Conscious black holes and the qualia of infinity

Transcript

00:00:03Hey everyone, welcome back to Waking Kosmos, Adrian here. Today I'm joined by Mike Johnson from the Qualia Research Institute, where they're looking at really one of the most enduring mysteries in philosophy of mind, which is how do we say anything about why it is that experiences like the red of roses, or the smell of roses, or the taste

00:00:30taste of cinnamon, have their very unique ways of appearing in consciousness, what we call their

00:00:36qualia. Why do we experience red as having this very specific qualia? And of course, going beyond

00:00:44maybe conjuring up a kind of evolutionary explanation about the signal that the color

00:00:49red might have had at different points in our evolutionary history about what's out there in

00:00:55the world. But to go beyond that and talk about these very distinctive topologies of experience,

00:01:01which seem to defy language. And that's really the mystery that Mike is focusing on. And he's

00:01:09going to talk about a theory that he and his colleagues at the Qualia Research Institute

00:01:14are exploring, which could potentially open a significant door into this space. It's called

00:01:22the symmetry theory of valence.

00:01:25And in this theory, the valence of an experience, which is essentially how positive we interpret an

00:01:31experience to be, is a function of its internal symmetry. So Mike and his colleagues are identifying

00:01:38symmetry as an important access point into this space of qualia, of experience. In the second half

00:01:47of today's episode, we discuss some of Mike's broader metaphysical views, including how any

00:01:54sufficiently critical theory of the qualia of experience can be used to identify the symmetry.

00:01:56can be directed at anything, including cosmological phenomena like black holes and even the Big Bang.

00:02:04So I hope you enjoyed today's episode as much as I did. Remember, if you're enjoying these

00:02:09open-minded conversations about consciousness and reality, you can support Waking Cosmos on Patreon

00:02:16and help me turn this into a full-time project. And for those of you who are already supporting

00:02:21my work, thank you for your kindness and generosity. Alright, I started out today by

00:02:28asking Mike, why should we care about the mystery of consciousness?

00:02:35So the question is, why is consciousness such an important topic? I think it's the important topic.

00:02:43I think that all questions of morality and value ultimately resolve to questions about

00:02:51consciousness.

00:02:51And it's kind of embarrassing that we don't actually have any really good theories in this space.

00:03:02You and I met at the Tucson Science of Consciousness Conference this year,

00:03:07and I think it would probably be fair to say that this is the most prestigious

00:03:11consciousness conference that there is at the moment. So what did you think about the

00:03:15conference this year? Did you feel like the discussion about consciousness is

00:03:20moving in a positive direction?

00:03:22Yeah, I think so. I think of this in terms of where on the process of turning consciousness

00:03:28research from alchemy into chemistry are we. And I would say that we're still fairly close

00:03:36to alchemy, but there are some tantalizing hints of maybe there are some things that

00:03:42we can formalize and empirically test. The most intriguing themes at the conference for me

00:03:51all formalism. So some of Giulio Tononi's grad students were there and talking about IIT.

00:03:59And we can talk about what is IIT. It's a form of panpsychism, right?

00:04:04Yes, yes. And it's a way of formalizing what it would feel like to be a system. And that system

00:04:11can be anything. You can apply IIT to a person, a dog, a car, an alien, and you get a well-formed

00:04:21So it's a very powerful approach to understanding consciousness.

00:04:26And it's based on the amount of integrated information that a system is capable of having.

00:04:33Yep, that's right. Brains have a lot of this sort of integrated information. There's a lot

00:04:37of integration between the neurons. One way to think about integration is how much each part

00:04:42of the system is causally affected by the other parts. And brains have a lot of that.

00:04:48Whereas, you know, something like a chair has

00:04:51very little integration between the neurons. And so it's a very powerful approach to understanding

00:04:51very, very, very little. So IIT does imply panpsychism,

00:04:57which would imply that everything

00:05:00is at least a little bit conscious, but there would be horrendous orders of magnitude of difference between like a chair and a dog, for instance.

00:05:09Right. Presumably we wouldn't identify a chair as having consciousness because it is itself not integrated. It's just kind of a collection of matter that we have a functional use for.

00:05:20Right, right. There's really no functional integration in a chair.

00:05:23So it's quite interesting that the general discussion about consciousness at the moment and at the highest levels of academia, we're taking this view that consciousness is something real and probably an important part of the way that the world is organized.

00:05:39Yeah. I think it's slowly catching on and I think it's catching on because literally it's the only way to make progress.

00:05:48That if there's something to be explained,

00:05:53the correct approach to it is let's formalize this. Let's formalize this thing that we're trying to explain.

00:06:00Whereas I think in years past people were more in favor of just trying to explain it away.

00:06:07And in that case then formalism would be the wrong approach.

00:06:11Right. So you've described yourself as a qualia formalist.

00:06:15Yes. That's right.

00:06:16So can you explain what that means and maybe include within that an explanation of what qualia is?

00:06:23Yeah. So qualia are the components of subjective experience.

00:06:29So I'm having the qualia of a little bit of pressure on my skin, a little bit of warmth from the air.

00:06:39I've got some green qualia and some white qualia and so on.

00:06:44And to say that I'm a qualia formalist means we should be able to precisely,

00:06:51in a very,

00:06:53rigorous way, we should be able to talk about what sort of sensations I'm having mathematically.

00:07:00To put this in, you know, big language,

00:07:04we should be able to derive a mathematical object that's isomorphic to my phenomenology.

00:07:10And that just means that it represents it perfectly.

00:07:13Right. And so by reducing phenomenology to a mathematical structure,

00:07:20that's not a materialist move.

00:07:23Not necessarily.

00:07:24You're still a realist about consciousness,

00:07:26but you're saying that this is the suchness of certain kinds of experience like redness and blueness.

00:07:33It should be a part of an intelligible map of qualia that we can understand

00:07:38and we should be able to expect that mathematics would be a place in which we could explore qualia.

00:07:47Right. And that's as far as I need to go,

00:07:51in terms of what assumption I'm making.

00:07:53I think there are a lot of assumptions that are required from my work.

00:07:55But yes, let's talk about this mathematically.

00:07:59Personally, though, I'm kind of partial to something on the order of dual aspect monism,

00:08:05where the physical world and the phenomenological world are sort of two shadows cast by underlying reality.

