Audio Player
Starting at:
Bernardo Kastrup, Donald Hoffman, Susan Schneider: Debate on Consciousness
March 10, 2023
•
33:36
•
undefined
Audio:
Download MP3
⚠️ Timestamps are hidden: Some podcast MP3s have dynamically injected ads which can shift timestamps. Show timestamps for troubleshooting.
Transcript
Enhanced with Timestamps
82 sentences
5,277 words
Method: api-polled
Transcription time: 32m 16s
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 was just launched this month it gives you gives me a front row access to the economist internal editorial debates where senior editors argue through the news with world leaders and policy makers and 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.
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 plants where everyone
The Institute
for Arts and Ideas produces numerous debates, lectures, and interviews on a broad range of topics including philosophy, science, politics, art. They asked me to host a debate of a sort between Donald Hoffman, Bernardo Kastrup, and Susan Schneider on whether machines are or can be conscious.
This is part one, and the rest of the episode can be viewed on the IAI.tv website, as well as on their YouTube channel, both linked in the description. Also, I was asked to host an interview one-on-one between myself and Donald Hoffman, and that part one should be exclusively on the Toe channel shortly, as well as the rest on the IAI website slash YouTube channel. Enjoy. My name's Kurt Jaimungal, I'm the host of
The YouTube channel Theories of Everything, which analyzes tolls from a theoretical physics perspective. And it's my distinguished pleasure to host this talk with Donald Hoffman, Susan Schneider and Bernardo Castro. Donald Hoffman is an American cognitive psychologist working at the University of California, Irvine. He's making waves with his new theory, suggesting that instead of presenting reality as it really is, quote unquote, that our perception is tantamount to a desktop user interface, enabling us to use reality effectively.
Bernardo Kastrup is a Dutch computer scientist and is one of the prominent defenders of metaphysical idealism, the notion that reality is essentially mental.
Bernardo has worked as a scientist in leading laboratories across the world, including CERN, searching for supersymmetry and finding the Higgs as an artifact, and the Phillips Research Laboratories, and he's a regular contributor to Scientific American. Susan Schneider is an academic and public philosopher, the William F. Dietrich Professor of Philosophy at Florida Atlantic University, and a recipient of the National Endowment for the Humanities Public Scholar Award.
She's also the co-director of the Machine Perception and Cognitive Robotics Lab at Florida Atlantic University, and her recent book, Artificial You and the Future of Mind, is now available. The question that we're going to use as a catalyst for this discussion is, should we accept that consciousness arises only in biological beings and that AI simply isn't made of the right ingredients? And we'll go in order of Don, Susan, and Bernardo, and welcome everyone. It's good to see you all. So, Don. So, I think the very question itself makes an assumption.
It assumes that space and time and physical objects are fundamental, that the fundamental reality of space and time and particles and quantum fields is not conscious, and that somehow consciousness is a latecomer in the universe, that somehow life emerged from
Non-organic starting points like particles and then eventually consciousness emerge. I think that the whole framework is wrong that our best science, quantum field theory with Einstein theory of gravity and evolution by natural selection tell us that space and time are not fundamental.
The elementary particles are not the fundamental reality and so the whole framework of the question, how does consciousness arise from physical systems, neurons or artificial intelligence is the wrong way to frame the question because space and time themselves are not fundamental.
to come up with a more fundamental framework to even address this question. What is beyond space and time? And physicists are finding new structures like decorative permutations beyond space and time. So they're going to change completely how we even think about this question. So my answer is that we're not thinking deeply enough about the question if we frame it that way. Susan. So I guess I would say no, first and foremost. I call my approach to machine consciousness the wait and see approach.
It's designed to be a middle ground between a view called biological naturalism, which says consciousness is intrinsically biological, and a view called techno-optimism associated with thinkers like Ray Kurzweil and other transhumanists. That view claims that conscious machines are inevitable as an outgrowth of sophisticated intelligence. I reject both of these views. First off, let me say
Consciousness in machines, the idea is conceptually and logically coherent. We see it in science fiction films all the time. There's nothing logically flawed about the idea. My problem, though, is with the idea that the techno optimists have that as AI gets smarter and smarter, consciousness will be an outgrowth of intelligent machines.
So I think that we don't know if building conscious AI is compatible with the laws of nature. We don't know if it's technically feasible, even if it is compatible with the laws of nature, to actually build conscious machines. And we don't know if AI companies would even want to do it. I mean, look at the mess with like LeMoy, right? He claimed that Google's Lambda system was conscious and it got him in a lot of trouble.
Google's Lambda system claimed to be conscious. These large language models, they're getting smarter and smarter. If they look to the public like they're conscious, I bet Google and the rest of big tech will try not to build conscious AI. Because if you're trying to build an AI service and people are convinced it's conscious, that looks like slavery.
Bernardo, I'm with Don that consciousness is not something that is generated. It's not a secondary or epiphenomenon of nature. It is primary, it is out there. But when we ask is a computer conscious, we mean something else we we mean to say that the computer has a private conscious in their life of its own. And I do think private conscious in their life arises, it emerges in nature, it's not something that was there from the get go.
So the question now is, okay, we know that we biological entities have private conscious in their lives. Does a computer or can a computer ever have private conscious in their life?
I would say we have absolutely no good reason to think that at all, certainly not idealism. Idealism would say everything is in consciousness, not that everything is conscious in the sense of having private conscious in their life of its own. The justification for the hypothesis that silicon computers will be conscious is based on some form of isomorphism or some form of similarity between the patterns of information flow in a biological brain and a silicon computer.
But if similarity is the crux of the argument, then we have to look at the concrete reality of a silicon computer and a brain and metabolizing that brain and realize that they are completely different.
and that's to find the similarity you have to take several steps of abstraction away from concrete reality until you find a level where there is an isomorphism but it's entirely arbitrary to say that at that level the similarity is defining and that all preceding levels don't matter although they show profound dissimilarities that those don't matter to me this is a form of belief i think
The chance that the silicon computer will be conscious is just about the same as the chance that my computer will pee on my desk if I run an accurate simulation of kidney function on it. Now, when it comes to urine, we understand the difference between a phenomenon and the simulation of the phenomenon. But when it comes to consciousness, for some reason, we don't. And I think the reason for that is that the human impulse for religion manifests itself in whatever way it can.
and transhumanism is an expression of the religious impulse and it abandons reason, abandons evidence and engages into a very suspicious form of thinking that is more about wishful belief than reason. All right, let's penetrate to the core of this dispute. What is consciousness and what best explains consciousness? We'll go in order of
bernardo susan and don and bernardo briefly if you don't mind explaining to the person or a thing being simulated in the machine does it not feel wetness in the wetness is not outside the machine but to what's being simulated if you could integrate that into your answer so when i use the word consciousness i mean phenomenal consciousness an entity is conscious if there is something it is like to be that entity
In other words, if the entity has experience at the most basic level. So for me, consciousness does not require higher level mental functions such as metacognition, self-reflection, introspection. No, if there is something is like to be something, then that something is conscious. Is there something it is like to be an entity in a computer simulation? No, I don't think there is anything it is like to be that entity. In the same sense that there is nothing it's like to be my computer mouse.
