. That there is no external reality. Everything is just a mathematical system."
},
{
"end_time": 5104.002,
"index": 192,
"start_time": 5075.623,
"text": " and that one mathematical system is in fact just as legitimate, just as real as any other one. It's that the reason a chess piece on a chessboard feels the laws of chess to be real rather than the laws of checkers is simply because it's having to right now be viewed as a token in a game of chess, but it could just as easily be in a game of checkers. From that perspective then, there is a mathematical system"
},
{
"end_time": 5134.104,
"index": 193,
"start_time": 5105.247,
"text": " which would have as some of its equations human beings looking out at experimental universes around them and noticing patterns which might be approximations in some sense or not of this underlying mathematics, but underlying it all is just that mathematics. So it's somewhat stepping to the side of saying"
},
{
"end_time": 5160.947,
"index": 194,
"start_time": 5134.838,
"text": " It's a single monolithic edifice that explains both philosophies, humans, our instruments, the patterns, and underlying reality. It's all just one single thing, and there is no room for anything else. It is informed by science, but it is not actually somehow something that is proven by science."
},
{
"end_time": 5186.408,
"index": 195,
"start_time": 5161.527,
"text": " You know how you said that you're unsure what it means for free will to have the metaphysical notion of free will? What the heck does that mean? I'm unsure what it means that mathematics is the ultimate reality. I don't understand Max Tegmark's ideas and I'm also unsure if you're claiming that you like those ideas and that's the one that you see as fitting reality most. I view it as in one fell swoop solving"
},
{
"end_time": 5206.613,
"index": 196,
"start_time": 5187.466,
"text": " Well, it doesn't completely solve the, the conundrums because for example, laws of mathematics, um, girdles and completeness theorem says there is no ultimate, um, uh, unviable consistency in the laws of mathematics themselves. So hear that sound."
},
{
"end_time": 5233.66,
"index": 197,
"start_time": 5207.585,
"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": 5259.787,
"index": 198,
"start_time": 5233.66,
"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."
},
{
"end_time": 5285.555,
"index": 199,
"start_time": 5259.787,
"text": " 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 slash theories, all lowercase."
},
{
"end_time": 5314.411,
"index": 200,
"start_time": 5285.555,
"text": " Go to Shopify.com slash theories now to grow your business no matter what stage you're in Shopify.com slash theories. It eats its own tail. And if one then says noisy, um, deter, um, deductive reasoning, you then get an even worse this oration of the foundations of that statement that all is mathematics disappears. What do you mean? Cause there's no such thing as a mathematics, but here's the sense in which"
},
{
"end_time": 5343.131,
"index": 201,
"start_time": 5315.196,
"text": " So those are extra extensions of the underlying first concept, which are very, very important, but before getting to them, before invoking Gödel and so on. I view ontic structural realism, Tegmark's elaborations, as explaining it all in the sense of quantum physics, that you, Kurt,"
},
{
"end_time": 5372.602,
"index": 202,
"start_time": 5343.951,
"text": " are part of a wave function that involves the screen you're looking at and everything else. It's the wave function of the universe and all aspects of your neurobiology and therefore your subjective notion of free will. All aspects of your scrutinizing the results of experiments involving what in your lab notebook is called an electron. All of that is arising"
},
{
"end_time": 5401.34,
"index": 203,
"start_time": 5373.08,
"text": " as simply solutions to Schrodinger's equation and that there is nothing else besides that. There is no underlying concrete reality that your mathematics is describing. It suffices to just dispense with it. You don't need that concrete reality. All those mathematics is not the mathematics we currently have in our textbooks. But that's fine because our textbooks themselves are simply"
},
{
"end_time": 5427.568,
"index": 204,
"start_time": 5401.817,
"text": " some of the characteristics, the projection operators of the quantum physics of the wave function of the universe. Have you ever heard the phrase, we shouldn't mistake the map for the territory? I don't believe I've heard the phrase. Yes. Or it's actually it's sometimes the map and the territory. Yeah, don't mistake them. Right. That's right. What did I say?"
},
{
"end_time": 5457.432,
"index": 205,
"start_time": 5427.91,
"text": " You said the map and the territory. No, I meant the map if I misspoke. Yeah, the map and the territory because in that case there's stuff that's not in the map that is in the territory and you can experimentally ascertain that. This point is that the map that we currently have is not the ultimate map but the ultimate map so to speak. The Ur map has in it, this is almost like a Borgesian kind of a"
},
{
"end_time": 5482.517,
"index": 206,
"start_time": 5457.773,
"text": " How do we know that all is math?"