00:08:14Right. And that's really what I'm sympathetic with as well.

00:08:17So the classic view of that might be Russellian monism,

00:08:21or maybe the idea that,

00:08:23the world as it is,

00:08:25is constituted from maybe exteriors and interiors,

00:08:30and that they imply each other.

00:08:32Right, right.

00:08:34And so if we assume that and take a look at physics,

00:08:38there are a lot of very, very wonderful, crisp, predictive formalisms in physics.

00:08:46And so to be a qualia formalist,

00:08:48it's just saying,

00:08:49well, maybe we should try to apply that same sort of thinking,

00:08:53scientific outlook to qualia.

00:08:56So I'm fascinated by your symmetry theory of valence,

00:09:00which you've been developing for several years.

00:09:03So it's a theory about qualia and an approach to understanding why qualia

00:09:07or specific experiences have distinct kinds of textures that they do,

00:09:13which is currently a huge mystery.

00:09:16And you framed this as kind of the problem that we are immediately confronted with after we solve,

00:09:23the hard problem of consciousness.

00:09:25Right.

00:09:26There's this question of if we could figure out consciousness,

00:09:31if we got a good quantitative theory of consciousness,

00:09:34then what?

00:09:35So at the Qualia Research Institute,

00:09:38we focus a lot on the interpretation problem.

00:09:42If I could figure out some mathematical object that represented what it feels like to be you,

00:09:49then how can we use that for useful things?

00:09:53One way to frame this is,

00:09:57what's the first line in the Rosetta

00:10:00Stone of consciousness?

00:10:01I love that.

00:10:01Yeah. Another way to frame it is what is the C. elegans of qualia? What is the simplest

00:10:07place to start? It's a big problem in that no one really knows. And if we could figure

00:10:16out just where to start, that would be halfway to solving it. And at QRI, we think we know

00:10:24where to start. And that's with this idea of emotional valence, or how pleasant or unpleasant

00:10:32an experience feels. So you have these two domains. You have the mathematical domain.

00:10:39So if you have a quantitative theory of consciousness, it tells you something mathematical about

00:10:46the phenomenology of a system. And then you have the phenomenological domain.

00:10:54And that's talking about what qualia does the system have. And there's this big issue

00:11:00of how do you map between these? How do you point to something in the math domain and

00:11:06then connect it to something in the domain of phenomenology?

00:11:11The approach that I've taken is to link symmetry in the mathematical domain with emotional

00:11:21valence.

00:11:24And that's the phenomenological domain.

00:11:26Valence being how positive we interpret an experience to be, right?

00:11:31Yes, that's correct.

00:11:32And so you've identified valence and then symmetry as a particularly interesting natural

00:11:39kind. Is it a natural kind?

00:11:41Yeah. So Plato and Aristotle talked about natural kinds and trying to carve reality

00:11:48at the joints. And this is, of course, it is the problem in consciousness research.

00:11:54Where do you even start? What words fit the contours of reality? And we do think that

00:11:59emotional valence is well defined across all conscious experiences and the simplest

00:12:05place to start.

00:12:06Right. I think typically people tend to lump valence in with the whole spectrum of emotions

00:12:14that humans are capable of having and presumably many other species. But you're saying actually

00:12:20valence is a fundamental action.

00:12:24in the space of qualia, right?

00:12:28Right. So emotions are complex. So Lisa Barrett has this interesting paper, Are Emotions

00:12:36Natural Kinds? And her answer is no, that they're complex bundles of phenomenology,

00:12:43behavior and dynamics. And it's very messy and complicated. But what I'm saying is that

00:12:52emotional valence is a building block of emotions. It's this atomic thing that you can build

00:12:58over complicated things out of.

00:13:00And you've suggested that this in some sense corresponds to the mathematical symmetry that

00:13:07is occurring in the brain.

00:13:08That's our hypothesis. Essentially, harmony in the brain feels good. Full stop. That is

00:13:17literally what feels good. Now, that's not to say that listening to high harmony stimuli

00:13:24or high constant stimuli always feels good.

00:13:28Right.

00:13:29But that's due to boredom causing dissonance in our brain so that we move on.

00:13:36So boredom is the brain introducing dissonance? That's interesting. Could you say more about that?

00:13:44So when the brain is too good at predicting its inputs, it gets bored. And it tells you basically,

00:13:54hey, you know, this is not worth paying attention to as much as before. And then it's like, oh, I'm

00:13:58bored.

00:13:58So there's this explore versus exploit dichotomy. And boredom is kind of advanced technology

00:14:06that the brain uses in order to shift from exploit back to explore.

00:14:11So we've already got software programs that can create very beautiful music just by implementing

00:14:18the underlying mathematics that music employs. So do you think that by using this theory

00:14:24about valence, we'd be able to create music that was

00:14:28algorithmically optimized to be like the most positive music that we could listen to?

00:14:32And presumably, if we could do that, what we'd get would be pretty extraordinary. So

00:14:38yeah, I'm curious what genre of music you think that would be.

00:14:42Yeah, that's a great question. I would say, I mean, first of all, just a disclaimer

00:14:49that everyone's brain is a little bit different. So there might not be one piece of music that

00:14:54works equally well.

00:14:55Right.

00:14:58So I think that's a really important question, which is, if this thinking about valence

00:15:00is correct, what accounts for our different tastes in things like music?

00:15:05Right. Essentially, I want to say the biggest component of this is what kind of

00:15:12mind music is already going on inside your head. And you want to harmonize with that.

00:15:19You may prefer songs with a rhythm or a tempo that's kind of calibrated for that.

00:15:28Hey, if you like heavy metal music, maybe something really interesting and maybe a little

00:15:32dissonant is going on in your head already. And I mean, it's also a function of what mood are you in

00:15:40right now. And some moods, you might like heavy metal and some moods, you might like something

00:15:45more upbeat or... But very importantly, the same thing is going on in the head of somebody who's

00:15:54listening to heavy metal and enjoying it versus listening to techno or dance or club or blues and

00:16:02enjoying it. And that's harmony in the brain. So you've suggested that it's kind of the

00:16:09mind's goal to have more harmony. What does being in a state of high symmetry or high harmony

00:16:18do for the brain? How does it orient a mind towards

00:16:24expanding or developing consciousness? Does it have a kind of role or is it just a fact about

00:16:29the way that mind exists in a kind of metaphysical sense? Yeah. So I would say it is a background

00:16:37kind of physical law, like gravity is universal. The symmetry theory of valence would be universal

00:16:44through in humans, in dogs and dinosaurs and conscious aliens. I would also say that just from

00:16:54an evolutionary perspective, that harmony in the brain seems to be a success condition.