Actually, there isn't even a mouse because not everything we have a name for in language is an actual entity in the ontological structure of nature. So I think the burden of argument is on those who think computers will be conscious because I don't think we have any reason to entertain the hypothesis seriously at all. I can't refute the hypothesis, but then again, I can't refute the flying spaghetti monster.
The burden is not on me to refute the flying spaghetti monster. People have to provide an argument for us to take that hypothesis seriously. I think conscious silicon computers are the same thing and they are not better epistemically at all. It's just the same kind of, well, if I may be myself, the same kind of nonsensical hypothesis.
Please answer the question, but also if you want to respond to any of the comments Bernardo had just made. I agree with Bernardo that it's the felt quality of experience, what it feels like from the inside to be that entity. And the big question is whether a machine could be conscious. Maybe I'll just address the issue from my perspective as the former leader of a big project at NASA and the NASA chair.
because there my job was to articulate the relationship between consciousness and intelligence and talk about how we might locate either one in the universe and one thing I want to caution against actually and I know all of us here we have very detailed metaphysical theories of consciousness is assuming from our theories that we can read in
to the search for conscious machines or conscious beings on other planets, because look at the search for life in the context of astrobiology. When you look at the definition of life that NASA endorses, for example, self-replicating and capable of biological evolution, it is
tremendously open, capable of Darwinian evolution in particular, to the possibility that life could look very different on other planets. It doesn't need to be recognizably life to us. The big thing that I think we need to do here when we search for other forms of consciousness is keep an open mind and realize that it's going to look very different than the human case. And I think the interesting question here from a theoretical standpoint
is how different we'll allow it to be before we say it's not consciousness at all, which is one thing I want to throw out there. So, Bernardo and I both thought, well, consciousness is essentially the phenomenal feel. Well, to me that does seem like a necessary ingredient in consciousness. And if a machine lacks it, then to me it doesn't seem conscious. It has some other qualities and maybe based on those it will
be dangerous or it will be of moral significance. But I think that that is one thing where if we were to look for consciousness elsewhere, we might take as a fundamental ingredient. But I think it will be very, very different when we get to the case of synthetic machines. Briefly, you said it was necessary to felt experience of consciousness. Is it also sufficient? If that's too long of an answer, then we'll get to Don.
Yeah, but I think that, you know, that's where you get into issues involving how to define that felt quality. I mean, for example, there are some people who define consciousness in such a demanding way, like Daniel Dennett in his paper, Quining Qualia, you know, where he listed like seven fundamental ingredients, or other people try to build in a whole theory, like a higher order theory of consciousness that intellectualizes consciousness.
And that's where I think we need to be incredibly open and humble and not import our own theories of consciousness into the debate too early. Okay, Don. I would agree that consciousness is something that's experienced, something phenomenal. So, for example, the taste of coffee, the smell of a rose, the feeling of a headache.
These are all specific conscious experiences. And so that's what I mean when I talk about conscious experiences. And what's remarkable is that my brilliant friends and colleagues who are studying this problem using cognitive neuroscience and or physical systems like artificial intelligences, theories like integrated information theory or orchestrated collapse of quantum states of microtubules and so forth. What's remarkable is that
When you look at all these theories, there is not a single specific conscious experience that has ever been explained. What is the pattern of integrated information that must be the taste of vanilla and could not possibly be the smell of a rose? What is the orchestrated collapse, the precise orchestrated collapse of quantum states and microtubules that must be the taste of chocolate and could not be the taste of garlic?
If you ask, and these are my friends, I've asked them in public at conferences and so forth, give me a specific conscious experience that your theory can tell us. These are supposed to be theories of conscious experience, so they should explain conscious experiences. So what experience can you explain? What integrated information must be the taste of chocolate or whatever one you can do? And what's remarkable is there is no physicalist theory
Or integrated information theory or work straight class, whatever. There is no theory of that type that can explain even one specific conscious experience. We're batting zero. Nothing. There's nothing on the table. So that's why I said it's premature to talk about. So how could computers boot up consciousness? We have no idea about how anything could boot up consciousness right now from a physical system. Literally nothing is on the table.
so so until we can solve that we don't know how neurons could possibly boot up consciousness there's nothing on the table so that's why this is a you know
I said it was premature to be asking whether, you know, only biological systems can boot up consciousness and not AIs. We have no theory about how biological systems can boot up consciousness. There's literally nothing on the table that makes sense. And I predict that there that's a principle problem. This is not simply these are bright people. These are good friends of mine. They're brilliant. But when you start with space and time and physical objects inside space and time, like particles or neurons or whatever,
Good luck trying to boot up a theory of anything else from that. So that's my sort of my diagnosis of the problem. So we have to go after
I think I agree with Bernardo, you know, starting with an idea in which consciousness itself is fundamental. Bernardo is going after that philosophically. As a scientist, I'm trying to go after it mathematically, try to get a mathematically precise theory of the very kinds of things that Bernardo is talking about, and then show how space and time are booted up precisely from a theory of consciousness. So get a dynamics of consciousness outside of space-time and boot up space and time. And for example,
Scattering amplitudes of particles get all of that without a hand wave, get all of that derived precisely from a theory of consciousness. So in other words, we've tried to start with space and time and particles and biology and try to boot up a theory of consciousness from that.
We've not gotten a single success, not one specific conscious experience. But if we start with consciousness being fundamental and the mathematical model of that, then I think we can boot up space and time and particles and show where they come up and we can back 100 on that. So I think we should start with consciousness and show space and time is just a user interface that some consciousnesses use. But if we start the other direction, our best science tells us good luck, you've got the wrong foundation. What if someone says that
There is no the taste of garlic or the taste of vanilla. The way that broccoli tastes to me is different than the way it tastes to you and you bring associations with it atop. So that to think of it as an absolute is you have to think of it more in terms of a relation. So what if someone says that and then also number two is has do you feel like your theory has made progress toward picking out a certain element and saying that this is the taste of garlic and it could not be any other way?