},
{
"end_time": 5508.114,
"index": 207,
"start_time": 5482.892,
"text": " We don't. Remember, I was very, very careful the way to phrase this, that one perspective, which is consistent and simple, it doesn't invoke unseen, as since going back to Kant, at least, Western philosophy has been prone to doing, or Plato, for that matter, with his Platonic ideals, it doesn't invoke all those unseen, because it's pointing out"
},
{
"end_time": 5538.268,
"index": 208,
"start_time": 5508.916,
"text": " that ultimately it's by flipping the perspective to say that we are phenomena generated by it rather than it is something external to us that we are trying to understand. It's a flip of the perspective. You start with the it and we are manifestations of it, projection operators as they call them in physics, of those pieces of the it. Rather than starting from ourselves in almost a Copernican and pre-Copernican"
},
{
"end_time": 5562.619,
"index": 209,
"start_time": 5538.66,
"text": " pre Copernican view of the earth as the center. We are the center trying to infer the outside territory. No, we are ourselves just one small little map in this much bigger map. And from that perspective, you don't need anything outside and everything is explained. All of these conundrums, you know, no free lunch. How does that allow me to even get going?"
},
{
"end_time": 5587.79,
"index": 210,
"start_time": 5563.404,
"text": " Yep, you're correct, little boy, that no free lunch doesn't allow you personally to get going. But that doesn't change the fact that here is a way that here is a, I'm not even sure what you would call it. It's not saying it's a reality. It's not saying it's a hypothesis. It's just a edifice of"
},
{
"end_time": 5619.104,
"index": 211,
"start_time": 5590.384,
"text": " interrelated explanatory phenomena of mathematics, which we don't even currently have, but which would actually do a jiu-jitsu move of all of our philosophical conundrums by simply saying that they are what come up in the neurobiology of these particular little systems that are being generated as part of the solution to the Schrodinger equation when they try to look outward at things outside of themselves."
},
{
"end_time": 5643.541,
"index": 212,
"start_time": 5620.179,
"text": " Sir, how would you respond to my thought, which is it sounds like we have quantities and then we say, well, all that exists are these quantities and not everything is quantitative. However, it seems to that's what I mean when I say that. Sorry. That's my that's my rendition of all is math. However, it seems like there are large classes of our experience which are qualitative that don't lend themselves, at least not readily."
},
{
"end_time": 5671.664,
"index": 213,
"start_time": 5643.916,
"text": " To me, it sounds like a large statement to say all of what we think of as qualitative, so anger, the sight and experience of red and blue, etc. So you're talking about qualia? Yeah, yeah. Not everything lends itself to a quantitative analysis. Okay, but are you specifically talking about like charmers and these people, their notion of what they call qualia? Sure. So I view that"
},
{
"end_time": 5702.449,
"index": 214,
"start_time": 5673.609,
"text": " conundrum is ultimately, ultimately vacuous. We think that we feel something. I mean, famously, Dan Dennett had a great put down of the paper. What does it feel like to be a hummingbird? I think it was that very famous paper. Dennett eviscerates it. But ultimately, the question, what is it"
},
{
"end_time": 5733.865,
"index": 215,
"start_time": 5703.968,
"text": " I don't see if there's any room for that question any more than there's any room for our belief, for one hemisphere's belief that's in control of the other one. That belief is wrong. It is demonstrably wrong. Neurobiology can presumably in some future incarnation explain why that hemisphere would come to that conclusion after first neurobiology has to even"
},
{
"end_time": 5761.664,
"index": 216,
"start_time": 5734.309,
"text": " figure out what it means to say that a particular piece of brain tissue is coming to a conclusion or not. But it is wrong. It has got nothing to do with reality, that qualitative experience of that hemisphere being in control. No, it's not. But we can now expect this future incarnation of neurobiology would be able to explain all aspects. Whatever question you might ask,"
},
{
"end_time": 5791.783,
"index": 217,
"start_time": 5762.637,
"text": " of that subjective experience, it would be able to provide the answer that there's not necessarily any room for anything else. You said this is yet to be demonstrated. We just see you see this as reasonable. Yes, neurobiology has not done this by a long shot. But I'm just saying that what's consistent with everything that we now know with everything that we can think"
},
{
"end_time": 5822.125,
"index": 218,
"start_time": 5792.312,
"text": " Let me tell you my thoughts quickly, I'm sorry to interject. When it comes to neurobiology and then we have neural correlates of certain qualitative experiences and then you can say those correlates are quantitative, so we found an association between quantitative and the qualitative, so therefore the qualitative ultimately reduces to the quantitative. Those ultimately rely on many, many, many, many, many reports of subjective experiences. So for example, I'll have to tell you"
},
{
"end_time": 5846.698,
"index": 219,
"start_time": 5822.5,
"text": " I flip the question on you."