00:17:03It's something that maybe we're not programmed to experience a lot,

00:17:08but we are programmed very, very heavily to seek it out because it's linked to what causes

00:17:17temporary harmony in the brain, such as eating a lot of calories or having good things happen to

00:17:24you or accomplishing your goals. The brain is geared toward that being a success condition.

00:17:33I guess what I'm still slightly unclear about is if we have this success condition and let's say

00:17:39that we do achieve it from time to time, does that then serve to be an optimal state to then

00:17:44make further inferences? And is there a kind of intrinsic value about a lot of symmetry and

00:17:51harmony, which projects in some teleScript orScript orScript orScript orScript orScript orScript orScript

00:17:54of mind in a positive direction. Right. Well, this gets into the question of can we get to

00:18:01ought from is. Ethics. Right. Ethics. I think this is the bridge that if we say, okay,

00:18:11all value boils down to consciousness and value of consciousness is based on the emotional valence

00:18:20of the experience. And the emotional valence is wholly dependent on the symmetry of the

00:18:28mathematical representation. Then it doesn't take much of a jump from there to say, okay,

00:18:35this actually gives us a prescriptive understanding of what we're here to do.

00:18:42There's a lot of evolutionary pressures and so on. But if we're here to optimize the universe's

00:18:50qualia,

00:18:51then we should focus heavily on symmetry and valence. Yeah, I would generally agree with that.

00:18:57And I don't think we can easily get away from the fact that all value and meaning are realized in

00:19:03conscious experience. I know I always say it, but maybe that's a necessary part of the universe. But

00:19:09consciousness is the way that value is realized in the world. And as we're starting to recognize

00:19:15this and begin to integrate it collectively, to me, this is leading us to orientate

00:19:20our choices and our ethics toward this vast, unrealized possibility space of possible

00:19:27conscious experiences, which are as real in many ways as tomorrow morning. So I think we can see

00:19:34ourselves in the present moment as making choices that prune away branches from what we are capable

00:19:41of realizing in that space. Right. I like the way that you put that, that they're as real as tomorrow

00:19:48morning. Thanks.

00:19:50I think it's really important to just know what's out there, to have a map of consciousness or the possible

00:19:58state space of conscious

00:20:00experience because we really don't right now, we don't even know how good the future could be because we don't have this map.

00:20:08Right, and in a sense I think we can say that by definition we don't know what the value is that can be realized in this possibility space.

00:20:17But the one thing I think that we can know in the sense that saying that it's real is that we know that there is incredible value that could be realized in consciousness.

00:20:26And we also know at the same time, that sophisticated

00:20:29As sophisticated and complex and developed as we are, we in all likelihood only occupy a very tiny corner in terms of the space of mind that could exist.

00:20:40And so there's an enormous space to be realized.

00:20:45I think that the possibility space of consciousness is in a very important sense our birthright.

00:20:56And we just have to figure out how to do that.

00:20:59And figure out what kind of thing it is and how to navigate it.

00:21:03Is it the case that we could be equally concerned about the space of consciousness or optimizing the qualia of the universe?

00:21:13Even if, for example, materialism was the case, if we went in the direction of Dennett,

00:21:18we could still say well you know for all practical purposes consciousness exists and the things that we care about are in consciousness.

00:21:25Even if the way that we ultimately formalize that is as a

00:21:29some kind of illusion. And I'm not a big fan of that view, but it strikes me that even if that's

00:21:35the case, we'd still have many of the same intuitions about ethics, because it would still

00:21:40reduce to concerns about conscious experience. Well, when I say consciousness, and when someone

00:21:50like Dennett says consciousness, we seem to be saying two different things. In a sense, I think

00:21:56you're right, that we can both at least use the same words in a conversation and seem to make

00:22:04sense. But on a more deeper level, they chart out very, very, very different universes.

00:22:12If consciousness is real, is formalizable, then moral value is objective. Whereas if consciousness

00:22:22is this kind of high-level abstraction,

00:22:26or leaky reification, then there's nothing objective about it. And we're kind of left to

00:22:34our own devices as to how to navigate this. And it sort of precludes any systematizing

00:22:40of this field. I think that if you want to point to a specific part of the brain and say,

00:22:49that's where the magic happens, then we do need a consistent story about how the magic happens,

00:22:56and I think any consistent story about how the magic happens will also be panpsychist.

00:23:06So let me ask you this. Do you think that panpsychism gives us a kind of a more intelligible

00:23:12account of reality? Does panpsychism make sense of, in addition to why it is that we have conscious

00:23:19experience, but other mysteries or unknowns that we have in science and about the universe?

00:23:26Right. So this question of what exactly should we be aiming for is important. And some people

00:23:34think that, well, a good theory of consciousness should match my intuitions. And I think that

00:23:40that isn't always a reasonable assumption. I think reality could have some surprises

00:23:46in store for us there. But I think that how I would put it is that a good theory of consciousness

00:23:54will be consistent.

00:23:56Such that, for example, gravity is a consistent theory. And ideally, it should be able to make

00:24:04novel predictions, bold novel predictions that we can try to falsify, and hopefully predictions

00:24:13which we can't get to if we don't assume consciousness is real.

00:24:18Do you have any intuitions about what kinds of predictions that panpsychism could make,

00:24:25which perhaps could even be...

00:24:26verifiable?

00:24:27Right. So Giulio Tononi's IIT does make some predictions in terms of which systems

00:24:36are conscious and which brain states are conscious. It can predict which people can wake up from

00:24:43comas based on the kind of background level of consciousness in their brain. It does this

00:24:49by sort of ringing the brain like a bell with TMS and then...

00:24:54Transcranial magnetic stimulation.

00:24:56That's right. That's right. And then seeing how long it rings,

00:25:00how long it takes to get back to whatever state it was in.

00:25:05And they're doing much, much better than anyone else predicting who will wake up from a coma, which is really impressive.

00:25:14So at QRI, we're trying to add another element to this.

