Episode Transcript
[00:00:03] Speaker A: Exploring ways to do cognitive science and cognitive neuroscience that do not rely exclusively on computations as a modeling tool or as a sort of organizational principle of the brain. And what should we think about the mind if the mind is not a computer?
Perhaps we should sort of learn to relinquish our hope of having single, elegant and physic like deductive theories in psychology in favor of scrappy, pluralistic and radically, in a way, radically disunified approach to the mouth. Maybe our mind are complex in the sense that there is no single perspective that captures all the relevant factors about them or the factors we care about about them.
Yeah, and it also seems the kind of science that's in a semi perennial foundational crisis like count from the 50s on we have had the cognitive revolution. Then we had connectionism, which was a revolution. Then we had dynamicism, that was a revolution. Then we have embodied cognition and that was a revolution.
[00:01:41] Speaker B: This is Brain Inspired Powered by the Transmitter hey everyone, I'm Paul Middlebrooks. That was Marco Fakhin, my guest today. Marco is a postdoctoral philosopher of neuroscience and cognitive sciences more broadly at the University of Antwerp.
He and his colleague Farid Zanoon recently hosted a workshop called Beyond Neurocomputationalism with themselves as speakers and a handful of other speakers, all, almost all of whom have been on Brain Inspired. In that workshop they discussed many topics around this sort of forever ongoing reassessment in neuroscience and philosophy about how best to think about cognition, the role of brains, the 4e cognition, embodied, inactive, embedded and extended.
Those are known together as 4E cognition, although there are more E's like ecological, ET cetera. They talked about how much biological detail matters for a good explanation, and so on. And the talks from that workshop are online and I link to them in the show Notes. So today Marco and I discuss how all that went and many of the topics and themes that I just mentioned, plus his own work and ideas along those lines. And I also link to his work that we discuss in today's conversation. So go to the shownotes at BrainInspired Co podcast 244 for those. You can also go to BrainInspired code to learn how to support this here podcast on Patreon and get some extra bells and whistles. Thanks for being here. Thanks for your support. If you do support Brain Inspired on Patreon, enjoy.
[00:03:26] Speaker A: Marco
[00:03:29] Speaker B: Beyond Neurocomputationalism so you recently held put together organized a workshop called Beyond Neurocomputationalism. When I was I was very interested in it. I wished I was there and the talks are now online, actually, so people can go and watch the talks and it looks like a really good time.
Did I miss out on a really good time?
[00:03:53] Speaker A: Yes, but I am, of course, a biased source on that.
I think it was a success as a workshop. I think we didn't expect it to be that much of a success.
We planned it to last two days, and we planned to receive about 20 submissions. They were one order, almost one order of magnitude more. Wow. And we thought about extending it to, say, three days, but it was not possible for logistical reason. And as we told the keynotes that the workshop lasted two days, they had planned around that.
So, yes, I think it was a success.
I hope it was important and formative for everyone who participated. I think we've done a great job, or at least a good job, in allowing various perspectives, including the more traditional computationalist one, to be represented.
And.
[00:04:58] Speaker B: Oh, I was going to ask about that because it seemed like with the speaker lineup, at least that it seemed everyone was sort of on the same page. Maybe, but maybe I was missing something.
[00:05:10] Speaker A: No, I don't think that everyone was on the same page. I think that Kevin Mitchell was still probably defending representations and perhaps also computations, at least as a modeling practice. I myself don't have any particular
[00:05:30] Speaker B: trouble
[00:05:31] Speaker A: with computation using computation in modeler.
[00:05:37] Speaker B: Well, that's. I think that's.
[00:05:38] Speaker A: Victor, tell us what the.
[00:05:39] Speaker B: What the workshop was about, because, I mean, it's interesting that you. What do you mean? You don't have trouble using computation in modeling? Because that's. That's different than thinking of the brain as computing, right?
[00:05:51] Speaker A: Yes. So that's.
That's interesting. So what the computation was about, ideally, was this.
Exploring ways to do cognitive science and cognitive neuroscience that do not rely exclusively on computations as a modeling tool or as a sort of, say, organizational principle of the brain.
[00:06:25] Speaker B: And
[00:06:29] Speaker A: what should we think about the mind if DMind is not a computer? That is probably how I would summarize the big theme of the workshop.
That said, wanting to explore what's outside usage of computational practices doesn't entail that you should abstain entirely from computational practices and computational modeling.
And I think that much of my resistance at least has to do with the view that the mind and the brain are computer.
So with this sort of metaphysical claim about the nature of minds and brains.
[00:07:15] Speaker B: Right.
[00:07:17] Speaker A: But it's also something I'm perhaps constitutively confused about, in the sense that I'm no longer sure that I understand what is being claimed when people Claim that the mind is a computer.
[00:07:38] Speaker B: Okay,
[00:07:41] Speaker A: so let's take one example. Galistel is probably still defending the view that somewhere in the brain there must be something like a tape of a Turing machine
[00:07:57] Speaker B: or Randy Gallistol that you're referring to. So I just want to make sure we get the full name in there.
[00:08:02] Speaker A: Oh yes, true.
Or something like that. Functionally like that. And that is a kind of way of thinking of the brain as a computer that I clearly understand.
And in this sort of picture, you expect your formalism, your computational formalism to be a sort of direct guide to the discovery of brain structures.
I'm not super sure that most people who call themselves neurocomputationalist think that way about the brain and computation, I think, but I can't speak for them. But from my outside perspective, what I see them doing is using a lot of maths to describe what happens in the brain.
But that doesn't come in and by itself with this sort of metaphysical commitments.
[00:09:08] Speaker B: No, they're two separate things. I mean, you have to use math to. To create models, but you don't have to. But it wouldn't make sense to avoid computational models to understand whatever aspect that you're trying to understand in the brain. Right.
Are we conflating those two things?
[00:09:29] Speaker A: I'm not sure. I think that sometimes we are and sometimes we are just not caring about whether we are confused or not. Insofar we discover stuff that we deem interesting and important, We have to use math. But using math in and by itself doesn't even amount to using computational theory.