Right, so those are both excellent points. So I would agree that there's probably, we can imagine there, my experience of the taste of coffee is very different from your experience of the taste of coffee. Some people don't like it and some people do, for example. So it seems that there's something very, very different that can happen there.
So a theory, a physicalist theory would need to explain first what's in common with all those people that we say that this is the taste of coffee, right? So we all, you know, say I'm tasting something like coffee. So what is, if we find some kind of neural correlates of that, great, we can find neural correlates of the taste of coffee.
And we can then see how those neural correlates vary from Bernardo to Don to Susan. And that's great. So we can see those differences. But if we're then going to have a scientific theory that says it's the neural activity or the orchestrated collapse of quantum states and microtubules or integrated information that's creating the taste of chocolate or the taste of coffee, then those theories
themselves are saying that they're going to do this. They're saying, we're going to start with the integrated information. We're going to give you the taste. Okay, well, no one told you you have to do that.
No one says you have to take on that problem, but they've said we're going to take on that problem. So then as a scientist, it's my, I'm very interested. Well, you can explain conscious experiences. So which conscious experience can you explain? And can you explain the variability, you know, Hoffman's taste of chocolate versus Susan's taste of chocolate? How, you know, what can you explain there? Well, right now, absolutely nothing. There's not a single experience that they can explain.
and again no one's put a gun to their head and said you have to start with physical systems and biology and boot up consciousness no one said you have to do that they're the ones that are volunteering to do it and i'm just saying they volunteered to do it and so so far they're batting zero um so maybe we need to look somewhere else well i forgot the second part of your question kirk is there any element of your is there a part of your theory that picks out as an element this is garlic and this is this cannot be not garlic
I start with conscious experiences, so every theory in science makes assumptions. Unfortunately, with all due respect, there is no theory of everything in science. Every theory makes certain assumptions, and given those assumptions, it'll explain other things, but no theory ever explains its own assumptions.
And so I start with conscious experiences and say, I'm going to start, if you grant me conscious experiences like the taste of coffee and so forth, end of dynamics, then I'll show you how I can boot up space and time. And the reason I think that's interesting is because the physicalist approaches are saying, grant me space and time and particles, grant me biology.
Oh, now grant me also these orchestrated collapses and these integrated information. And then you'll have to also grant me the experiences because I can't explain them. I have to stipulate the experiences too. I have to stipulate that there's this correlation between this experience and this integrated information or this orchestrated collapse. So the stipulating space and time, the integrated information and the experience. I say, let's just stipulate the experiences. I'll explain all the rest from that. So by Occam's razor, it's a much simpler approach and therefore to be preferred.
Great. Now on to phase two, AI is constantly improving. I think we can all agree on that. The possibility of artificial intelligence is around the corner. Now, sorry, artificial general intelligence, perhaps we don't agree. But let's imagine that's the case, that general artificial intelligence is around the corner. Would a machine with such an intelligence, a human level intelligence, thus have human level consciousness? We'll go Susan, then Bernardo, then Dawn, and Susan, feel free to comment and everyone else feel free to comment on what you've heard before.
Okay, yeah. To Donald, so we can't wait for a theory of consciousness to be discovered. And just like we can't wait for a theory of the nature of space-time emergence to be discovered, so while I, like you, see the two as connected, we have urgent issues right now arising all the time concerning the question of machine consciousness. So I think we have to actually separate out the issues
and figure out here and now ways of dealing with the fact that there are machines claiming to be conscious and there are intelligent humans like Blake Lemoine who is saying that the machines they're creating seem conscious. So that's just the first thing I want to mention. But you know the question itself is tricky because it brings up
Artificial general intelligence, Kurt. So you're asking about the possibility of AGI or artificial general intelligence and whether consciousness would fall out of that. And what I want to do though is just stress that I don't actually think there'll be anything like AGI ever created.
at least in the normal sense that people have of AGI. So AGI is usually considered to be human level, so something that's functionally isomorphic to a human. AI companies have already beat humans in various domains. They're not going to dumb down on their machines.
right? I think that we'll see very soon what I call savant systems. I have several questions for you just on savant systems written down and we can get to that at some point. It's such an interesting thing because I mean just as with a human savant there are strange deficits but they're beyond human capacities so too I think the first general intelligent systems that we'll see they'll be like that and they'll be made up of AI services and they'll be highly distributed
They'll be so different from us that we cannot assume that they're conscious. To get back to the end of your question, you're wondering, well, would these kind of entities, whether they be AGI's or Savant systems, be conscious? Well, I think we need to conceive of a situation where there's something like a global Google brain combined
processing of different AI services owned by Google that spatiotemporally distributed and that's the savant system. And why believe anything like that would have the felt quality of experience unless one dilutes down the notion of that to include panpsychism, the view that everything's conscious. I mean, I just can't see that sort of a system being conscious, but again, I have a wait and see approach. I'm in principle open-minded.
Do you mind explaining why you singled out Google? Is that just for instance, or did you actually mean that it would be a Google brain? I mean, both. I mean, I think Google is a soft monopoly. It has a lot of very promising AI services. It has very intelligent, large language models. But, you know, and they have the aspiration, unlike open AI, which is also building very intelligent, large language models. And I think we have to wonder
If there's already a savant system instantiated across the planet right now that we just don't know about. And indeed, how would we even identify one if it exists? Okay. Bernardo. I think intelligence and consciousness are completely different notions. One can be objectively measured. That is intelligence as it is defined in AI and computer engineering, computer science.
Intelligence just means a certain way of processing data that leads to effective solutions to problems or answers to questions without the user having to give a recipe for how to produce the answer or how to produce the solution. So it's basically an objective way of processing data that leads to solutions. Consciousness is what it is likeness. It is the felt qualities of experience. These are completely different things.
So even if we achieve general AI, and I think we will, I mean, I worked in AI myself, I think we will achieve general AI, that doesn't entail or imply or even suggest that that goes hand in hand with private conscious in their life at all. Why would it? It's a completely arbitrary association that we make.
People might say, well, Bernardo, but if an AI claims to be conscious and behaves just like a human and passes the Turing test, wouldn't we have to take it seriously? I would say, not at all, because it was constructed to look and behave like a human. I mean, we can construct a doll tomorrow that looks like a human.
Will that similarity be grounds for us to suspect that the doll might be human? Of course not. It was constructed to look like it in exactly the same way that the chat bot looks like a human talking doesn't provide any grounds whatsoever for us to think that it may have conscious private conscious in their life like a human because it was made to imitate a human. It was deliberately constructed for that purpose.