00:25:20So our work is compatible with IIT, but we also add emotional valence to the mix.

00:25:28So for instance, we think that we have a method by which you can put somebody in a neuroimaging system like an fMRI,

00:25:40and we can objectively tell you how happy or sad they are.

00:25:45Right. And yeah, this strikes me as something that could be amazingly useful and could fundamentally change how we interact with each other.

00:25:55I mean, if it was possible for us to have a kind of representation of what other people's mental states are in, and our own as well, of course,

00:26:05there would certainly be times when this would be something that we would want to share.

00:26:09And obviously, there would be many times when we would probably want to keep this private.

00:26:13But assuming that your theory is correct, and the degree of harmony in our brains does correspond to how positive we feel,

00:26:22imagine technologies which will eventually change.

00:26:25But if they actually come out of that recognition, they could be abused.

00:26:29But I can imagine that fundamentally changing how we interact with each other.

00:26:35And in another respect, I think if we could glimpse at more of the experience of other people,

00:26:42and maybe the pain that other people are in and other beings,

00:26:46I imagine that we would be much more quick to help them.

00:26:50And who knows what this technology would look like.

00:26:53But I think if we had a focus on the focus of the focus.

00:26:55sense of someone's emotional valence in real time it's kind of an extraordinary impact on the way

00:27:01that we interact with each other right like a 21st century mood ring exactly yeah and if there is an

00:27:09objective fact of the matter about how we feel and if if we can kind of find the right lens by which

00:27:15to view this and i think um atisoi's work and qri's work constitutes that lens um then that

00:27:25change human interaction and to just take a step back if it's possible to do what iit claims to be

00:27:33able to do which is simply to point at things in the world and be able to say yes this system is

00:27:39conscious and to a certain degree this is of course not something that we know how to do at

00:27:44the moment you know i just assume that you're conscious because you seem conscious and you know

00:27:49maybe i have a 10 percent less confidence that an animal has consciousness and

00:27:55things get a little bit more squirrely when i look at a plant or a cell but my intuition

00:28:01philosophically is that there's a lot of consciousness going on in the world around us

00:28:05but if we could have a formal way of knowing that and develop what tononi calls the qualia scope

00:28:12and look around the world and see where consciousness is being instantiated in the

00:28:18world i mean that would be huge maybe even transform society and you know specifically

00:28:25in coma patients like you were saying we would be able to know if there was someone

00:28:30still inside there is there a conscious mind inside that brain or are these anesthetics truly

00:28:38effective in the way that we think they are or are they just simply blocking memory for example

00:28:42and so formalism about consciousness give us a lot of answers to questions that we really care about

00:28:47but this additional step a formalism about valence that could be the first line of the

00:28:55rosetta stone of consciousness so yeah it's definitely fascinating and even if it turns

00:29:00out to be incorrect that would be independently interesting as well because we'd be able to ask

00:29:06why and if not a mathematical symmetry then what sort of formalism should we be pursuing

00:29:12yeah well it's it's building a new science and i always go back to the alchemy to chemistry example

00:29:19it's a good one yeah yeah and it's this idea of you know alchemy was

00:29:25like clearly something cool was happening there but people were very very confused about exactly

00:29:31how to think about reality and then over time and with a lot of effort we you know figured out this

00:29:38periodic table and you know electron shells and it started to become systematic and i think we're

00:29:45we're just at the beginnings of that in consciousness and it's incredibly exciting i i

00:29:51wake up in the morning and i see it i'm like

00:29:55what kind of thing can we systematize today or or this year

00:30:00Nerd.

00:30:01Absolutely.

00:30:04Yeah, it's amazing to me that we find ourselves at this time in history.

00:30:09It's quite exciting, really, to live in a time where we haven't actually got answers to these really fundamental questions.

00:30:16And in many ways, we're not even really asking the question in the right way yet in terms of ethics.

00:30:23And we don't really know how to delineate from the kind of parochial ethics that we've got as a result of our evolutionary history as primates and particular kinds of cultural idiosyncrasies that we have.

00:30:36And those ethical norms, for example, which may be universalizable and they may apply to societies of minds wherever they're instantiated in the universe.

00:30:47And yet we don't have any way of thinking about that clearly yet.

00:30:52Yeah.

00:30:53I mean, just in the last 10 years with IIT coming onto the scene, we're beginning to be able to ask the right questions.

00:31:03And it is exciting to be living in an era with real questions that haven't been answered yet.

00:31:10And I would agree that asking the correct questions is at least half the battle.

00:31:17I feel like a lot of people haven't kept up on the literature and realized how

00:31:23quickly things are moving.

00:31:25And there's a lot hanging in the balance.

00:31:27There's a lot at stake.

00:31:28There's a non-trivial chance that the whole enterprise of life on this planet could come to an end.

00:31:34There's at least a certain degree to which it's true to say that solving these problems, this navigation problem of ethics, for example, of what it is we're trying to achieve together here in the universe is going to be important for the survival of the entire enterprise itself.

00:31:50And so the extent to which we can do what you're doing is going to be very important.

00:31:53you're advocating which is to orient ourselves to how do we optimize qualia in the universe

00:31:58we're moving beyond this very species specific parochial view about human preferences to this

00:32:07space of consciousness which is in many ways what we really are would you agree with that

00:32:13is consciousness what we really are yes um absolutely and and i would consider myself a

00:32:20qualiaist in in terms of uh like what is real what should be optimized i think a lot is hanging

00:32:28in the balance here so um nick bostrom talks about in super intelligence that there are trade-offs

00:32:35that we can make um in terms of increasing efficiency and um sort of with the with

00:32:44transhumanism and advanced technology and ai then more things are on the table there are

00:32:50more things that are on the table and more things are on the table and more things are on the table

00:32:50that we can trade for greater efficiency and he makes this uh he has this wonderful quip that

00:32:57we should make sure not to trade off the wrong things uh and by which he means consciousness

00:33:06because then we might end up building a disneyland with no children i love that it's very similar to

00:33:13what um uval noah harari says at the end of his book homo deus he basically says you know at some

00:33:19point if we keep moving down this

00:33:20hyper technological path which is presumably inevitable we can't change course really even

00:33:26if we wanted to and at some point we're going to be confronted with the choice of what do we value

00:33:31most information processing or consciousness and of course it's a rhetorical question because the