So there's a lot of math applied to the brain that's at least to my eye, not immediately computational or something that I can't immediately sort of cash in in terms of computational functionalism, computational formalisms like Turing machines or automata, this sort of thing is not really clear to me. Whether counts just as the application of computational formalism to the brain or just a more general practice of modeling the brain using maths.
And a for theory is not clear to me whether the fact that 90% of perhaps even more of neuroscientists call themselves computationalists and apply math to the brain should be read off as signaling any sort of commitment of any genuinely health commitment to the view that the brain is a computer or that the mind is a computer.
[00:11:03] Speaker B: Being a neuroscientist, my claim would be that probably upwards of 90 higher than 90% don't even think about it and just have that implicit assumption baked in, because that's how we were taught essentially.
So I'm not sure that it's a, it's not something that's carefully, carefully considered.
[00:11:28] Speaker A: I think I do have the same impression and I'm very sorry, I have never managed to put it in the form of a publishable philosophical arguments, but I do have the growing impression that the reason why we are talking about computations and representations most of the time is purely sociological.
And it's just, it has to do with the fact that we have been raised that ways, that way during our time at the university when we started neuroscience. And also because trying to not use the computational and representational lexicon opens one to be charged with behaviorism and that's considered extremely bad. It's borderline crime and no one wants to be charged with behaviorism. So for good measure, let's talk in terms of computational representations.
[00:12:36] Speaker B: Well, what's driven me kind of batty lately is I want the term computation to mean a definitive thing and you know, like Turing computation for example, or we need to have different modifiers before the term computation because I think a lot of people, neuroscientists think that well, whatever the brain is doing, it is computation. It's just whatever, whatever the brain is doing, we're going to call that computation. And that's unsatisfactory to me because that means computation is trivial.
[00:13:16] Speaker A: Maybe not. That means that computation becomes inescapable, becomes what definitionally becomes what you will find by looking at the brain. If you define computation in terms of what the brain does.
And I mean I don't have any super strong argument to force people not to define computation in that way other than presumably we would like claims to the effect that the brain compute and the brain represent to be empirical claims, that is claims that are true in virtue of what really happens in the brain.
Yeah, so there seems to be something intuitively wrong in using that definition of computation.
But, but if you look at the work of say Vicente Raja, which was one of the speaker at the Beyond Neurocomputationalism, he has this very nice and I think productive idea of motifs, of conceptual motifs that are I think his own very words usefully open ended ideas that everyone can spin and recount in its own peculiar way.
And he thinks that computation is such a motif. And so it's usefully open ended idea that you can play in the language of Turing machine and using automata formalisms, but also you can play using artificial neural networks.
You can play using what? Piccinini, Valtiero Piccinini, a defender of Computationalism called sui generis computation, which is the specific type of computation in the brain.
And if, as a matter of fact, the concept of computation works that way.
[00:15:22] Speaker B: It's trivial, then it's trivial. Actually. You're not gonna.
[00:15:25] Speaker A: But we have to make peace with ourselves and live with it.
[00:15:31] Speaker B: Yeah.
[00:15:32] Speaker A: It seems something that we have no.
No sort of way to rationally push out from our talk and discursive practices because it's can fit in whatever space we allow for it.
[00:15:47] Speaker B: That's right. But I think maybe one of the dangers is it's easy to slide into, like, computational functionalism and just say, well, okay, the mind is whatever the computations are.
You know, like the consciousness is computation. And then we think all of a sudden, oh, AI must be conscious because there's a lot of computation going on.
And if that's the wrong way to understand the mind, then it's a poor way to explain brain processes.
[00:16:24] Speaker A: That's hard. That's hard also because I'm a bit susceptible on computational functionalism.
Perhaps in some ways. I still like the view, at least when applied to consciousness, it seems to me more demystifying than many alternatives.
What I'm going to say is perhaps this, that computational functionalism is at least in a sense, something you could, in a way, try and check for, in the sense that if you have a computational equivalent of me which is not conscious, then you have falsified computational functionalism. Good luck in checking whether my computational equivalent really is conscious or not.
But it's something that you could in principle falsify.
The problem I have with computationalism being played as a motif, if it is indeed played as a motif, is that you can't even do that.
There is nothing that is gonna in principle falsify or count as evidence against it, because whatever you find, it will be computation definitionally.
So I think that there is still some sort of subtle difference there.
[00:17:58] Speaker B: Well, I thought that you were more susceptible to the 4e approach to understanding cognition. So where do you sit now? You've thrown me because now you've just told me that computational functionalism might be the right way to go about it. But I thought you had a lot of sympathies with the four E embodied and active, extended.
What's the fourth?
[00:18:29] Speaker A: Embedded.
[00:18:30] Speaker B: Embedded, yeah.
Approach to cognition.
[00:18:34] Speaker A: Or ecological, if you want. Because embedded is not.
[00:18:37] Speaker B: There's lots of. There are more E's. I know.
[00:18:39] Speaker A: Yes. So am I a 4E guy?
Maybe? Yes.
I've written stuff in favor of 4 ecognition.
[00:18:50] Speaker B: Right.
[00:18:51] Speaker A: I don't think that 4 ecognition is necessarily inimical to computational functionalism.
Andy Clark surely is.
4e person is the person who invented the extended minds. So it's one of the E's. The big E's is his, but is in some sense a computational functionalist.
So where do I sit?
I don't know. I don't feel super pressured to decide whether I really am an ecological person or an inactive person or more traditional computational functionalist.
Maybe I'm in between.
I'm.
[00:19:42] Speaker B: Well, you can't be there. That's impossible. You can't do that. Right. You have to choose a side. Is my understanding
[00:19:49] Speaker A: who could be.
Maybe would have been a smart move to choose a side, at least when it comes to academic marketing. And I think that at the end of the day, I am generally perceived as a 4e person, but I'm not necessarily inimical with everything computational functionalism and computationalism has to say.
I don't have very big problems, ethical problems, in trusting computational models and computational modeling practices.
And I am not super worried that we are gonna miss out on something about consciousness if it turns out that the essence of consciousness can be appropriately captured by computational models. That's probably my position. I strongly lean for E.
But I'm not a fanatical 4e person.
Would that be an acceptable answer?
[00:20:59] Speaker B: Sure, but. So my perception is that 4e is becoming more popular.