},
{
"end_time": 5876.459,
"index": 220,
"start_time": 5848.916,
"text": " I wouldn't even say these are claims. My statements do not have to do with current neurobiology, or even necessarily that neurobiology will get there. It's much more the following. It's a challenge to you. If, in fact, there is something inherent about the subjective that"
},
{
"end_time": 5905.316,
"index": 221,
"start_time": 5877.261,
"text": " would not be reducible to mathematical properties. Given, especially in light of the Bin-Libet experiments, the split-blades and all the other related zombie within kind of phenomena, how could one even speak about it without reducing it to being a"
},
{
"end_time": 5934.616,
"index": 222,
"start_time": 5906.118,
"text": " mathematical thing what is even it how can we even say anything how can we have a discussion with unless it's going to be defined precisely and i don't know how to define something precisely without defining it precisely it's it's a challenge how do you other than sitting around rather than having an aesthetic experience which we can have how do we actually engage in it in a reasoning manner"
},
{
"end_time": 5963.865,
"index": 223,
"start_time": 5935.964,
"text": " Let's have fun here. When you're saying that you don't know how we can speak unless we define our words precisely, well,"
},
{
"end_time": 5992.637,
"index": 224,
"start_time": 5964.957,
"text": " You and I can speak about viewing art. For example, right now you've got a guitar back there. I presume that you play. I don't wish I could but I could imagine that we could"
},
{
"end_time": 6020.06,
"index": 225,
"start_time": 5992.91,
"text": " have an interaction concerning music. And if in a different life where those particular regrets returning to a previous theme did not exist in me, I might be able to have some facility on and we could do it. We could jam together or whatnot. That would be us communicating. It would all be legitimate. It would be valid. It's a different form of communicating from when we are trying to ascertain jointly"
},
{
"end_time": 6044.77,
"index": 226,
"start_time": 6020.538,
"text": " using reasoning processes, something concerning, I wouldn't even use the word existence because that by itself is a fraud exercise, but so it's a different realm of discourse. You and I can very definitely talk about what does it feel like to experience beautiful music,"
},
{
"end_time": 6069.735,
"index": 227,
"start_time": 6045.691,
"text": " What does it feel like to, I mean, red is a simple minded one, but to experience beautiful art. I have often described how to me, uh, I have been ambushed in my life by paintings, um, turn a corner in a gallery or a museum. And literally the phrase that keeps coming to my mind, I keep repeating is I'm being ambushed. I'm just knocked flat. I'm stunned by it."
},
{
"end_time": 6098.422,
"index": 228,
"start_time": 6070.384,
"text": " It's a physical reaction by that actual painting. To give you a simple example, it's things like Monet's and Van Gogh's. Because it's so beautiful that you're in awe? Cézanne's in particular. I never particularly appreciated Cézanne until when I first saw one in the flesh, then it ambushed me because there's so many details of it that just don't reproduce well. When you read about a Cézanne piece, Simon's Scent, I think it was,"
},
{
"end_time": 6123.558,
"index": 229,
"start_time": 6098.712,
"text": " in the art books, it's all intellectual that he had this effect on art history and so on and so forth. Okay, so you can read it like that. But then when I saw it, wow, I was knocked out of my socks. And then more generally, I can talk with you or anybody else about how for me to actually judge a painting, the way that I judge it, the process I go through is what can it teach me?"
},
{
"end_time": 6151.408,
"index": 230,
"start_time": 6124.275,
"text": " If I myself were trying to redo that painting and try to change some things, would I actually learn, no, wait, you can't do that because of these aspects that you weren't even sensitive to. Does it have its own three-dimensional reality, that painting, or is it just a flat caricature? I don't mean necessarily sense of perspective. I mean, does it exist? I understand. So we can have that discussion."
},
{
"end_time": 6181.408,
"index": 231,
"start_time": 6152.21,
"text": " Does that have anything to do with Schrodinger's equation? Well, I mean, you know, ultimately, if you want to flip things around, yeah, sure, David, you can have a different set of discussions where you're saying that these aesthetic experiences in your head are just neurobiology and so on and so forth. Sure, you can. And if you want to, go off and do it. But that's a different conversation we're having from talking about that art piece, Ambushing Me. It's irrelevant to talk about it in terms of Schrodinger's equation. Sure, yes, but who cares?"