00:33:38answer is without consciousness we don't really have the we don't have a perspective by which to

00:33:43gauge the value of the answer and so oh yeah i wanted to ask you uh this it relates to um

00:33:50what what you're talking about here so because it seems more or less um more and more that we're

00:33:55outsourcing our cognition into technology uh for example you know with our phones nobody really

00:34:03thinks that our phones are conscious right but you know we're all glad that we don't have to

00:34:07remember phone numbers anymore so i guess what i'm interested in getting your opinion about is

00:34:12what do you think is kind of the optimal line when it comes to outsourcing our cognition could we go

00:34:18too far with this and

00:34:20we certainly want consciousness there at some level right i do think that there are things that

00:34:28we shouldn't outsource just if we outsource more and more to unconscious processes then pretty soon

00:34:37we won't be useful consciousness won't be useful that which actually has value won't be useful

00:34:43so this is really uh another one of your areas which you've spent quite a lot of time thinking

00:34:48about and you mentioned to me

00:34:50that one of the initial reasons which brought you out here to the bay area was

00:34:55ai safety and so why is ai safety something that you've given

00:35:00a lot of attention to.

00:35:02So I tend to believe that AI is dangerous, both in terms of a GI, artificial general

00:35:08intelligence, and just narrow AI.

00:35:11You can do a lot of damage with computers, basically.

00:35:15And the smarter the computer is, the more damage you could do.

00:35:19So yeah, I moved out to the Bay to focus on AI safety.

00:35:23I then sort of realized that consciousness was even less studied than AI safety.

00:35:32And I felt like it was my duty to help change that.

00:35:36An open question here is, what do we do after we win?

00:35:40If we figure out how to build very intelligent AI, and we basically have a blank check to

00:35:48live in any sort of future that we want, what sort of future should we want to live in?

00:35:53And I think this goes straight back to consciousness.

00:35:56To understand this, we have to understand what has value, what the possibility space

00:36:03is.

00:36:04Yeah, I very much agree.

00:36:06And, you know, I think there is this real question that, you know, what are we going

00:36:09to do when we have computers which are more intelligent than we are?

00:36:12And, you know, what kind of potential risks and gains, like, could be waiting for us down

00:36:18the AI path?

00:36:19And I certainly think that we should take very seriously the idea that some kind of

00:36:23self-sufficiency, you know, is a very important part of the AI, and that we should be able to

00:36:25If it gets loose, so to speak, it could be enormously destructive.

00:36:30And of course, you know, there's Nick Bostrom's example in which a rampant artificial intelligence

00:36:35turns the entire cosmic horizon into paperclips.

00:36:40And people love that example.

00:36:42But the much more immediate concern, presumably, is the risk that a superintelligence would

00:36:48pose to us if it wasn't aligned with either our preferences or intrinsic interests of

00:36:56consciousness to develop and rarify. Right, so I think human survival is

00:37:04important. I also think that sort of the grand arc of what kind of qualia are

00:37:12instantiated by the universe is the important thing. And I think that we're

00:37:19in a very critical time point right now. I actually think that if we don't solve

00:37:27AI, then there's a grand opportunity cost there. And it's possible that that

00:37:36future will be permanently less good than if we do solve consciousness before

00:37:43we develop AI.

00:37:45I suppose if super AI is sort of inevitable, I guess the challenge that we

00:37:52have is to direct it to becoming something that we wouldn't mind replacing

00:37:55us.

00:37:56Right. And you know, whether or not it has consciousness seems to be central to

00:37:59that.

00:38:00I think that's very important. So there's this question of a, how do we prevent

00:38:06AI from killing us and using our atoms for something else? And that's a problem.

00:38:12But also, if we figured out that to some extent, how do we teach AI how to be kind

00:38:19to us, how to be benevolent? And there are all these approaches from what's called

00:38:26indirect normativity. And it's just kind of looking at how humans treat each other,

00:38:31and then extrapolating that for teaching an AI. And I think that's a very imperfect

00:38:38approach for a lot of reasons.

00:38:40It's very specific to our particular time and species and so on.

00:38:45Right. There's this big divide in ethics on this question. And it's this question of,

00:38:54is goodness preference satisfactory?

00:38:56Right.

00:38:57Like, if you want some ice cream, I give you some ice cream, and then you're happy.

00:39:05So should we define things in terms of satisfying people's preferences? Or should we define

00:39:12things in terms of making sure they have good qualia? I'm of the second camp. I think that

00:39:19people often don't know what they want, or what they want won't actually make them happy.

00:39:24And we can have second order desires.

00:39:26About our preferences, like that maybe we know that ice cream isn't that good for us.

00:39:30Sure. But even that doesn't directly connect to phenomenology. And I think phenomenology is what

00:39:36literally actually matters.

00:39:38And with respect to this question in ethics, that what we should care about is either

00:39:43satisfying our preferences, or as you mentioned, creating good qualia. Is creating good qualia,

00:39:50is that synonymous with developing the possibility space of consciousness that we're,

00:39:56capable of sampling the value of?

00:39:59I

00:40:00would generally agree.

00:40:02So I'm working on a post about utopia and what makes a good society.

00:40:09And my conclusion is that there are two components to this and people usually only focus on one

00:40:16or the other.

00:40:17But we really do need to bring in both.

00:40:21And the first component is just having a functional civilization, a working society in which

00:40:28people can engage in positive some dynamics.

00:40:31And so being able to keep the lights on, you can't have nice things without civilization.

00:40:37And then the follow up to that is that just living in a healthy society does not necessarily

00:40:47or causally make people happy.

00:40:49Obviously, I think it would very much contribute to this.

00:40:52But I think in that sense, if we have a social

00:40:58surplus, if our society is running well, and then we can decide what to put our surplus

00:41:04in, I think it makes a lot of sense to optimize the universe's qualia.

00:41:11I think that's what literally has intrinsic value.

00:41:16Good qualia.

00:41:17And good qualia is more than just a lot of valence.

00:41:21It's how do we...

00:41:23I suppose we have to leave the interpretation of what is good qualia.

00:41:28And what is not good qualia to consciousness itself.

00:41:30Right.

00:41:31I think that it's kind of early days to say anything very firmly.

00:41:37But I think that a good starting point here would be experiences that have very high emotional

00:41:47valence.