I don't know if that's in philosophy though, or in neuroscience. It's hard to say. And I'm curious, as a philosopher, how do you. Where is 4e right now? Are people liking it more?
I guess as it becomes more popular, its critics also become more vocal. But. So how do you see 4e right now in terms of acceptability, popularity?
[00:21:36] Speaker A: That's a nice question. And I think that the right answer is I can't answer that question in that form because I don't think that 4ecognition is being received and manipulated, so to speak, as a single bundle of ideas.
I think that many aspects of it are having different lives and different degrees of acceptance. So extended cognition apparently is accepted by the majority of analytic philosophers. In the last field Paper Surveys, 2023, I think was published, it's published by David Purcha and David Chalmers.
51%, roughly 51% of analytic philosophers said yes to extended cognition.
[00:22:35] Speaker B: What is extended cognition?
[00:22:39] Speaker A: The actual piece of data. Extended cognition is the mainstream view.
And extended cognition is, as I understand it, the view dots the physical gears and the physical machinery of at least some cognitive processes is not entirely encased in your scope. That stuff that's outside your brain, physically outside your brain can be, at least in some cases, a gear in the sort of machinery that makes you think, makes you think has to be understood by broadly including feeling and perceiving and so on and so forth. So the go to example is of course the smartphone, which is happily substituting for a lot of our memory like we used to. I'm old enough to remember the grand old days of people having to memorize numbers with their phone numbers with their brain.
Nowadays I don't think I actually know my own phone number anymore. I just copy paste it from WhatsApp.
[00:24:03] Speaker B: So that'd be an age related problem.
[00:24:07] Speaker A: I hope I'm not that old yet.
[00:24:09] Speaker B: Yeah, yeah.
[00:24:11] Speaker A: So that's, that's the idea and I think that's.
That is becoming the majority view, at least in philosophy.
Details can sort of shift from one philosopher to another.
Standardly said that there are three waves or three big ways to think about how the mind extends. There should be a so called first wave in which the mind extends. If something external to the brain, something external to the brain is doing something that's roughly isomorphic, at least computationally isomorphic to what would happen in the brain and that's typically attributed to Andy Clark. Then there is a second wave that's typically attributed to Richard Menari from Macquarie and Mark Rowlands and other people where what stressed is the complementary between inner neural processes and outer processes. There's also a third way which is more inactive and that stresses how the mind itself is constituted by external interaction with the environment.
And that perhaps is also open to the idea that our phenomenal consciousness is extended in this way, that the machinery that makes us conscious is not entirely in the brain.
And that is a view that seems to be roughly well accepted amongst philosophers.
The idea of embodiment is in some forms well accepted. I think that's.
I would say that most neuroscientists accept the existence of modal bodily representations and that's such modal, broadly modal and bodily representation play a relevant role in our cognitive processing. They're not just the exhaust fumes of our neural activity.
So that seems an idea that's broadly accepted.
I agree.
[00:26:19] Speaker B: I think that among the four E's, embodied is the most accepted in the neuroscience community. I think, yes,
[00:26:30] Speaker A: I have that sense too, or maybe already people arguing for supermodel representations that are hostile to it. But other than that, I think it has been broadly accepted. The idea of affordances has been broadly Accepted affordances are everywhere, but they have also been mutated in an important way. They have been in many cases turned into mental representations of sorts.
[00:27:01] Speaker B: Yeah, that's kind of a tricky thing that neuroscience has done. People like Louis Favela are pretty upset about that. He's an ecological psychology advocate. And that's tough. You know, it's like the term computation, right.
All these terms get change, you know, semantic drift, et cetera. But people sort of take a sexy term like affordances, which was I think coined by Gibson. Yeah, by James Gibson, and means a very specific thing. And then it's kind of a unique term and then people just kind of have brought it into the computationalist approach. Right, so an affordance is the essentially the what is available for action essentially in an organism's environment.
Yes, but it has become a. And so that has nothing to do with like you don't represent an affordance.
It's not something internal in the brain, but it has been brought into the brain by modern neuroscientists.
[00:28:09] Speaker A: Yes.
[00:28:09] Speaker B: Yeah.
[00:28:10] Speaker A: So that's an idea that maybe has been borrowed, maybe has been appropriated, depending on your point of view. This is an idea with a, I would say a good amount of popularity when it comes to an activism.
Things are, I think more dire in activism, in and by itself is already a sort of tripartite thing.
There are sort of three big families of inactivism.
There's the so called radical and activism by Eric Main here in Antwerp and Dan Hotel, the University of Wollongong in Australia.
And I think that the sort of most perspicuous way to look at it is to look at it as a ideal continuation of a Wittgensteinian research program applied to cognitive science. I'm not sure that Eric Mine or Dan Hodo would use that description, but I also don't think that would reject it outright.
It's a sort of mostly philosophical movement aimed at getting a sort of concept of clarity about what's being said by cognitive scientists.
And especially Dan Hatto has been campaigning lately in talks clarifying that his project was mostly negative in the sense of let's eliminate the confusions, let's not get trapped in concept of quagmires.
What I'm doing is this sort of work. I'm not willing to put forth a new research program or a semi empirical research program to substitute cognitive science.
So there's that and I think that's not very well.
It has not been very well received by cognitive scientists. But perhaps because they expected something more positive, something more Like a positive research program.
So there is that there is so called sensory motor and activism which is typically associated to the names of Alvano and Kevin Oregan. Alvinoi, I think in Berkeley, Kevin Oregan, I'm not really sure, I think in Canada, but I don't remember specifically which university.
And their original idea, as I understand it, is to explain the character of phenomena of consciousness.
That is why does experiencing stuff feels the way it feels in terms of what they called sensory motor contingencies. That is specific, regular and low like patterns of interaction between the agent and the environment and specific ways in which that interaction changes our sensory stimulation when we experience.
I think that their idea of sensory motor contingencies has been fruitful and has had a career similar to that of affordances and that it has been appropriated by computationalists.
Most model of sensory motor continuances I can think of are computational indeed.
And I think that their original view of accounting for the character of consciousness has been broadly ignored, perhaps by the larger cognitive science community in the sense that it has been discussed.