},
{
"end_time": 6205.094,
"index": 232,
"start_time": 6181.8,
"text": " That's got nothing to do with the actual discussion we're having about what art pieces have that they can teach you. Conversely though, if I am instead in a different mood, I've woken up on a different side of the bed and I'm discussing philosophy, formal reasoning, what is the ultimate nature of reality, then"
},
{
"end_time": 6232.398,
"index": 233,
"start_time": 6205.486,
"text": " all of the other stuff about being ambushed it just the question gets reduced to is all of that in theory at least explainable or is there something there that is not even in theory explainable as being a meta as an epiphenomenon an effect of mathematical Schrodinger's equation. It's a completely different way of of engaging with the same subject domain."
},
{
"end_time": 6252.944,
"index": 234,
"start_time": 6232.961,
"text": " And what I'm saying is that when I'm over there wearing my metaphysics hat, if I'm just going to push metaphysics to be what I personally, David Walport, I'm interested in, which is all the way down the road reason I see, I see. Okay, I see. So let me put it like this. It sounds like what you're saying is if one is making a claim that"
},
{
"end_time": 6277.346,
"index": 235,
"start_time": 6253.387,
"text": " your brain, David, that you have some free will and it supervenes the laws of physics. Sorry, it doesn't super, the physics supervenes on that your free will somehow determines the laws of physics, at least momentarily enough for a neuron to fire differently. And so therefore you have free will. So you have some, your free will has some ontological status. And then you're saying, well, okay, where along the chain from electrons to cells to"
},
{
"end_time": 6306.715,
"index": 236,
"start_time": 6277.671,
"text": " Neurons to, well, neuron is this type of stuff, to the whole brain configuration, to your behavior, where on that chain was physical law broken? Because you can look at any given one of those chains and you can say, well, look, there's no violation of physical law. And so if the nebulous person is making the claim about some nebulous concept like free will, you'd better show me where it comes in. So far, the laws of physics are not violated at any level as far as we've examined it. Is that what you're saying? Essentially, yeah."
},
{
"end_time": 6312.637,
"index": 237,
"start_time": 6307.517,
"text": " As opposed to do I have a feeling of free will? Yes, I do."
},
{
"end_time": 6342.841,
"index": 238,
"start_time": 6313.268,
"text": " okay now what I would say to that is I'd still say that we don't know and I don't think you're making the claim that you know either so I don't think that we're speaking differently let me put a bit of a monkey wrench so when we look and we examine at any one of these features so we look at let's say the electron bound with a proton we say there are no laws of physics being violated here this is all explainable with what we have currently as our models or at least we have the idea that some other model may exist in the future and then cells and so on so on it's as if you're looking individually at small parts but still"
},
{
"end_time": 6371.271,
"index": 239,
"start_time": 6343.609,
"text": " The counterclaim could be that you show me that the laws of physics are obeyed for someone making a decision across the whole spectrum and not looking at it individually across different people in different circumstances and saying here is now violated. It would be almost as if what you're doing is looking at a small billiard balls bouncing around in a small environment saying there's no time here. There's no time here. There's no time here. But yet you're saying you're saying that there's time at the higher level. There can be. There can be a merger. I mean, Phil Anderson's more is different."
},
{
"end_time": 6401.869,
"index": 240,
"start_time": 6371.903,
"text": " You can have laws, I mean, that's what condensed matter physics is all about. Parisi just won the Nobel Prize for spin glass models, which is thermodynamic limit, an infinite number of phenomena that arise strictly speaking only when you have an infinite number of these interactors, but it's all being governed by mathematics. And so you're correct. I'm not saying that I now, and maybe not humanity ever, would even be able to fill in all those steps."
},
{
"end_time": 6430.862,
"index": 241,
"start_time": 6402.858,
"text": " I'm saying that there is a consistent picture in which those steps are part of a mathematics, which may forevermore be beyond humans. But I'm then going a little bit further and saying, give me provide for me anything other than essentially an elaborated version of"
},
{
"end_time": 6458.575,
"index": 242,
"start_time": 6433.234,
"text": " describing the experience that you have when looking at a beautiful painting, something that is not mathematics. I don't even see what one could call free will. I just don't even understand what else there could be besides mathematics and then aesthetics. I don't see that there's"
},
{
"end_time": 6490.23,
"index": 243,
"start_time": 6461.613,
"text": " I don't see how we can be doing reasoning, reasoning rigid. Iron cloud, we know that with the things that iron clouds can come up with, given all this stuff about noise deterministic reasoning, deductive reasoning, but how we can actually come to any kinds of conclusions if we're not using logic, let me say that second order logic, that's part of mathematics."