00:41:47If the thing that ultimately matters in the universe is positive valence, is our

00:41:54goal in some sense to tile the entire universe?

00:41:58I think there's this separation of philosophically what the correct answer is, and practically

00:42:11what we should do today.

00:42:12And I think these diverge on this question.

00:42:17I do think that experiences with positive valence are literally what seems to have value.

00:42:24And yes, we should optimize for that.

00:42:28But only in the longest of terms.

00:42:31I mean, it would be easy for humankind to just decide, oh, let's optimize for positive

00:42:37valence.

00:42:37We kind of burn ourselves out in this flash of light, and then everything's dark.

00:42:43I'm more of a fan of taking the long view.

00:42:47Taking the long, long, long, long view.

00:42:50The cosmological perspective.

00:42:52The cosmological perspective, exactly.

00:42:54And seeing the value in...

00:42:57Yeah.

00:42:58The value in civilization, in things that keep things working, that keep the lights

00:43:03on.

00:43:03I like to say that without civilization, we can't have any nice things.

00:43:10Let me ask you this, because it's something which I perhaps don't quite fully understand

00:43:15about this view of positive valence.

00:43:19So there's two things about it which I perhaps don't probably understand.

00:43:23The first thing is that we can certainly feel very good.

00:43:27We can feel very good.

00:43:28And it be to our detriment.

00:43:29So we can become addicted to opium or something like that, and we can wither away, but we're

00:43:37in total bliss.

00:43:38And on the other hand, there's a lot of challenging experiences, let's say, which aren't associated

00:43:45with a lot of positive valence and may feel pretty rough.

00:43:48And yet, at a later time, we metabolize those experiences having enormous amounts of value.

00:43:56I think so.

00:43:57Yeah.

00:43:58Suffering can be redemptive.

00:44:00It can put us on a better path.

00:44:02It can allow us to grow in ways that we would not have grown without it.

00:44:08Which isn't to say that we should fall in love with suffering.

00:44:12I think it serves a very important functional role in the human psyche.

00:44:17And also just, I mean, if you're at a local maxima, then to get anywhere better, you have

00:44:23to go down before you get up.

00:44:26At the same time, you're not alone.

00:44:29things about human experience do hurt a lot. So if we can replace those things with things that

00:44:40fulfill the same function, but are more pleasant, less unpleasant, less painful, then that would be

00:44:47kind of a viable path forward. I think our brains are pretty good at using suffering for

00:44:55kind of redemptive purposes often, but

00:45:00you also have edge cases where things break. You know, you have people who suffer from depression for 30 years. You have a grieving process which lasts a decade or

00:45:13So, I think that, um, not all suffering is redemptive.

00:45:19And if we focus on those forms of non-redemptive suffering, uh, that's the low hand for it

00:45:26here.

00:45:28But I want to take a bit of a left turn here because I really want to get into some of

00:45:32your cosmological, uh, speculations about consciousness.

00:45:35And you've agreed to talk about these even though, you know, we're definitely in the,

00:45:39well into the realm of, of speculation.

00:45:41But that is kind of what this podcast is about.

00:45:44And so this is, it's definitely a place for open-mindedness.

00:45:47So, uh, and I find particularly, uh, some of what you've, uh, speculated about in the

00:45:52appendix of your paper, uh, absolutely fascinating.

00:45:55And so, and it's the idea that, you know, if qualia is in some sense all around us, uh,

00:46:01for example, if some essentially idealist or panpsychist theory of consciousness is

00:46:05true, then there's presumably a lot more experience that's going on in the universe than just

00:46:11us or, or even life on this planet.

00:46:14And of course, we're not just talking about aliens.

00:46:17Uh, and I think for example, you speculated that the big bang, uh, probably felt pretty

00:46:22awesome.

00:46:23Can you talk about that?

00:46:25Yeah.

00:46:26So, um, yeah, just to, to touch on the speculative nature of this, um, I kind of stuffed all

00:46:33this stuff in the, uh, the last appendix of my paper.

00:46:36Um, I definitely wanted to include it, but, uh, it wasn't sort of core to my argument.

00:46:41So for, uh, for listeners who maybe are a little leery of, you know, biting the whole

00:46:48bullet here, you know, you don't, you don't have to, but at the same time, if we have

00:46:55a, um, sort of an objective theory of consciousness that properly crisply formal theory of consciousness,

00:47:04then we should be able to apply it to anything literally.

00:47:08And so the same theory of consciousness, which gives us a crisp formal result, if we pointed

00:47:14at a human brain should give just as good of a result, just as precise a result if we

00:47:22pointed at, you know, a dog or a dog.

00:47:25A dinosaur, a chair, a tree.

00:47:28Um, but it should also give good result if we pointed at cosmological phenomenon, right?

00:47:37Such as the big bang or such as stars or black holes.

00:47:41So this would be the qualia scope that Tononi talks about, right?

00:47:46So we take our qualia scope and we start directing it out into the cosmos.

00:47:51Where do you see the consciousness lighting?

00:47:55Yeah.

00:47:55Coming up in this field of view.

00:47:57Right.

00:47:58So assuming that, uh, Tononi is right in that integrated information is kind of the magic

00:48:05that generates consciousness.

00:48:07Then in my appendix, I list a number of possibilities.

00:48:12One is maybe big things have a lot of qualia.

00:48:16So maybe stars, but especially black holes, there's just so much mass in a black hole.

00:48:26if they have hair, so to speak, that could sort of encode integrated information or allow integrated information,

00:48:35then presumably it does feel like something to be a black hole.

00:48:40So there's something about black holes which could have integrated information.

00:48:45Certainly they have information, but integrated information in a black hole.

00:48:50Right. I think based on what we know about black holes, that's a live possibility.

00:48:57And presumably it would be stretched across the surface.

00:49:00Right. And potentially it could be a lot more conscious than any one person.

00:49:06Wow.

00:49:07And to an extreme order of magnitude, perhaps.

00:49:11Yeah.

00:49:12Many orders of magnitude.

00:49:13I really want to get back to why it is that you speculated that the Big Bang could have felt really good.

00:49:20And if so, was that the kind of omega point and that we are sometimes just living in the aftermath of that and the main event has already passed?

00:49:30Right. So I do mention that in my paper as, you know, very speculatively.