Seems to me to be laying sort of dormant right now.
[00:32:26] Speaker B: Well, we, we really enjoy ignoring philosophers of all sorts. So, you know.
No, no, no, I'm not one of
[00:32:37] Speaker A: those people who enjoy it too.
[00:32:39] Speaker B: Do you feel ignored as a philosopher?
[00:32:45] Speaker A: That's a tricky question.
I think that the answer is it depends on which sort of cognitive scientists are we talking about.
And I think there are the majority of cognitive scientists that work in a largely a theoretical manner or that they are not interested in checking the validity of or testing or just thinking about broad metaphysical framework about the mind or DNA.
[00:33:22] Speaker B: And that's fine in some respects, right?
[00:33:24] Speaker A: You can entirely fine.
Also there are the majority, and I think that the majority of our discoveries about the mind have been produced by this sort of people in this sort of war.
And perhaps their ignorance is bliss in the sense that you do your experimentation, you care about your experimentation and its technical details and then you let
[00:33:50] Speaker B: the
[00:33:51] Speaker A: big picture to be dealt with by someone else. That's a perfectly fine division of sort of academic and intellectual work.
So those people ignore us, but I don't think it's problematic.
What is more problematic is when cognitive scientists want to do metaphysical or metaphysically adjacent theorizing and ignore philosophers. But I think that's in the limits of the reasonable.
They are not ignoring philosophy altogether.
They are not willing to dive head first in the super subtle philosophical minutia, but at least the one I can think about. They do know their Daniel Dennis under Fred Dresky and Ruth Milliken.
[00:34:52] Speaker B: So
[00:34:55] Speaker A: I wouldn't say that philosophy has been ignore when it matters.
[00:35:01] Speaker B: As a practicing neuroscientist who is also interested in the metaphysical questions, I often when I'm reading philosophy I do get sort of bogged down in the minutiae and the details of like the semantic arguments and the positions that are being defended.
And I just want to go in and get what I need and go out. And that seems impossible in most philosophical works. I'm not criticizing you at all because I enjoy a lot of philosophy.
But then even like you know, when you're, when you were explaining extended cognition or inactive, you know, everything has 12 different camps and they all have their own subtleties, you know, and it's like it's too much. I just want to do my experiment and get on with the day. Right, Yeah.
[00:35:49] Speaker A: I mean that makes sense. And I would say an unavoidable effect of the fact that our intellectual labor is divided and is divided for efficiency. Right.
You want people doing the hard stuff on the field, to be focused on the hard stuff on the field and on the specific math they need to use, the specific experimental technique they want to use.
And you want theories to be as subtle and as analytic as possible, even when that means taking into account wild and far fetched scenarios.
Perhaps we are not good at making those two groups of people communicate, but I don't have a solution to offer either. Right.
[00:36:43] Speaker B: Yeah.
[00:36:45] Speaker A: And perhaps there is no solution insofar our interests are different.
Like take as an example medicine.
There exists philosophy of medicine.
And some of the questions that philosophers of medicine discuss are what are diseases?
What does it mean to say that something is a disease? Does a disease necessarily involve some sort of pathogen that's attacks your body and causes disruption to your body? In that case, obesity couldn't be possibly be a disease, for example.
And that's a perfectly fine discussion.
But you surely don't want surgeon to pose meat operation to contemplate whether obesity is a disease.
And in an important sense we don't even care that surgeons spend their free time reading philosophers of medicine.
And I don't think that there is any sort of social societal pressure for them to do so. In this sense, the sort of societal infrastructure of cognitive science is odd in that it is expected for practitioners on the field to spend their free time reading the big theory and engage with philosophy.
That's quite odd. That's not what tends to happen with most sciences. Right. Is that right though?
[00:38:27] Speaker B: I don't know that it is expected.
Yeah.
I don't know.
[00:38:33] Speaker A: Well, if you think about the cognitive accident, philosophy is there is one of the edges. So it is a part of cognitive science.
And that's very weird because no one would think that philosophy of medicine is part of medicine. It's a separate intellectual endeavor that's in a way a second order type of inquiry that takes that, that observes the first order kind of inquiry. Medicine, the biological sciences.
[00:39:11] Speaker B: That's kind of an interesting example because like the brain, right. Medicine is about depending on what kind of medicine it is, it has kind of the same problems, like you're treating a complex system.
So when you bathe it in some drug, for example, you, you will, you might have longitudinal side effects that I don't know if philosophers are the right people to address this sort of thing. But, but medicine kind of treats the body not holistically, but as a domino, a set of dominoes or billiard balls. Right. And, and you kind of go in with a, like a baseball bat and just swing at the disease. And if it works, if it improves symptoms by 12% or whatever, then it is the treatment. Right. So. So I'm not sure that medicine shouldn't be paying more attention to philosophy of medicine.
And maybe, maybe cognitive science has it right then because you know, we're trying to understand this thing that is ineffable, the mind. Right.
And so I think you need all hands on deck.
[00:40:26] Speaker A: Perhaps my gut reaction is to say that yes, you're right, but my perhaps more analytically trained mind points out that cognitive science is the odd one out. And it's also presumed, perhaps the set of science or scientific endeavor where progress seems harder or less swift.
[00:40:55] Speaker B: Yeah.
[00:40:55] Speaker A: And it also seems the kind of science that's in a semi perennial foundational crisis like counts from the 50s on we have at the cognitive revolution. Then we had connectionism, which was a revolution. Then we had dynamicism, that was a revolution. Then we have embodied cognition and that was a revolution.
Then there was 10 years ago or something, predictive processing. And that seemed to solve the minds and provide us a sort of complete way to observe the mind, brain and everything, living system and so on and so forth.
[00:41:37] Speaker B: Don't forget the Bayesian brain. That was big when I was coming up. Yeah.
[00:41:41] Speaker A: Yes. It was sort of not sneaked in, in predictive processing for me. But they are different.
That doesn't happen with medicine. Like they got germ theory and it has been there for roughly 200 years and it doesn't seem to change that much. Or maybe it does, but I don't know it from the outside. It seems way more conceptually stable than cognitive science.
[00:42:14] Speaker B: We're continuously in crisis. I like that.