},
{
"end_time": 6516.834,
"index": 244,
"start_time": 6491.015,
"text": " Second order logic is part of mathematics and I don't see how we can, we can't even define the terms in something like what is called in mathematical logic a language where you've got an associated set of axioms and so on and so forth. I don't even see how we can be speaking about it in those terms. I don't see what philosophy can be if it's not that."
},
{
"end_time": 6536.476,
"index": 245,
"start_time": 6518.387,
"text": " I see. When we're talking about how some of these philosophical claims are volutinous, they're bleary, they're opaque, dubious, you don't know what the heck it means because they're speaking ambiguously. And then you're saying, well, I don't know what it means. I understand what mathematics would mean if mathematics was fundamental. For me,"
},
{
"end_time": 6563.046,
"index": 246,
"start_time": 6537.381,
"text": " Please help clarify for me. What the heck does that mean? Because to me, that's just equally, I can understand the statement that mathematics describes what we've seen so far. But then to say that mathematics is what we've seen so far, I don't understand what that is is what is the mathematics and then when we say it's a rule, what is a rule we can play that game. So can you help me understand the ontology? Um, this is the"
},
{
"end_time": 6591.63,
"index": 247,
"start_time": 6564.121,
"text": " So first of all, assuming that we're not worrying here about things like Girdle's incompleteness theorem or the other things about math eating its own tail. Sure. We're not trying to do, for example, Pressberg arithmetic, which would allow you to avoid Girdle's incompleteness theorem or anything. That's one of the things that I find so deep"
},
{
"end_time": 6621.186,
"index": 248,
"start_time": 6592.91,
"text": " flaw or attribute in me is that the more I can be smashed on my ass by the universe realizing that I have been misconstruing my limitations for being those of the universe the more I feel I'm making progress the more I actually that's my aesthetic"
},
{
"end_time": 6653.302,
"index": 249,
"start_time": 6625.794,
"text": " I don't think I can answer your question and that to me, and I understand your question and feel your question. And the fact that I cannot answer it to me is a beautiful, stunning illustration of the fact that the deficiency is in me. It's a feature, not a bug. It's a feature, not a bug. Yup. Um, things like, um,"
},
{
"end_time": 6683.763,
"index": 250,
"start_time": 6654.053,
"text": " I don't know how much you know about algorithmic information complexities. No, I have a question about that. I would like to get to that later. Some of algorithmic information complexity, I view it as the most, the only true philosophy metaphysics, and by far the deepest that we've ever come across. Gödel was, his result was one of the most important philosophical results, precisely because, to my thinking, aesthetically, because it was a"
},
{
"end_time": 6711.817,
"index": 251,
"start_time": 6684.889,
"text": " incontrovertible demonstration of how all that we were so sure, we all thought the same as David Hilbert did, and we were wrong. And whether we can understand it or not is irrelevant. And this is a similar kind of a thing. I do feel that question you're asking and the fact that I don't understand what it means to give it as an answer,"
},
{
"end_time": 6729.462,
"index": 252,
"start_time": 6712.978,
"text": " doesn't mean that that conclusion, which is causing that question of yours, that the conclusion is wrong, it's a limitation in me. That I can't grapple with what the conclusion even means, just like Hilbert couldn't grapple with what Gödel meant."
},
{
"end_time": 6757.056,
"index": 253,
"start_time": 6731.237,
"text": " Okay, speaking of limitations, I know you said that you have to, you're keeping an eye on the clock. So how about I give an outline of what we have left to talk about and we can determine how we're going to parse our time to most efficiently get through these. Okay. Okay. And I have, we're not even third way done so that we'll have to pick and choose some of these. Or would it make sense to have a follow up? We can definitely have a follow up. Yeah. Given the time, how about we then do"
},
{
"end_time": 6777.125,
"index": 254,
"start_time": 6757.824,
"text": " I'm not sure how you want to do the video splicing on your end. We'll just wear the same clothing and pretend it's all one session. Yeah, we could do that. Much much of my clothing this time of year tends to be the same anyway. Yeah. But yeah, but if we could then just schedule another time. Yeah."
},
{
"end_time": 6787.927,
"index": 255,
"start_time": 6777.381,
"text": " Okay, so I'm assuming you have to go in about 10 minutes. Is that correct? Yeah. Okay, so let me at least go through the questions and then we don't have to answer them. But let me just outline to Laplace's demon I wanted to know."