So that provides grounds for no philosophical speculation at all. So I think that the so-called ethical problems of dealing with AIs that claim to be conscious is equivalent to the ethical problem of judging the ethical imperatives of the flying spaghetti monster. It's the same thing. In other words, it's a non-existing problem that we make up
to maybe give expression to our repressed religious impulse or to entertain ourselves.
But it's just not a thing at all. It's just not there. AIs are tools, and we have to have safety measures when we are using powerful tools. That's no different for AI. A nuclear power plant is a tool to produce energy. It can also kill us. It can be dangerous. So we have to put safety mechanisms in place to protect us against the tool we created. AI is the same thing. It's a very powerful tool. It can go berserk.
So we have to put safety measurements in place, not to get hurt by the tool AIs that we use. That doesn't mean that we have an ethical problem in our hands at all, not any more than we would have an ethical problem about shutting down a nuclear reactor and killing the thing. We don't talk about that problem in these terms.
Well, I think that AI systems will become so brilliant, so complicated that we will not be able to even understand
them.
This clip was brought to you by Brilliant.org. If you're familiar with TOE, you're familiar with Brilliant. But for those who don't know, Brilliant is a place where you go to learn math, science and engineering through these bite sized interactive learning experiences. For example, and I keep saying this, I would like to do a podcast on information theory, particularly Chiara Marletto, which is David Deutsch's student has a theory of everything that she puts forward called constructor theory, which is heavily contingent on information theory. So I took their course on
random variable distributions and knowledge and uncertainty in order to learn a bit more about entropy now there's this formula for entropy essentially hammered into you as an undergraduate which seems to have fallen from the sky however when you take brilliance course it was the first time that i could see that it's an extremely clear and intuitive formula that is to say that
It would be unnatural to define it in any other manner. Visit brilliant.org slash toe that is t o e to get 20% off the annual subscription and I recommend that you don't stop before four lessons. I think you'll be greatly surprised that the ease at which you can now comprehend subjects you previously had a difficult time rocking at some point. I'll also go through the courses and give a recommendation in order.
The podcast is now concluded. Thank you for watching. If you haven't subscribed or clicked on that like button, now would be a great time to do so as each subscribe and like helps YouTube push this content to more people. Also, I recently found out that external links count plenty toward the algorithm, which means that when you share on Twitter, on Facebook, on Reddit, etc.
It shows YouTube that people are talking about this outside of YouTube, which in turn greatly aids the distribution on YouTube as well. If you'd like to support more conversations like this, then do consider visiting theories of everything dot org. Again, it's support from the sponsors and you that allow me to work on toe full time. You get early access to ad free audio episodes there as well. Every dollar helps far more than you may think. Either way, your viewership is generosity enough. Thank you.
▶ View Full JSON Data (Word-Level Timestamps)
{
"source": "transcribe.metaboat.io",
"workspace_id": "AXs1igz",
"job_seq": 8680,
"audio_duration_seconds": 1936.32,
"completed_at": "2025-12-01T01:11:34Z",
"segments": [
{
"end_time": 26.203,
"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 culture, they analyze finance, economics, business, international affairs across every region."
},
{
"end_time": 53.234,
"index": 1,
"start_time": 26.203,
"text": " I'm particularly liking their new insider feature was just launched this month it gives you gives me a front row access to the economist internal editorial debates where senior editors argue through the news with world leaders and policy makers and 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."
},
{
"end_time": 81.954,
"index": 2,
"start_time": 53.558,
"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? Jokes aside, Verizon has the most ways to save on phones and plants where everyone"
},
{
"end_time": 96.971,
"index": 3,
"start_time": 81.954,
"text": " The Institute"
},
{
"end_time": 112.5,
"index": 4,
"start_time": 96.971,
"text": " for Arts and Ideas produces numerous debates, lectures, and interviews on a broad range of topics including philosophy, science, politics, art. They asked me to host a debate of a sort between Donald Hoffman, Bernardo Kastrup, and Susan Schneider on whether machines are or can be conscious."
},
{
"end_time": 132.142,
"index": 5,
"start_time": 112.5,
"text": " This is part one, and the rest of the episode can be viewed on the IAI.tv website, as well as on their YouTube channel, both linked in the description. Also, I was asked to host an interview one-on-one between myself and Donald Hoffman, and that part one should be exclusively on the Toe channel shortly, as well as the rest on the IAI website slash YouTube channel. Enjoy. My name's Kurt Jaimungal, I'm the host of"
},
{
"end_time": 159.565,
"index": 6,
"start_time": 132.551,
"text": " The YouTube channel Theories of Everything, which analyzes tolls from a theoretical physics perspective. And it's my distinguished pleasure to host this talk with Donald Hoffman, Susan Schneider and Bernardo Castro. Donald Hoffman is an American cognitive psychologist working at the University of California, Irvine. He's making waves with his new theory, suggesting that instead of presenting reality as it really is, quote unquote, that our perception is tantamount to a desktop user interface, enabling us to use reality effectively."
},
{
"end_time": 176.067,
"index": 7,
"start_time": 159.565,
"text": " Bernardo Kastrup is a Dutch computer scientist and is one of the prominent defenders of metaphysical idealism, the notion that reality is essentially mental."
},
{
"end_time": 202.466,
"index": 8,
"start_time": 176.067,
"text": " Bernardo has worked as a scientist in leading laboratories across the world, including CERN, searching for supersymmetry and finding the Higgs as an artifact, and the Phillips Research Laboratories, and he's a regular contributor to Scientific American. Susan Schneider is an academic and public philosopher, the William F. Dietrich Professor of Philosophy at Florida Atlantic University, and a recipient of the National Endowment for the Humanities Public Scholar Award."
},
{
"end_time": 231.681,
"index": 9,
"start_time": 202.466,
"text": " She's also the co-director of the Machine Perception and Cognitive Robotics Lab at Florida Atlantic University, and her recent book, Artificial You and the Future of Mind, is now available. The question that we're going to use as a catalyst for this discussion is, should we accept that consciousness arises only in biological beings and that AI simply isn't made of the right ingredients? And we'll go in order of Don, Susan, and Bernardo, and welcome everyone. It's good to see you all. So, Don. So, I think the very question itself makes an assumption."
},
{
"end_time": 247.79,
"index": 10,
"start_time": 232.005,
"text": " It assumes that space and time and physical objects are fundamental, that the fundamental reality of space and time and particles and quantum fields is not conscious, and that somehow consciousness is a latecomer in the universe, that somehow life emerged from"
},
{
"end_time": 265.691,
"index": 11,
"start_time": 248.422,
"text": " Non-organic starting points like particles and then eventually consciousness emerge. I think that the whole framework is wrong that our best science, quantum field theory with Einstein theory of gravity and evolution by natural selection tell us that space and time are not fundamental."