What's your take on predictive processing?
There's a problem with it, I do
[00:42:24] Speaker A: think that's suffering from its incredible success in the early 2010s where there was a huge influx at least among philosophers of people working on predictive processing, trying to use it as a sort of way to solve any problem about the mind.
I think that Freeston work on the free energy principle fomented this sort of hype concerning predictive processing being dissolved or the free energy principle being this sort of ultimate theory of everything.
[00:43:07] Speaker B: Well, it didn't start off that way though.
[00:43:09] Speaker A: It didn't start off that way, but it ended up being that way. And I think that it started to accumulate problems quite fast.
Problems and objection. Quite fast.
One of which is that no one was quite able to understand what was going on with Markov blanket's anti formalism that was ever changing and the fact that Carl Friston was constantly producing outputs and preprints that were making conflicting statements. And I think that the interest waned in part because of dots, in part because of philosophical analysis. Here the paper by Elle Brunenberg is probably the one that made the biggest splash. The Emperor knew Markov blankets and that sort of was a call to do publicly admit that what has been the most discussed theory of the mind and brain had some problems and need some rethinking.
And when it comes to me, I think that in its more maximal version predictive processing doesn't give you a good theory of the brain. If what you are interested in is understanding the sort of fine structure of the brain, why near areas differ from each other, for example, is something I don't see predictive processing offering an answer to.
[00:45:02] Speaker B: Why? How does it fall short? I mean, I can probably state this and then you can correct me, or I can just let you state why it fall short.
[00:45:10] Speaker A: I think that the reason why it falls short is that if you take predictive processing in its sort of maximalist form seriously, what's relevant about the brain is that it does prediction error minimization.
The explanation must cite prediction error minimization and so forth.
[00:45:35] Speaker B: The explanation, the single explanation of the brain.
[00:45:38] Speaker A: Exactly, it must.
Then I don't see how you're able to explain why V1, V2, V3 and M1, M2, M3 are differently shaped.
They are connected different way, they have different morphological and anatomical characteristics. Aren't they doing in the relevant sense the same thing. Then why do they differ so much?
And I don't see predictive processing in that form answering that.
I think that predictive processing and the cluster of ideas associated to predictive processing are useful. I don't think that they are the only thing that matters.
This sort of decade of hyper focusing on prediction error minimization may sort of have harvested or the most easily accessible theoretical fruits. From the perspective superhaps, it could be advisable to do something else for a while.
I mean, go back to it.
[00:46:52] Speaker B: Well, in your view, is there gonna do we need a unified theory that explains it all or is it more likely that we're going to have a more perspectival group of different cuts into understanding brain and mind?
[00:47:12] Speaker A: Well, the universal, elegant, simple theory that explains it all.
Super attractive is super attractive.
The question is, is it likely?
Are we likely to stumble upon it?
And perhaps the answer is yep,
[00:47:32] Speaker B: predictive processing sort of like is attractive also. Right. It seems to make a lot of sense as a one of the candidate unified theories.
So I can understand why people will be super excited about something like that. Because if you have, if that's the mechanism, ah, now you know what to look for. Now you can actually go and perform experiments and look for these sorts of things, right?
[00:47:58] Speaker A: Yeah, but it didn't really work out.
[00:48:01] Speaker B: No.
[00:48:05] Speaker A: So perhaps the best we can have is a mosaic of different theories and different insights and perhaps we should sort of learn to relinquish our hope of having single, elegant and physic like deductive theories in psychology in favor of sort of scrappy, pluralistic and radically, in a way, radically disunified approach to the mind.
Maybe our mind are complex in the sense that there is no single perspective that captures all the relevant factors about them or the factors we care about about them.
[00:48:56] Speaker B: Do you think that modern artificial intelligence has helped us understand brains and or minds?
It was a sharp turn there. Sorry, but because I want to get on at some point we're going to talk about your a leaving work which is.
Yeah, but I want to. But we'll get there. But I kind of wanted to get your take on AI in general because you know, there's the term neuro AI has become a popular term in the neurosciences and you know, there's all this sorts of work from people like Jim DeCarlo and Dan Yemen, many of the many people using deep neural networks as proxies as models for brains and finding alignment between the activities of the units and the model and the activity of single neurons in Brains and that's, you know, I'm not describing, I'm not doing any justice to the amount of work and the amount of cool stuff that's been found. But then you have, you know, because we were talking about predictive processing and oh, it's cool because then you know what to look for. Well, something that came out of AI is like, oh, backpropagation works, therefore it must be in the brain. And there's been a lot of work like looking for back propagation in the brain and it's sort of a backwards thing because it's this artifact that we produced and has shown that like here's a really inefficient but workable solution to learning in this model. So therefore it must be in the brain. Whereas you know, we, we built these things originally from a very, very simplistic idea of what neurons might be doing and really hasn't changed that much since then. So I'm, I'm curious what your viewpoint is on neuro AI in general and or using models like that to understand brains or are the models too complicated themselves?
[00:50:49] Speaker A: I have a super strong opinion about that and I'm very ignorant about the, that line of research.
Of course calling upon computational system has helped us do research and about the mind. So it has had a positive impact in that sense.
One worry is that we should start look elsewhere and consider the mind from different perspective and the brain from different perspective to get a grip on other source of phenomena or to get a better grip on what we already know. For example, chemical diffusion and of chemical diffusion in the brain is something you sort of ignore or downplay if you only do computational modeling or only do deep artificial neural networks as models of the brain. But it's presumably important deep artificial neural networks focus on neurons.
And yet the modeling of neurons and making inferences about neurons, if you use them as a proxy to investigate the brain, that is the brain is not made up only by neurons.
There are glial cells and there are people suggesting that glial cells are not cognitively nerves.
I don't know how you could plug this into an artificial neural network,
[00:52:34] Speaker B: but
[00:52:34] Speaker A: that's something, it's worth investigating.
So I guess that the answer is yes. But as every model, they have a price. If you pay attention to something, you need to disattend other aspects of that very same thing.
[00:52:50] Speaker B: Yeah, I find it kind of fascinating that.