},
{
"end_time": 6812.005,
"index": 256,
"start_time": 6788.37,
"text": " about how your inference limits on inferences, limits on inference devices, what that has to say about Laplace's demon, but outline that for the audience. And then Joseph Rakoveka apparently has a substantially simplified argument involving Turing machines and free will, which like we just talked about free will, I haven't looked at his, and I want to know if you could comment like what was his argument? How is it simpler? I don't I don't know that particular one. Let's forget about that then."
},
{
"end_time": 6827.466,
"index": 257,
"start_time": 6812.449,
"text": " Cross-validation. How can cross-validation be used to exploit the scientific method? I think that one we should definitely do before you leave. And then we've got to get to the definition of observers. We can save that for later because that's quite lengthy conversation. And then I wanted to talk about Kolmogorov complexity"
},
{
"end_time": 6849.104,
"index": 258,
"start_time": 6827.466,
"text": " And you mentioned that there are different bounds, which was extremely interesting on Kalmogorov isn't bounded. And that's intuitively obvious why it's not bounded. But then thermodynamic complexity is bounded. Well, I don't know what thermodynamic complexity is. And I wanted you to explain to me in the audience what it is. And then I also wanted to know if you know about assembly theory. Have you heard about assembly theory?"
},
{
"end_time": 6866.817,
"index": 259,
"start_time": 6849.718,
"text": " You mean the stuff of annoyments? No, no. It's recent. It's by Lee Cronin, who's a chemist. Oh, oh, Lee's stuff. Yeah, oh yeah, Lee is quite the... He is quite a character, Lee. Yeah, yeah, yeah. Okay, so he's not calling it assembly theory, is he?"
},
{
"end_time": 6882.056,
"index": 260,
"start_time": 6867.108,
"text": " Okay, logical irreversibility versus thermodynamic irreversibility, what's the difference between those? Then I also wanted to know if the way that you can quantify an observer, which is extremely interesting, is that related to a notion of intelligence? So we could talk about that later."
},
{
"end_time": 6909.224,
"index": 261,
"start_time": 6882.944,
"text": " You also mentioned the most profound results in philosophy come out of algorithmic information theory, and that's where my question for you was, what the heck is algorithmic information theory? Can you give some examples? Why do you say that it's the most profound? And then we have several audience questions. So we have Edward Lee of Berkeley. Professor Kevin Knuth has a question for you. He has quite a few, and he also said to say hi to you. Okay, I'll say hi back."
},
{
"end_time": 6935.52,
"index": 262,
"start_time": 6909.565,
"text": " Let me put it this way. This is some of the most"
},
{
"end_time": 6967.142,
"index": 263,
"start_time": 6937.432,
"text": " near and dear to my heart would be a glib thing. I view these as like the most profound things that any human minds can be engaging with. And I am sure that just as in so many other things, I'm just full of fecal matter, shall we say, about tons of things. And if I don't, you're helping to have these kinds of considerations and cognitive wrestlings"
},
{
"end_time": 6998.029,
"index": 264,
"start_time": 6968.422,
"text": " um uh conveyed out to a broader set of people who can join this mosh pit of wrestling with these things the more the better more power to you independent of me let me fall under a bus tomorrow but these are the well let me not but whatever but these issues i think yeah they should be engaged with and i think that's you know i am"
},
{
"end_time": 7006.51,
"index": 265,
"start_time": 6998.541,
"text": " Something I do passionately agree with, and I thank you very, very much for helping to have them be engaged with in a broader."
},
{
"end_time": 7033.643,
"index": 266,
"start_time": 7007.381,
"text": " Well, thank you for providing me some material to engage with. Okay, now with regard to cross validation and exploiting the scientific method, why don't we get to that quickly and then we'll wrap up for today. Okay, so you said that in one of your talks, we can use cross validation to exploit the scientific method. First, what is cross validation, you have to explain that to the audience. And then how does it undermine science? Yeah, it's actually I'm not sure what the exact what was it basically cross validation is the scientific method formalized. So"
},
{
"end_time": 7059.121,
"index": 267,
"start_time": 7034.633,
"text": " One aspect of the form of the scientific method, like so many glib phrases, it's glib and leaves a lot out. The scientific method, we put far more posterior probability, we assign far more credence in a theory if it predicts phenomena that we have not already observed and then are found to be true."
},
{
"end_time": 7090.316,
"index": 268,
"start_time": 7060.503,
"text": " So, you know, famously, for example, the Eddington and Einstein, the observation of during the eclipse and so on and so forth. So another way to say that is that if I give you two theories, A and B, and I ask both of them to make a prediction about something that neither of them has yet seen, A predicts it accurately and B does not, then we say let's go with A on making any subsequent predictions."