},
{
"end_time": 282.142,
"index": 12,
"start_time": 265.691,
"text": " The elementary particles are not the fundamental reality and so the whole framework of the question, how does consciousness arise from physical systems, neurons or artificial intelligence is the wrong way to frame the question because space and time themselves are not fundamental."
},
{
"end_time": 311.391,
"index": 13,
"start_time": 282.551,
"text": " to come up with a more fundamental framework to even address this question. What is beyond space and time? And physicists are finding new structures like decorative permutations beyond space and time. So they're going to change completely how we even think about this question. So my answer is that we're not thinking deeply enough about the question if we frame it that way. Susan. So I guess I would say no, first and foremost. I call my approach to machine consciousness the wait and see approach."
},
{
"end_time": 338.234,
"index": 14,
"start_time": 311.698,
"text": " It's designed to be a middle ground between a view called biological naturalism, which says consciousness is intrinsically biological, and a view called techno-optimism associated with thinkers like Ray Kurzweil and other transhumanists. That view claims that conscious machines are inevitable as an outgrowth of sophisticated intelligence. I reject both of these views. First off, let me say"
},
{
"end_time": 366.51,
"index": 15,
"start_time": 338.712,
"text": " Consciousness in machines, the idea is conceptually and logically coherent. We see it in science fiction films all the time. There's nothing logically flawed about the idea. My problem, though, is with the idea that the techno optimists have that as AI gets smarter and smarter, consciousness will be an outgrowth of intelligent machines."
},
{
"end_time": 396.647,
"index": 16,
"start_time": 366.954,
"text": " So I think that we don't know if building conscious AI is compatible with the laws of nature. We don't know if it's technically feasible, even if it is compatible with the laws of nature, to actually build conscious machines. And we don't know if AI companies would even want to do it. I mean, look at the mess with like LeMoy, right? He claimed that Google's Lambda system was conscious and it got him in a lot of trouble."
},
{
"end_time": 421.783,
"index": 17,
"start_time": 396.937,
"text": " Google's Lambda system claimed to be conscious. These large language models, they're getting smarter and smarter. If they look to the public like they're conscious, I bet Google and the rest of big tech will try not to build conscious AI. Because if you're trying to build an AI service and people are convinced it's conscious, that looks like slavery."
},
{
"end_time": 451.34,
"index": 18,
"start_time": 422.534,
"text": " Bernardo, I'm with Don that consciousness is not something that is generated. It's not a secondary or epiphenomenon of nature. It is primary, it is out there. But when we ask is a computer conscious, we mean something else we we mean to say that the computer has a private conscious in their life of its own. And I do think private conscious in their life arises, it emerges in nature, it's not something that was there from the get go."
},
{
"end_time": 462.21,
"index": 19,
"start_time": 451.732,
"text": " So the question now is, okay, we know that we biological entities have private conscious in their lives. Does a computer or can a computer ever have private conscious in their life?"
},
{
"end_time": 488.166,
"index": 20,
"start_time": 462.568,
"text": " I would say we have absolutely no good reason to think that at all, certainly not idealism. Idealism would say everything is in consciousness, not that everything is conscious in the sense of having private conscious in their life of its own. The justification for the hypothesis that silicon computers will be conscious is based on some form of isomorphism or some form of similarity between the patterns of information flow in a biological brain and a silicon computer."
},
{
"end_time": 502.739,
"index": 21,
"start_time": 488.677,
"text": " But if similarity is the crux of the argument, then we have to look at the concrete reality of a silicon computer and a brain and metabolizing that brain and realize that they are completely different."
},
{
"end_time": 528.012,
"index": 22,
"start_time": 503.046,
"text": " and that's to find the similarity you have to take several steps of abstraction away from concrete reality until you find a level where there is an isomorphism but it's entirely arbitrary to say that at that level the similarity is defining and that all preceding levels don't matter although they show profound dissimilarities that those don't matter to me this is a form of belief i think"
},
{
"end_time": 556.664,
"index": 23,
"start_time": 528.49,
"text": " The chance that the silicon computer will be conscious is just about the same as the chance that my computer will pee on my desk if I run an accurate simulation of kidney function on it. Now, when it comes to urine, we understand the difference between a phenomenon and the simulation of the phenomenon. But when it comes to consciousness, for some reason, we don't. And I think the reason for that is that the human impulse for religion manifests itself in whatever way it can."
},
{
"end_time": 582.517,
"index": 24,
"start_time": 557.022,
"text": " and transhumanism is an expression of the religious impulse and it abandons reason, abandons evidence and engages into a very suspicious form of thinking that is more about wishful belief than reason. All right, let's penetrate to the core of this dispute. What is consciousness and what best explains consciousness? We'll go in order of"
},
{
"end_time": 606.63,
"index": 25,
"start_time": 582.517,
"text": " bernardo susan and don and bernardo briefly if you don't mind explaining to the person or a thing being simulated in the machine does it not feel wetness in the wetness is not outside the machine but to what's being simulated if you could integrate that into your answer so when i use the word consciousness i mean phenomenal consciousness an entity is conscious if there is something it is like to be that entity"
},
{
"end_time": 635.009,
"index": 26,
"start_time": 606.869,
"text": " In other words, if the entity has experience at the most basic level. So for me, consciousness does not require higher level mental functions such as metacognition, self-reflection, introspection. No, if there is something is like to be something, then that something is conscious. Is there something it is like to be an entity in a computer simulation? No, I don't think there is anything it is like to be that entity. In the same sense that there is nothing it's like to be my computer mouse."
},
{
"end_time": 659.497,
"index": 27,
"start_time": 635.009,
"text": " Actually, there isn't even a mouse because not everything we have a name for in language is an actual entity in the ontological structure of nature. So I think the burden of argument is on those who think computers will be conscious because I don't think we have any reason to entertain the hypothesis seriously at all. I can't refute the hypothesis, but then again, I can't refute the flying spaghetti monster."
},
{
"end_time": 683.336,
"index": 28,
"start_time": 659.906,
"text": " The burden is not on me to refute the flying spaghetti monster. People have to provide an argument for us to take that hypothesis seriously. I think conscious silicon computers are the same thing and they are not better epistemically at all. It's just the same kind of, well, if I may be myself, the same kind of nonsensical hypothesis."