So, you know, I was brought up in computational neuroscience and I was in graduate school recording extracellular spiking neurons in non human primates and Everything you learn is about the action potential and that must be the currency of the mind. That's how everything is done. And it's. I just find it fascinating that like, okay, here's this electrical signal that looks different than other kinds of electrical signals and can travel far distances. So we're going to put absolutely all of our effort into counting these things and making models that talk about the shape of the action potentials. And that's like all we measure. I mean, again, I'm not doing justice to many, many great scientists who do other sorts of neuroscience, but I just find it fascinating that how much of our effort is devoted to counting spikes.
[00:53:54] Speaker A: They are surely one of the most impressive things that happens in the brain, for sure. So it makes sense.
We had a very nice talk by David Colasso and the paper is now, I think out in Brain and Behavioral Science about the role of metabolism in computational model.
[00:54:20] Speaker B: Oh, yeah, yeah, yeah.
[00:54:22] Speaker A: He's by David Colasso and Philip Howis. I hope I pronounced their name correctly and in the right order as they appear on the paper.
If not, I'm sorry.
And their point is? Well, look, we should at least test our computational models and constrain our computational models with metabolic considerations, which is in a way just a point of good sense. Of course, the right computational model that captures what really happens in our mind must be metabolically efficient or at least sustainable.
[00:55:07] Speaker B: Part of their argument is that it just, it should be within the energy bounds that a physical brain would use, right?
[00:55:15] Speaker A: Yes, I think so.
But they're also calling to other usages of metabolic consideration, for example, to arbitrate between models to perform model choice.
And that's a very nice idea. I don't know how it will be integrated in the actual practice.
[00:55:37] Speaker B: I don't think it will.
[00:55:38] Speaker A: He's already integrated in the actual practice.
[00:55:43] Speaker B: That's the thing about like computational functionalism, right? Is we don't.
You don't, you don't care about energy. It's all about the computation. So it doesn't matter if it takes a sun to run the thing. If it's doing the same thing that your energy efficient brain is doing, it must be thinking.
So I'm not sure how accepted like an approach like that will be.
[00:56:09] Speaker A: Sure that's right. But presumably neuroscientists are trying to model our brains and our minds and not brains or minds in general.
[00:56:21] Speaker B: Well, I think we're trying to model computations. I think we're trying to model the computations, quote unquote, that our brains perform, not our Brains.
Right. Because if we were trying to model the brains, we would do more than spikes.
[00:56:36] Speaker A: That's nice point.
My sort of philosophical gut reaction to that, to your claim are just trying to model computation is wow, that can't be true. Otherwise everyone would be fine with very big lookup tables. Those are modest of computation too.
But I'm not really sure it would be the appropriate reaction.
I think that right now the game is that of identifying plausible models. And then perhaps the choice between plausible models on the grounds of timing consideration or metabolic consideration might be postpone to a second moment when we will know more.
Maybe that's the sort of idea of division of labor that's taking place.
Or maybe not. Maybe we just didn't pay attention to metabolic and chemical and other sort of consideration at all and we should start to integrate them right now.
[00:57:46] Speaker B: The problem is it's a lot of work to integrate everything in a complex system.
[00:57:50] Speaker A: Yeah, it's a lot of work. And it's also hard to master all that knowledge.
[00:58:03] Speaker B: Oh, it's impossible. Yeah.
[00:58:04] Speaker A: So yeah.
And this means that there are physically, there aren't people able to perform that work either alone or in group.
[00:58:15] Speaker B: Right, Right. Maybe that's where AI will help. Perhaps.
[00:58:21] Speaker A: Could be.
[00:58:23] Speaker B: In your talk at the, at the workshop you talked about the difference between a Platonic brain and a protean brain.
[00:58:30] Speaker A: Yeah.
[00:58:31] Speaker B: And presented an alternative. And I don't. I think that was the first time I'd heard of Proteus. So can you describe, can you, can you describe the who. Who Proteus is? We all know Plato, I think, but what the issues are and where you come out
[00:58:49] Speaker A: who Proteus is. So Proteus is a Greek, I think, deity which was famous for shape shifting away from problems.
And the idea there is that.
Well, we are all familiar with a quote unquote Platonic brain model where function maps onto neural structures in a way that's super abstract and entirely rigid. Like Ha. Broca's area does language full stop.
[00:59:24] Speaker B: Verification finds edges in visual scenes. Yeah.
[00:59:29] Speaker A: Diffusiform face area.
Detect faces full stop.
[00:59:34] Speaker B: Thalamus is just a pass through though.
[00:59:37] Speaker A: Exactly.
The subcortical stuff is mysterious.
Works in mysterious weight on us.
So that way has been largely rejected. I think that even the people that look for individualized functions in the brain are looking for more finely grained defined function to be mapped on physically smaller and more specialized neural area. So that view of the brain has been described. But it seems to me that at least when it comes to embodied cognition and embodied neuroscience, radical Embodied neuroscience that generated a sort of antipathy for the idea of mappings structures and functions in the brain. And I don't think that that's a positive thing. I think that even if you think that the brain is a complex system and that really is undecomposable and thus it makes no strictly speaking sense to understand that bit of the brain in opposition of that bit of the brain, it's still at least heuristically useful to decompose it and to assign more or less intelligible and stereotypical functions to bits of the brain.
Is it?
[01:01:11] Speaker B: Go ahead.
[01:01:13] Speaker A: And as a reaction to the standard picture of the brain, embodied cognition tends I think to go in the protean direction and to.
[01:01:21] Speaker B: You think it's swung too far in the other direction?
[01:01:24] Speaker A: Yes, I think it swung too far. It's. It tends to want too far in the other direction and to either ignore the problem or just say yeah, everything can do everything. If the context is right or if you train the brain well enough and thoroughly enough.
That doesn't really seem to be the case. Even the I think source is the famous author of the reward ferret experiment. Even theory wire ferrets weren't able to discriminate visually discriminates with their. I don't remember precisely with their non visual cortex. Visually they were able to perform some gross discrimination but not the whole thing.
And yeah, ignoring this sort of limitation, it may cause embodied cognition to shoot itself in the foot.
There is something like being over correcting and resulting and ending up saying unacceptable stuff because you are over correcting. And I think my talk and that paper will be aimed to avoid shooting ourselves in the footwell for corrupting.