},
{
"end_time": 7120.93,
"index": 269,
"start_time": 7091.527,
"text": " I'm going over and above. And so that can be viewed as part of the scientific method. Now, let's see what that might mean in the context of machine learning. Well, if I want to choose among a bunch of machine learning algorithms, one way to do it is to give them all a bunch of data and hold out some data, let them train on the data they have, and then just see which one does best on the data it doesn't have."
},
{
"end_time": 7150.759,
"index": 270,
"start_time": 7121.834,
"text": " And this is actually done in contests, bake-offs, they're sometimes called, all the time. You know, we're going to choose between the following character recognition algorithms by giving them all of this data set, pulling out some other examples, and then seeing who does best. You can also, though, do that yourself, so to speak, by saying, I've got myself, I've got a bunch of different algorithms that I can use. I've got a fixed data set. Let me"
},
{
"end_time": 7177.381,
"index": 271,
"start_time": 7151.374,
"text": " play like I'm holding a bake-off, train all these algorithms, a part of my data set, on part of my data set, see which one does best on my held out data, take that one, now train on everything and use it to make my subsequent predictions and use that as to choose among different algorithms, just me myself as a solo machine learner, or to set parameters of a single algorithm."
},
{
"end_time": 7207.79,
"index": 272,
"start_time": 7178.592,
"text": " That what I've just described to you is what's called cross validation. It is part and parcel of machine learning algorithms. It is part and parcel of the scientific method. It does not get around the no free lunch theorems. That in point of fact, you can actually show, and this is some really strange stuff that I would be very happy to talk about, but it'll require some care in the discussion, that the"
},
{
"end_time": 7232.551,
"index": 273,
"start_time": 7208.49,
"text": " algorithm of anti-cross validation, of choosing the scientific theory that does worst at predicting, there are just as many priors for which that is going to outperform the scientific method as the other way around. So very strange, it almost makes no sense when I just say it that way."
},
{
"end_time": 7259.667,
"index": 274,
"start_time": 7233.029,
"text": " So what would be an example of a statement that we believe is scientifically true, but you believe it was founded by inadvertently exploiting something? Everything in science. Whenever I cross the street, everything, this is what I meant by why the no-freelance theorem is sometimes just my notes to myself, TP. Everything. The anti-reasoning by no-freelance, anti-reasoning rather than reasoning."
},
{
"end_time": 7289.684,
"index": 275,
"start_time": 7260.196,
"text": " in terms of inductive reasoning. Let me state what cross validation is one more time simply to the audience so that some of the people who weren't able to follow can follow. Cross validation, if you've heard on a previous podcast, I believe I talked about Kryptonstein, there's a rule following paradox, so it's similar to this. Imagine you have a kid, a child, you're teaching how do you add digits. So you have a set of a thousand cards and each of them is an arithmetic statement, a statement about arithmetic. So two plus five is seven, three to the power of four is"
},
{
"end_time": 7318.643,
"index": 276,
"start_time": 7290.742,
"text": " 81 and then you show the child the set of statements that are correct and you show them the set of statements that are incorrect so some of these are true plus five equals minus one and then you say and then you want to find out is the child following a rule but the child could have simply if you give them the whole deck the child could simply memorize all of these cards and you can show them a card two plus three equals five and you can say is this correct and they'll be like that's correct and then you don't know are they following the rules of arithmetic or have they simply"
},
{
"end_time": 7344.07,
"index": 277,
"start_time": 7320.282,
"text": " So one way you can get around that is you take, let's say, 70% of your cards and you say, learn some rule, child. I'm not even going to tell you what the rule is. That's the point. I want you to learn, figure out how to add by looking at these examples. And I'm saving 30% of these here so I can show you new cards and say, is this one correct or incorrect? And then judging by that, I can say whether or not you're following the rules correctly."
},
{
"end_time": 7371.988,
"index": 278,
"start_time": 7344.514,
"text": " Okay, so that's what cross validation is. Now, how does that undermine science? So that's essentially the scientific enterprises you've outlined it. So how does the no free lunch theorem relate to that? Say it one more time, please. It sounds like what you're saying is, well, I could come up, the child can come up with any rule and it would still be, I could still find some situations in which that rule is valid. Just as many. And in fact, you could have whatever the child"
},
{
"end_time": 7400.811,
"index": 279,
"start_time": 7372.5,
"text": " says, I could define a new algorithm which is take whatever the child says and go with the opposite, so to speak, and there are going to be as many priors, as many universes, we're talking about the scientific method, in which that anti-algorithm is actually going to end up doing best on yet unseen situations rather than the original ones."