},
{
"end_time": 711.783,
"index": 29,
"start_time": 684.155,
"text": " Please answer the question, but also if you want to respond to any of the comments Bernardo had just made. I agree with Bernardo that it's the felt quality of experience, what it feels like from the inside to be that entity. And the big question is whether a machine could be conscious. Maybe I'll just address the issue from my perspective as the former leader of a big project at NASA and the NASA chair."
},
{
"end_time": 740.657,
"index": 30,
"start_time": 712.654,
"text": " because there my job was to articulate the relationship between consciousness and intelligence and talk about how we might locate either one in the universe and one thing I want to caution against actually and I know all of us here we have very detailed metaphysical theories of consciousness is assuming from our theories that we can read in"
},
{
"end_time": 762.688,
"index": 31,
"start_time": 741.084,
"text": " to the search for conscious machines or conscious beings on other planets, because look at the search for life in the context of astrobiology. When you look at the definition of life that NASA endorses, for example, self-replicating and capable of biological evolution, it is"
},
{
"end_time": 791.698,
"index": 32,
"start_time": 763.148,
"text": " tremendously open, capable of Darwinian evolution in particular, to the possibility that life could look very different on other planets. It doesn't need to be recognizably life to us. The big thing that I think we need to do here when we search for other forms of consciousness is keep an open mind and realize that it's going to look very different than the human case. And I think the interesting question here from a theoretical standpoint"
},
{
"end_time": 820.998,
"index": 33,
"start_time": 792.329,
"text": " is how different we'll allow it to be before we say it's not consciousness at all, which is one thing I want to throw out there. So, Bernardo and I both thought, well, consciousness is essentially the phenomenal feel. Well, to me that does seem like a necessary ingredient in consciousness. And if a machine lacks it, then to me it doesn't seem conscious. It has some other qualities and maybe based on those it will"
},
{
"end_time": 851.118,
"index": 34,
"start_time": 821.152,
"text": " be dangerous or it will be of moral significance. But I think that that is one thing where if we were to look for consciousness elsewhere, we might take as a fundamental ingredient. But I think it will be very, very different when we get to the case of synthetic machines. Briefly, you said it was necessary to felt experience of consciousness. Is it also sufficient? If that's too long of an answer, then we'll get to Don."
},
{
"end_time": 880.23,
"index": 35,
"start_time": 851.783,
"text": " Yeah, but I think that, you know, that's where you get into issues involving how to define that felt quality. I mean, for example, there are some people who define consciousness in such a demanding way, like Daniel Dennett in his paper, Quining Qualia, you know, where he listed like seven fundamental ingredients, or other people try to build in a whole theory, like a higher order theory of consciousness that intellectualizes consciousness."
},
{
"end_time": 904.377,
"index": 36,
"start_time": 880.759,
"text": " And that's where I think we need to be incredibly open and humble and not import our own theories of consciousness into the debate too early. Okay, Don. I would agree that consciousness is something that's experienced, something phenomenal. So, for example, the taste of coffee, the smell of a rose, the feeling of a headache."
},
{
"end_time": 932.244,
"index": 37,
"start_time": 904.855,
"text": " These are all specific conscious experiences. And so that's what I mean when I talk about conscious experiences. And what's remarkable is that my brilliant friends and colleagues who are studying this problem using cognitive neuroscience and or physical systems like artificial intelligences, theories like integrated information theory or orchestrated collapse of quantum states of microtubules and so forth. What's remarkable is that"
},
{
"end_time": 959.138,
"index": 38,
"start_time": 932.654,
"text": " When you look at all these theories, there is not a single specific conscious experience that has ever been explained. What is the pattern of integrated information that must be the taste of vanilla and could not possibly be the smell of a rose? What is the orchestrated collapse, the precise orchestrated collapse of quantum states and microtubules that must be the taste of chocolate and could not be the taste of garlic?"
},
{
"end_time": 985.811,
"index": 39,
"start_time": 959.735,
"text": " If you ask, and these are my friends, I've asked them in public at conferences and so forth, give me a specific conscious experience that your theory can tell us. These are supposed to be theories of conscious experience, so they should explain conscious experiences. So what experience can you explain? What integrated information must be the taste of chocolate or whatever one you can do? And what's remarkable is there is no physicalist theory"
},
{
"end_time": 1010.35,
"index": 40,
"start_time": 986.459,
"text": " Or integrated information theory or work straight class, whatever. There is no theory of that type that can explain even one specific conscious experience. We're batting zero. Nothing. There's nothing on the table. So that's why I said it's premature to talk about. So how could computers boot up consciousness? We have no idea about how anything could boot up consciousness right now from a physical system. Literally nothing is on the table."
},
{
"end_time": 1018.763,
"index": 41,
"start_time": 1010.623,
"text": " so so until we can solve that we don't know how neurons could possibly boot up consciousness there's nothing on the table so that's why this is a you know"
},
{
"end_time": 1046.527,
"index": 42,
"start_time": 1019.053,
"text": " I said it was premature to be asking whether, you know, only biological systems can boot up consciousness and not AIs. We have no theory about how biological systems can boot up consciousness. There's literally nothing on the table that makes sense. And I predict that there that's a principle problem. This is not simply these are bright people. These are good friends of mine. They're brilliant. But when you start with space and time and physical objects inside space and time, like particles or neurons or whatever,"
},
{
"end_time": 1061.254,
"index": 43,
"start_time": 1046.527,
"text": " Good luck trying to boot up a theory of anything else from that. So that's my sort of my diagnosis of the problem. So we have to go after"
},
{
"end_time": 1089.701,
"index": 44,
"start_time": 1061.817,
"text": " I think I agree with Bernardo, you know, starting with an idea in which consciousness itself is fundamental. Bernardo is going after that philosophically. As a scientist, I'm trying to go after it mathematically, try to get a mathematically precise theory of the very kinds of things that Bernardo is talking about, and then show how space and time are booted up precisely from a theory of consciousness. So get a dynamics of consciousness outside of space-time and boot up space and time. And for example,"
},
{
"end_time": 1105.23,
"index": 45,
"start_time": 1090.213,
"text": " Scattering amplitudes of particles get all of that without a hand wave, get all of that derived precisely from a theory of consciousness. So in other words, we've tried to start with space and time and particles and biology and try to boot up a theory of consciousness from that."