00:49:36I like how you assign probabilities to some of these things as well.

00:49:41Status, highly speculative.

00:49:43Sure, sure. You have to. But I felt like it was wrong not to go there.

00:49:49So there's this.

00:49:50This idea of eternal inflation, that the Big Bang is sort of ongoing.

00:49:56It's this infinite process that doesn't get depleted.

00:50:00This special sort of matter that's decaying into normal matter but doesn't itself deplete.

00:50:07And this tends to be described in the literature as a very high symmetry process. And so if there is consciousness associated with this process, then I'm speculating, and this is complete speculation, but it should probably feel pretty good.

00:50:28That would be the understatement of the century.

00:50:30Right. Do you have a view about where the universe might be heading with respect to the development of qualia?

00:50:38Could it be that we could be heading towards another kind of singularity of infinite complexity of consciousness and mind at some point if, say for example,

00:50:49and I feel like I speculate about this a lot on this podcast, but could it be that life eventually saturates the entire universe

00:50:58in some way and optimizes all of the matter in the universe and maybe it creates Max Tegmark's perceptronium?

00:51:05How do you see the idea that everything is in some sense inevitably heading toward a kind of conscious omega point?

00:51:14Yeah. I outline a few possibilities in the paper of if humans aren't the center of qualia in the universe.

00:51:23That if there's more qualia than just human qualia out there, then where's the higher quality?

00:51:28How do you see the universe operating?

00:51:28Large scale phenomenon like black holes, small scale phenomenon like virtual particles popping in and out of existence all the time,

00:51:37quantum fuzz is one way to describe that, and then the past, the Big Bang, if it created any qualia, it would be creating a lot of qualia

00:51:47and the future, that it could be that the universe is just starting to wake up and there's a lot of qualia in its future.

00:51:58I tend to think of a U-shaped curve.

00:52:02I think it's plausible that the Big Bang created a lot of qualia.

00:52:06And then we are sort of a qualia god chatter.

00:52:11I love it when you say that.

00:52:12Yeah. Re-coalescing 14 billion years after the main event.

00:52:17And then I think this process could accelerate and the U could go back up.

00:52:24And in some sense, that might be another Big Bang.

00:52:28In a way.

00:52:28So this would be a, what was the name of that cosmology?

00:52:32The Big Bounce.

00:52:34But it's a big bounce between omega points of consciousness.

00:52:38Yeah. I mean, in general, I just want to say that this is very speculative.

00:52:43But I'm always willing to bite the bullet and kind of plug things in and see where we get.

00:52:51And I think that you do get some interesting implications for cosmology.

00:52:58And long term future thinking, depending on what theory of consciousness is true.

00:53:05Right. So in principle, we should be able to extrapolate any formalized theory of consciousness all the way out into cosmology.

00:53:13Right. Right. Exactly.

00:53:15Let's return to the symmetry theory of valence.

00:53:18And perhaps you could re-summarize it briefly.

00:53:21What do you see is next for this theory?

00:53:23And how do you see it being tested?

00:53:26Right.

00:53:28So just to summarize briefly, the symmetry theory of valence suggests that harmony in the brain literally is pleasure.

00:53:36That's the short version.

00:53:37The somewhat more formal version is that...

00:53:41It's a very short version.

00:53:42Yeah. Yeah.

00:53:44Symmetry in the mathematical representation of phenomenology literally is pleasure.

00:53:51And that's a metaphysical truth in a way.

00:53:55Right.

00:53:56That would be true.

00:53:58But if you look at the theory of the universe across all conscious beings, whether they're humans or aliens halfway across the galaxy.

00:54:04Like a law of the universe, like gravity.

00:54:06And this is a theory which is testable.

00:54:09Right.

00:54:10And that's what kind of makes it special.

00:54:15Because it's testable.

00:54:17It's falsifiable.

00:54:18So we're saying that harmony in the brain feels good.

00:54:23Just...

00:54:24that is what feels good. So if we can put somebody in an fMRI and measure this, the amount of

00:54:31harmony in their brain, we should be able to tell how happy they are even better than they can.

00:54:39Right. I was going to mention this earlier when we first talked about this. You know, people's

00:54:43own self-reporting of their experience isn't always the best indicator of what they're

00:54:47experiencing. So, for example, you know, we could be wrong about how stressed we are.

00:54:54It's a surprise when people have a nervous breakdown because they're not fully aware of

00:54:57just how much stress they're in.

00:55:00And yet to say that we may not necessarily be the best judge of our own conscious experience

00:55:06seems very counterintuitive.

00:55:08Right, right.

00:55:09I mean, in the short term, we're going to have to use people's self-reports as the

00:55:14gold standard just to validate the theory.

00:55:17But in the long run, I think we can improve on that.

00:55:20Yeah.

00:55:21Yeah. And again, this is a substrate independent view about the nature of consciousness. So it's

00:55:28an approach to looking at qualia, which doesn't necessarily force us to pick a team when it comes

00:55:35to our particular theory of consciousness. And you choose to plug in Tononi's integrated

00:55:41information theory into this because it makes the right kinds of prediction. But in principle,

00:55:47we could have a different view about what consciousness is. And we could still,

00:55:51in principle, plug this into the symmetry theory of valence and see how it works.

00:55:58Right. We're pretty careful to be theory agnostic in terms of we don't have to pick a horse

00:56:07in order to talk about the interpretation problem. We assume that trying to derive a

00:56:15formalism for consciousness is a very important thing. And we don't have to pick a horse in

00:56:18order to talk about the interpretation problem. We assume that trying to derive a formalism is

00:56:18the correct approach. But we don't assume anything about which specific formalism might be true.

00:56:26So integrated information theory is probably the leading theory right now. But it could be true,

00:56:34or it could be false. And we don't necessarily have a need to come down on the other side there.

00:56:41Right. So to bring in another potential theory of consciousness,

00:56:48maybe we believe that we need a quantum theory of consciousness and that it's important that our

00:56:53theory of consciousness pays attention to the quantum level of reality and all of the weird and

00:57:01interesting stuff that happens down there. And if so, Hamroff and Penrose might be getting closer

00:57:08with their orchestrated objective reduction theory of consciousness. So if this turned out to be

00:57:15the right formalism or closer to the right form of consciousness, then we might be able to do a lot of

00:57:18to the right formalism. Could we plug this into your symmetry theory of valence? Would that work?