[01:02:46] Speaker B: So I mean it's like a reasonable middle ground essentially is what you're arguing for
[01:02:53] Speaker A: perhaps. Yes, I hope.
I think that's the way in which I would like to the paper to be perceived is in terms of guys don't and girls don't hypercorrect, not in terms of the brain. Must be something that's a bit. A bit also structurally rigid and non plaster.
[01:03:19] Speaker B: Where does. Because you brought up after Phrenology, Michael Anderson's book and so he kind of has this view that I'll use the term heterarchy, that it's not a hierarchy in the brain, that the brain is a heterarchy where different parts of it can be in control based on the context and needs at a given time and form these different brain areas can form these coalitions on the fly and it's all a big dynamic Beautiful mess.
And so. But is that. That seems to be kind of in the middle to me as well. Is that too far?
[01:03:53] Speaker A: It is in the middle for me, but I don't think that that is the go to understanding of the brain.
Embodied philosophers of mind and philosophers of mind, neuroscience. Oh, you.
[01:04:09] Speaker B: Oh, okay. I think. Okay.
[01:04:11] Speaker A: I think that what they have in mind is much more mushy in lack of a better term.
Like Anderson has this idea of workings, perhaps is the terms he uses in after Phrenology that is a sort of hardcore unchangeable function that you can map onto individual neural area and that working can give rise to a variety of different psychological processes or can be used in a way, a variety of different psychological processes depending on who the partners of the neural area are in a given occasion.
So there is this sort of rigidity in Anderson's view, even if I don't think he would particularly like the way I describe this, but I think it's there.
I think that the culprit here is sort of Clark's view of the brain in his surfing uncertainty, where he probably was trying to say something like Anderson to sort of embody Anderson's view, but he ended up, I think, literally saying, well, everything the brain does is prediction error, minimization.
Bits of the brain are recruited by the quote, unquote expected precision mechanism, depending on what's needed at the moment and how hard the cognitive task is. And so the idea of functional specialization and of different, in a way, neural organs has been partially lost when the view was repeated.
[01:06:03] Speaker B: But isn't a more normal view of predictive processing that, yes, you still have visual cortex, but what you're predicting is the visual world? And in, I don't know, like barrel cortex in a rodent, what you're predicting is a whisker, deflection, et cetera. So there is specialization that is amenable in predictive processing. But you're saying that. That people like Andy Clark have just taken it to its extreme and say, we don't even want to talk about visual cortex, for example. It's all prediction. And that's. That's where we need to stop describing it.
[01:06:40] Speaker A: I think that there's something true about what you said, and that's the kind of reaction I get from the predictive processing people when I talk about this sort of stuff.
The problem is that it seems to end just after the sensory cortexes are over because there is a lot of quote unquote associative cortex in the brain, and it doesn't seem to be predicting anything specific. So good luck in telling me what the specific function of those bits of the cortex is if you adopt that strategy.
And also there are explicit quote by Carl Friston that perhaps due to rhetoric flourishing is on print saying that there is no motor cortex. The motor cortex just is sensory cortex like the rest of the brain.
And if that is has to be taken face value, then it seems that the idea of cortical at least specialization is a bit thrown out of the window.
You just as deminimization of sensory prediction error. That's all there is.
[01:07:58] Speaker B: So do you have.
I'm going to put you on the spot and just ask how just in the grandest sense how you view neuroscience. Is neuroscience succeeding?
Is the field as a whole? Does it look foolish?
Is it. Does it showing some promise? Are we doing the right thing? Are we doing the wrong things? Do you have a view on this?
[01:08:26] Speaker A: I don't think I have any super big view. We are making progress partially because brains are so interesting and rich and minds are so interesting and rich that borderline impossible not to make progress.
I don't think he's a fool's Ireland.
I think as any model based bit of model based science, it has blind spots.
It has alternatives that are left unexplored.
It has aspects of sort of quote unquote brute experimentation without any care for the bigger theoretical picture.
[01:09:14] Speaker B: Do we need to care about the bigger theoretical picture?
[01:09:18] Speaker A: Depending on what your goal, what your epistemic goals are, yes or not? If your epistemic goal is to produce data and collect data and to do good experiment, maybe not.
If your goal, and I think that philosophers have this sort of epistemic goals of producing a theory in some legitimate sense of the term that. Yes.
But I'm not sure that everyone who works in science needs to be interested in the big theory.
So my very unsatisfactory answer is. Well, it depends.
[01:10:00] Speaker B: Depends. Yeah, sure, I accept that.
I wonder what the right proportion of people though is. Is for caring for something like that. There's probably an optimal. It was probably like 80, 20 or something where 20% of the people should care.
[01:10:19] Speaker A: That could be.
I don't know what the proportions are nor what's an optimum is.
But maybe there needs to be also no sort of, how can I say numerical splits in the community in the sense that you need to have a group of people doing that and a group of people doing brutish experimentation.
You can have a single group of people doing theoretical war only a fifth of their time sure. Only the last five years of their career.
Why not?
At the end of the day, I think that our goals are largely, our epistemic goals are largely driven by our pre theoretical attitude and curiosity and by what we feel good at and what we think we are successful at. So if you don't care for theorem, for reading big books, you're gonna be sort of a theoretical experimenter and that's fine.
I don't think that there is any strong reason to force people to do anything else.
[01:11:42] Speaker B: Have you chosen the right path thus far?
[01:11:46] Speaker A: I think so.
I'm happy with looking at the bigger picture. I think I enjoy doing philosophy of neuroscience and cognitive science.
It's rich.
[01:11:59] Speaker B: It's so rich, right?
[01:12:02] Speaker A: Yes. I don't think it's richer than cognitive science or neuroscience itself,
[01:12:11] Speaker B: but I meant as a philosophical endeavor. Right. It's, it's, it seems richer than the philosophy of medicine, for example. Nothing against the philosophy of medicine at all.
[01:12:20] Speaker A: I mean, I guess it's, I'm essentially prone to say yes, I can't understand how there is people with scientific interest that doesn't do what I do, because of course what I do seems very interesting and essential to me, but they could say the same. And sure, I guess that any different at this level is not something that you can't reason through or that's motivated by any sort of reason that can be sort of rationally altered.