},
{
"end_time": 1132.483,
"index": 46,
"start_time": 1105.23,
"text": " We've not gotten a single success, not one specific conscious experience. But if we start with consciousness being fundamental and the mathematical model of that, then I think we can boot up space and time and particles and show where they come up and we can back 100 on that. So I think we should start with consciousness and show space and time is just a user interface that some consciousnesses use. But if we start the other direction, our best science tells us good luck, you've got the wrong foundation. What if someone says that"
},
{
"end_time": 1157.739,
"index": 47,
"start_time": 1133.353,
"text": " There is no the taste of garlic or the taste of vanilla. The way that broccoli tastes to me is different than the way it tastes to you and you bring associations with it atop. So that to think of it as an absolute is you have to think of it more in terms of a relation. So what if someone says that and then also number two is has do you feel like your theory has made progress toward picking out a certain element and saying that this is the taste of garlic and it could not be any other way?"
},
{
"end_time": 1176.067,
"index": 48,
"start_time": 1158.951,
"text": " Right, so those are both excellent points. So I would agree that there's probably, we can imagine there, my experience of the taste of coffee is very different from your experience of the taste of coffee. Some people don't like it and some people do, for example. So it seems that there's something very, very different that can happen there."
},
{
"end_time": 1198.234,
"index": 49,
"start_time": 1176.442,
"text": " So a theory, a physicalist theory would need to explain first what's in common with all those people that we say that this is the taste of coffee, right? So we all, you know, say I'm tasting something like coffee. So what is, if we find some kind of neural correlates of that, great, we can find neural correlates of the taste of coffee."
},
{
"end_time": 1221.288,
"index": 50,
"start_time": 1198.234,
"text": " And we can then see how those neural correlates vary from Bernardo to Don to Susan. And that's great. So we can see those differences. But if we're then going to have a scientific theory that says it's the neural activity or the orchestrated collapse of quantum states and microtubules or integrated information that's creating the taste of chocolate or the taste of coffee, then those theories"
},
{
"end_time": 1230.469,
"index": 51,
"start_time": 1221.698,
"text": " themselves are saying that they're going to do this. They're saying, we're going to start with the integrated information. We're going to give you the taste. Okay, well, no one told you you have to do that."
},
{
"end_time": 1253.37,
"index": 52,
"start_time": 1230.862,
"text": " No one says you have to take on that problem, but they've said we're going to take on that problem. So then as a scientist, it's my, I'm very interested. Well, you can explain conscious experiences. So which conscious experience can you explain? And can you explain the variability, you know, Hoffman's taste of chocolate versus Susan's taste of chocolate? How, you know, what can you explain there? Well, right now, absolutely nothing. There's not a single experience that they can explain."
},
{
"end_time": 1277.125,
"index": 53,
"start_time": 1253.37,
"text": " and again no one's put a gun to their head and said you have to start with physical systems and biology and boot up consciousness no one said you have to do that they're the ones that are volunteering to do it and i'm just saying they volunteered to do it and so so far they're batting zero um so maybe we need to look somewhere else well i forgot the second part of your question kirk is there any element of your is there a part of your theory that picks out as an element this is garlic and this is this cannot be not garlic"
},
{
"end_time": 1298.882,
"index": 54,
"start_time": 1277.637,
"text": " I start with conscious experiences, so every theory in science makes assumptions. Unfortunately, with all due respect, there is no theory of everything in science. Every theory makes certain assumptions, and given those assumptions, it'll explain other things, but no theory ever explains its own assumptions."
},
{
"end_time": 1316.169,
"index": 55,
"start_time": 1299.206,
"text": " And so I start with conscious experiences and say, I'm going to start, if you grant me conscious experiences like the taste of coffee and so forth, end of dynamics, then I'll show you how I can boot up space and time. And the reason I think that's interesting is because the physicalist approaches are saying, grant me space and time and particles, grant me biology."
},
{
"end_time": 1342.892,
"index": 56,
"start_time": 1316.169,
"text": " Oh, now grant me also these orchestrated collapses and these integrated information. And then you'll have to also grant me the experiences because I can't explain them. I have to stipulate the experiences too. I have to stipulate that there's this correlation between this experience and this integrated information or this orchestrated collapse. So the stipulating space and time, the integrated information and the experience. I say, let's just stipulate the experiences. I'll explain all the rest from that. So by Occam's razor, it's a much simpler approach and therefore to be preferred."
},
{
"end_time": 1372.756,
"index": 57,
"start_time": 1342.892,
"text": " Great. Now on to phase two, AI is constantly improving. I think we can all agree on that. The possibility of artificial intelligence is around the corner. Now, sorry, artificial general intelligence, perhaps we don't agree. But let's imagine that's the case, that general artificial intelligence is around the corner. Would a machine with such an intelligence, a human level intelligence, thus have human level consciousness? We'll go Susan, then Bernardo, then Dawn, and Susan, feel free to comment and everyone else feel free to comment on what you've heard before."
},
{
"end_time": 1403.507,
"index": 58,
"start_time": 1373.592,
"text": " Okay, yeah. To Donald, so we can't wait for a theory of consciousness to be discovered. And just like we can't wait for a theory of the nature of space-time emergence to be discovered, so while I, like you, see the two as connected, we have urgent issues right now arising all the time concerning the question of machine consciousness. So I think we have to actually separate out the issues"
},
{
"end_time": 1432.875,
"index": 59,
"start_time": 1403.916,
"text": " and figure out here and now ways of dealing with the fact that there are machines claiming to be conscious and there are intelligent humans like Blake Lemoine who is saying that the machines they're creating seem conscious. So that's just the first thing I want to mention. But you know the question itself is tricky because it brings up"
},
{
"end_time": 1452.995,
"index": 60,
"start_time": 1433.2,
"text": " Artificial general intelligence, Kurt. So you're asking about the possibility of AGI or artificial general intelligence and whether consciousness would fall out of that. And what I want to do though is just stress that I don't actually think there'll be anything like AGI ever created."
},
{
"end_time": 1471.937,
"index": 61,
"start_time": 1453.66,
"text": " at least in the normal sense that people have of AGI. So AGI is usually considered to be human level, so something that's functionally isomorphic to a human. AI companies have already beat humans in various domains. They're not going to dumb down on their machines."
},
{
"end_time": 1498.968,
"index": 62,
"start_time": 1471.937,
"text": " right? I think that we'll see very soon what I call savant systems. I have several questions for you just on savant systems written down and we can get to that at some point. It's such an interesting thing because I mean just as with a human savant there are strange deficits but they're beyond human capacities so too I think the first general intelligent systems that we'll see they'll be like that and they'll be made up of AI services and they'll be highly distributed"
},
{
"end_time": 1522.773,
"index": 63,
"start_time": 1498.968,
"text": " They'll be so different from us that we cannot assume that they're conscious. To get back to the end of your question, you're wondering, well, would these kind of entities, whether they be AGI's or Savant systems, be conscious? Well, I think we need to conceive of a situation where there's something like a global Google brain combined"
},
{
"end_time": 1550.503,
"index": 64,
"start_time": 1523.592,
"text": " processing of different AI services owned by Google that spatiotemporally distributed and that's the savant system. And why believe anything like that would have the felt quality of experience unless one dilutes down the notion of that to include panpsychism, the view that everything's conscious. I mean, I just can't see that sort of a system being conscious, but again, I have a wait and see approach. I'm in principle open-minded."