00:57:25Yes. So orchestrated object reduction, ORC-OR, is definitely a player in this space.

00:57:35Maybe we ought to summarize it very briefly.

00:57:38Yeah. Essentially, with ORC-OR, consciousness is the collapse of quantum superpositions in the brain,

00:57:48especially having to do with the microtubules of neurons.

00:57:53It's also a panpsychism as well.

00:57:55Yes. I think that ORC-OR is in an interesting position in that it has a lot of the pieces

00:58:05of the puzzle to formula or attempt to formula phenomenology. But it doesn't make the final leap.

00:58:14It gets 90% of the way there. But it's still a formula. And so I'm going to spend a little bit of time

00:58:18it just doesn't go all the way. So IIT has this really clever way to draw the boundary

00:58:26of a conscious system, and then it takes the information inside this boundary, and then it

00:58:33rearranges it to represent phenomenology in order to precisely represent what it feels like to be

00:58:39that system. But ORCOR, it draws the boundary around a conscious system, but it doesn't do

00:58:48with the information in the boundary. It doesn't output a formal representation of phenomenology

00:58:55like IIT does. And I don't think that there's anything intrinsically holding ORCOR back

00:59:03on this. Penrose and Hamaroff just haven't made that final step.

00:59:08Right. So in principle, ORCOR could be developed to make this more clear boundary.

00:59:14I think so. And I would certainly encourage them,

00:59:18to do that. Cool. Let's finish this off somehow.

00:59:23Psychedelics and QRI? Yeah. Let's talk about psychedelics and QRI.

00:59:29What is to you so interesting about psychedelics and what it can teach us about qualia?

00:59:35Yeah. So a lot of times when a system is functioning normally, a lot of its internal

00:59:43workings are hidden. And psychedelics are interesting in that they

00:59:48perturb systems in fairly predictable ways, repeatable ways. And then suddenly things

00:59:56aren't working normally and things that you just took as

01:00:00as givens as eternal facts about consciousness or the mind are suddenly different. So they're a very useful tool by which to try to chart out the state space of consciousness.

01:00:14Right. One of the things that strikes me about the psychedelic experience is how it's an example of just how wired we are to experience symmetry, especially in the visual aspect.

01:00:28So if you're going to have a very potent psychedelic experience, the one thing that you rely on is that you're going to be experiencing a lot of symmetry and a lot of tessellating patterns and a lot of apparent meaning in those patterns.

01:00:41Perhaps that suggests that, you know, maybe symmetry is representative of the structure of the mind in some fundamental way.

01:00:48Yeah. My colleague Andrés Gómez-Emilsson and also someone by the name of Steve Lahar

01:00:56Yeah.

01:00:58have looked into this symmetry and psychedelic connection as well. I think there's something there. And I think there's sort of the brain naturally sort of tries to

01:01:10symmetrify things, symmetrify inputs, in a sense, trying to compress them with specific symmetry patterns.

01:01:20Right.

01:01:21And I think with some psychedelics, you can kind of viscerally experience that.

01:01:25Right.

01:01:27If the symmetry theory of valence is the appropriate way to go about understanding how it is that we feel the ways that we feel, and especially under psychedelics,

01:01:40potentially an application for symmetry theory of valence is that we could take someone that was having, for example, a particularly challenging psychedelic experience,

01:01:52and we could introduce or try and

01:01:55evoke more symmetry in their experience.

01:01:59Yeah, there's a lot there. I think it's absolutely true that if the symmetry theory of valence is true, it should have a lot of immediate practical applications.

01:02:13And I think, you know, rescuing people from bad trips would be a good example.

01:02:19I could imagine a good NowThis video about, like, new theory of qualia.

01:02:23Yeah.

01:02:24Saving people from the precipice of infinity at Burning Man this year.

01:02:29Right. Right.

01:02:30Of course, you'd have to have, like, Andreas there with some sand plates on top of some speakers and showing people. That would really…

01:02:37Yeah. We sort of already do this. I think that, like, set and setting are seen as very important.

01:02:44And the set and setting that are often encouraged are very sort of harmonious, not distanced.

01:02:53So this idea is already in play.

01:02:58But one thing that it could maybe offer over and above what's out there right now is more of a multimodal approach

01:03:08and trying to harmonize the different patterns of symmetry that you're exposed to.

01:03:15If you have some sort of a visual pattern and then you have some sort of a different audio pattern,

01:03:23it's important not just to have them be symmetrical or harmonious, but to have them symmetrically or harmonious with each other.

01:03:33Right.

01:03:33And I think that keeping the symmetry theory of valence in mind as you undergo a journey could itself have a lot of use.

01:03:45Well, Mike, thank you so much for joining me today and having this conversation.

01:03:49I find the way that you think fascinating.

01:03:53I'm very intrigued by your research with symmetry and valence.

01:03:58Where can people go to find out more about your work and ideas and the Qualia Research Institute?

01:04:05Yeah. So just let me say thank you as well for having me.

01:04:09Real pleasure.

01:04:10Yeah. Thanks.

01:04:11You can check out our work at qualiaresearchinstitute.org.

01:04:16We have a blog on there and kind of our core research.

01:04:21My colleague, Andres Gomez-Emelson, has a blog at qualiacomputing.com, which is also excellent.

01:04:28I'll put links to all those in the descriptions as well.

01:04:32Well, Mike, thanks again for joining me on the Waking Cosmos podcast.

01:04:36And I really hope we can do this again sometime soon.

01:04:40Thanks.

01:04:40Me too.

01:04:43Okay. I hope you enjoyed today's episode with Mike Johnson from the Qualia Research Institute.

01:04:48If you're listening on YouTube, let me know what you thought about today's topics in the comments.

01:04:53And if you enjoyed this episode, hitting like helps us out.

01:04:56And if you're not already, now is also a great time to subscribe and

01:05:00join us on this adventure. Remember if you want to support Waking Cosmos and these open-minded conversations about consciousness and reality, you can help me turn this into a full-time project by supporting me on my Patreon page and the link to that is in the description. Thank you.

01:05:17Alright, I will see you next time for more episodes of Waking Cosmos, exploring the nature of consciousness, reality, and life's place in the universe.

01:05:28That's about it from me. I hope you have a beautiful day.