},
{
"end_time": 7425.862,
"index": 280,
"start_time": 7402.517,
"text": " and it becomes very, very pointed when it's things like... It means that, for example, there are going to be as many situations in which your algorithm does worse than random guessing as in which it does better, no matter what your algorithm is."
},
{
"end_time": 7456.391,
"index": 281,
"start_time": 7426.476,
"text": " Now, this implies to me that scientific methods should not work, yet it seems to. So why? The problem. Yep. The problem. TP. Because your very statement, it seems to, is based upon the data set that has worked so far. But then when we go to off-train, the no-free lunch says doesn't mean the fact that it's worked so far well in the past has anything to say about how well it will work in the future. It's just like evolution has worked so well in the past, has no implications about how well it will work"
},
{
"end_time": 7482.841,
"index": 282,
"start_time": 7456.971,
"text": " in the future. So it eats its own tail. The fact that anything that we have used go up to a meta level, a meta algorithm of some sort, where I'm choosing between strategies for choosing between strategies doesn't matter. In fact, there's a philosopher, Carl Schurz, who just came out with a book on meta induction, it's making a big"
},
{
"end_time": 7508.933,
"index": 283,
"start_time": 7483.473,
"text": " um splash or waves you know whatever the um aquatic metaphor of choice and um he's got a whole chapter on no free lunch and i actually just wrote a commentary um on that there were some that were solicited and it's going to be appearing in um the journal general philosophy of science or something like that i forget the precise name where he claims that his results um get around no free lunch but no they don't"
},
{
"end_time": 7539.906,
"index": 284,
"start_time": 7510.23,
"text": " He's talking about meta induction at this level. So you're doing induction about induction about induction to try to essentially break the Gordian, you know, break free your binders, but they don't break free the binders, you're still stuck. And we have no reason to believe the scientific method. But I sure as hell believe it. Listen to the earlier part of this podcast, I was going a whole hog, I was the most dyed in the wool, conservative dogmatic scientific reasoner you could come up with."
},
{
"end_time": 7570.23,
"index": 285,
"start_time": 7540.247,
"text": " And then the next breath, I'll tell you that it's got no legitimacy. By the way, would you consider math to be scientific or do you consider science to be separate from math? Oh, this is like you asked me before about being an artist versus and so on and so forth. I don't I don't make those kinds of distinctions. I will. Yeah. The reason I was asking is if you did consider it, then it would be as if what you're saying is by using science, I've undermined science."
},
{
"end_time": 7597.398,
"index": 286,
"start_time": 7571.408,
"text": " Girdle, by using math, undermined math. Yep. Well, Girdle undermined a part of math, whereas you're undermining the entire scientific enterprise. Yep, pushing it further, pushing it further. And then things like the inference devices work, go even further. They, in some senses, undermine the notion of epistemology, full stop."
},
{
"end_time": 7626.63,
"index": 287,
"start_time": 7598.063,
"text": " of even knowing something, no matter how that information is arrived at, even by luck. And this gets to another very, very much tongue-in-cheek definition of free will, which was very, very much tongue-in-cheek. And then the stuff on noisy deductive reasoning would say that, oh, but if math itself, including that reasoning that you're using to come to these other theorems, is itself a noisy process inherently,"
},
{
"end_time": 7661.664,
"index": 288,
"start_time": 7632.005,
"text": " Well, you got to get going. And so for the people who are watching, if you have any questions for David, he'll be coming back on the theories of everything podcast at some point. So leave your questions below. We'll be talking about the same. Well, forget about that. I can add I can change that later. Professor, thank you so much."
},
{
"end_time": 7692.329,
"index": 289,
"start_time": 7662.363,
"text": " Well, thanks. This is a lot of fun. All right, we'll do. Well, thanks very, very much. Yeah, well, thank you, man. Interesting. And the more actually from my sitting in my chair, the more I actually feel that you, Kurt, are engaging in actually revealing what it is that you're actually thinking. That makes it more interesting, more what? More, more interesting, but also more stimulating."
},
{
"end_time": 7715.674,
"index": 290,
"start_time": 7692.671,
"text": " Again, recall that this was part one of a two-part conversation."
},
{
"end_time": 7726.869,
"index": 291,
"start_time": 7715.947,
"text": " In this second part, we take audience questions, as well as hearing from Kevin Knuth, Annel Seth, Karl Fursten, and Chris Langan, who all have questions posed to the great David Wohlpart."
},
{
"end_time": 7748.268,
"index": 292,
"start_time": 7729.087,
"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.