},
{
"end_time": 1579.906,
"index": 65,
"start_time": 1550.896,
"text": " Do you mind explaining why you singled out Google? Is that just for instance, or did you actually mean that it would be a Google brain? I mean, both. I mean, I think Google is a soft monopoly. It has a lot of very promising AI services. It has very intelligent, large language models. But, you know, and they have the aspiration, unlike open AI, which is also building very intelligent, large language models. And I think we have to wonder"
},
{
"end_time": 1607.193,
"index": 66,
"start_time": 1580.469,
"text": " If there's already a savant system instantiated across the planet right now that we just don't know about. And indeed, how would we even identify one if it exists? Okay. Bernardo. I think intelligence and consciousness are completely different notions. One can be objectively measured. That is intelligence as it is defined in AI and computer engineering, computer science."
},
{
"end_time": 1635.469,
"index": 67,
"start_time": 1607.637,
"text": " Intelligence just means a certain way of processing data that leads to effective solutions to problems or answers to questions without the user having to give a recipe for how to produce the answer or how to produce the solution. So it's basically an objective way of processing data that leads to solutions. Consciousness is what it is likeness. It is the felt qualities of experience. These are completely different things."
},
{
"end_time": 1657.722,
"index": 68,
"start_time": 1635.879,
"text": " So even if we achieve general AI, and I think we will, I mean, I worked in AI myself, I think we will achieve general AI, that doesn't entail or imply or even suggest that that goes hand in hand with private conscious in their life at all. Why would it? It's a completely arbitrary association that we make."
},
{
"end_time": 1680.947,
"index": 69,
"start_time": 1658.148,
"text": " People might say, well, Bernardo, but if an AI claims to be conscious and behaves just like a human and passes the Turing test, wouldn't we have to take it seriously? I would say, not at all, because it was constructed to look and behave like a human. I mean, we can construct a doll tomorrow that looks like a human."
},
{
"end_time": 1708.404,
"index": 70,
"start_time": 1681.237,
"text": " Will that similarity be grounds for us to suspect that the doll might be human? Of course not. It was constructed to look like it in exactly the same way that the chat bot looks like a human talking doesn't provide any grounds whatsoever for us to think that it may have conscious private conscious in their life like a human because it was made to imitate a human. It was deliberately constructed for that purpose."
},
{
"end_time": 1732.176,
"index": 71,
"start_time": 1708.404,
"text": " So that provides grounds for no philosophical speculation at all. So I think that the so-called ethical problems of dealing with AIs that claim to be conscious is equivalent to the ethical problem of judging the ethical imperatives of the flying spaghetti monster. It's the same thing. In other words, it's a non-existing problem that we make up"
},
{
"end_time": 1738.831,
"index": 72,
"start_time": 1732.841,
"text": " to maybe give expression to our repressed religious impulse or to entertain ourselves."
},
{
"end_time": 1766.391,
"index": 73,
"start_time": 1739.138,
"text": " But it's just not a thing at all. It's just not there. AIs are tools, and we have to have safety measures when we are using powerful tools. That's no different for AI. A nuclear power plant is a tool to produce energy. It can also kill us. It can be dangerous. So we have to put safety mechanisms in place to protect us against the tool we created. AI is the same thing. It's a very powerful tool. It can go berserk."
},
{
"end_time": 1786.8,
"index": 74,
"start_time": 1766.391,
"text": " So we have to put safety measurements in place, not to get hurt by the tool AIs that we use. That doesn't mean that we have an ethical problem in our hands at all, not any more than we would have an ethical problem about shutting down a nuclear reactor and killing the thing. We don't talk about that problem in these terms."
},
{
"end_time": 1816.305,
"index": 75,
"start_time": 1786.8,
"text": " Well, I think that AI systems will become so brilliant, so complicated that we will not be able to even understand"
},
{
"end_time": 1816.886,
"index": 76,
"start_time": 1816.647,
"text": " them."
},
{
"end_time": 1846.783,
"index": 77,
"start_time": 1817.978,
"text": " This clip was brought to you by Brilliant.org. If you're familiar with TOE, you're familiar with Brilliant. But for those who don't know, Brilliant is a place where you go to learn math, science and engineering through these bite sized interactive learning experiences. For example, and I keep saying this, I would like to do a podcast on information theory, particularly Chiara Marletto, which is David Deutsch's student has a theory of everything that she puts forward called constructor theory, which is heavily contingent on information theory. So I took their course on"
},
{
"end_time": 1866.817,
"index": 78,
"start_time": 1846.783,
"text": " random variable distributions and knowledge and uncertainty in order to learn a bit more about entropy now there's this formula for entropy essentially hammered into you as an undergraduate which seems to have fallen from the sky however when you take brilliance course it was the first time that i could see that it's an extremely clear and intuitive formula that is to say that"
},
{
"end_time": 1888.541,
"index": 79,
"start_time": 1866.817,
"text": " It would be unnatural to define it in any other manner. Visit brilliant.org slash toe that is t o e to get 20% off the annual subscription and I recommend that you don't stop before four lessons. I think you'll be greatly surprised that the ease at which you can now comprehend subjects you previously had a difficult time rocking at some point. I'll also go through the courses and give a recommendation in order."
},
{
"end_time": 1909.36,
"index": 80,
"start_time": 1888.541,
"text": " The podcast is now concluded. Thank you for watching. If you haven't subscribed or clicked on that like button, now would be a great time to do so as each subscribe and like helps YouTube push this content to more people. Also, I recently found out that external links count plenty toward the algorithm, which means that when you share on Twitter, on Facebook, on Reddit, etc."
},
{
"end_time": 1936.323,
"index": 81,
"start_time": 1909.36,
"text": " It shows YouTube that people are talking about this outside of YouTube, which in turn greatly aids the distribution on YouTube as well. If you'd like to support more conversations like this, then do consider visiting theories of everything dot org. Again, it's support from the sponsors and you that allow me to work on toe full time. You get early access to ad free audio episodes there as well. Every dollar helps far more than you may think. Either way, your viewership is generosity enough. Thank you."
}
]
}No transcript available.