[01:12:58] Speaker B: I kind of wonder how philosophy folks sort of choose the topics that they become interested in. And that part of that is just sort of probably what you're learning from your advisors and what your advisors are interested in. I'm thinking that's, that's kind of the way it goes in neuroscience frequently. Right. So you have an advisor and they study a particular part of the brain with a particular methodology, and lo and behold, that's what you end up doing as well. Right, but so, so I don't know if you want to speak to that, but I was going to segue into and ask you how you got interested in what you call a leaving, which is a corollary to believing something, but in terms of the sociological aspects of how people interact with artificial intelligence, entities, machines.
So how did you get interested in that? And tell me more about why we should care about it.
[01:13:58] Speaker A: Okay, So I got interested in that sort of things because a friend of mine, a person who did the PhD with me, Giacomo Zanotta, now he's working at the Milan Polytechnic and started to do philosophy of AI and one day come at me with a project For a small paper on treating, on the idea of treating AI companions and to a lesser degree, robotic companions as effective artifacts, artifacts that we use to regulate our emotion.
He got me on board and I started working on that too.
And the, the concept of elite came, came into our concept or radar, when we started to wonder, well, what's going on in the person mind when they interact with robotic or AI companions as if they were real?
[01:15:03] Speaker B: Is, is alief, Is it like artificial believing? Is that what it's supposed to point to?
[01:15:09] Speaker A: No, it's more of purely associative belief.
Purely associative reacting, I would say.
So the idea comes from alif, comes from Tamar Gendler.
And her starting point are cases where you act out something patently irrational like you forget your wallets.
A friend lends you a couple of dollar bills, you take them, you thank them, and you say, okay, I'm gonna store them in my wallet. And you try to reach for it.
This is clearly rational. You ask the money because you don't believe you have your wallet with you, but you still automatically reach for it. And she had this idea of, in a way postulating a novel mental state to account for this sort of behavior.
And she called this alif. And those are sort of, as I understand them at least, effective behavioral disposition that are sort of quasi automatically triggered by what you see about the ddd, what you perceive.
And the idea is that since especially AI agents are pretty decent in mimicking the kind of outputs you perceive by interacting with a person, at least linguistically, that the encountering that output triggers in you this sort of effective behavioral response that are associated with interacting with a real person.
Even if you don't believe that they are a real person, even if you are deep down convinced that there is no one you are interacting with, you are sort of animalistically pulled towards behaving as if you were interacting with a full blown social agent.
And I think that there is a decent case for arguing that.
The paper in which we do has been accepted, so it will be published soon in Phenomenology and the Cognitive Science. Thank you for allowing me to publicize my own work.
And we argue for that, and I think that that's a better account that what to me looks like the prevalent alternative, which is a form of fictionalism where you sort of pretend play with the computer or with the chatbot, that there is someone there and someone you are interacting with.
That account strikes me as very implausible phenomenologically, like when I talk with ChatGPT I don't feel like I'm pretending anything.
And it also has makes prediction that seem to be false. Like one big thing about pretends is that it tends to be quarantined.
There is a moment in which you start pretending and stop pretending. And there are in an important sense rituals that's quote unquote rituals that define these pace of pretends. So when you have even children pretend playing, sort of free floating manner, there's always the child that starts, ah, let's do that, I'm a cowboy and you are the indoor or that, I don't know, I am Optimus prime and you are Megatron or whatnot.
And then there is a moment in which pretense ends and that is sort of negotiated too by saying I don't know, let's play tennis in the water or whatnot or my mom calls me I need to go home.
That doesn't tend to happen with how we use AI companions. The scary thing is that we integrate them in our life in a sort of seamless way and perhaps as a sort of side effect of that our interaction with AI companion have serious consequences that pretends typically don't have.
They reflect and reverberate in the rest of your quote unquote serious life in a way that what you do when you pretend doesn't.
That was our motivation to sort of propose an alternative account.
[01:19:53] Speaker B: So this is.
Is this a way to assess the potential dangers of our interaction with, with like complete integration of AI into our lives? Because I'm curious like how if it's the same thing.
So before cell phones there were these little devices that were navigator devices and they would have, they would still say in 100ft turn left or whatever.
And I remember my dad talking, you know, responding to the commands and stuff. Is that the same, is that a leaving, Is that the same thing as the full blown interactions that we're having
[01:20:35] Speaker A: that could be part of or behavior partially motivated by an idea.
The fact that apparently most people say hello and thank you and please to chatgpt is sort of.
[01:20:51] Speaker B: I get aggravated when I do that.
[01:20:54] Speaker A: Automatism. Yeah, yeah. I stopped because ChatGPT is so inefficient that I'm no longer prone to treat it in a nice manner in any way.
[01:21:06] Speaker B: Did you swim the other way? You treat it poorly now.
[01:21:09] Speaker A: No, I treat it functionally as a tool.
[01:21:13] Speaker B: Yeah, me too.
[01:21:17] Speaker A: But our point was just in trying to sort of understand at a very coarse grain level what's happening in the user brain.
Does it have a sort of impact on how to assess the danger that those system pose?
Well, maybe yes, because if we are right and this sort of behaviors and behavior disposition are automatically triggered, you can't ask people just to behave better, to use AI more wisely, or just to remind themselves periodically that there is no one really there.
And so perhaps if that's the case, if ALIF is indeed the right explanation and we want to avoid people thinking that were at least behaving as if they were interacting with people when they are not, and perhaps we should design AIs and robots that do not look like people and then don't trigger that pull the same sort of emotional and behavioral lever levers that they pull right now.
[01:22:38] Speaker B: Well, I wonder if there's a generational difference as well. Like I worry about my children who will have grown up, you know, interacting all the time with these things and already there are examples of self harm, you know, people being encouraged to do unhealthy things via their conversations with their buddy AI or whatever. And that seems like insane to me, but I, but I also worry that my children will have, will have less of a filter. I'm not sure where that sits in the a leaf on the ALIEF spectrum, but I worry that they're going to treat the artifacts as real entities.
[01:23:22] Speaker A: If we are right and the design practices don't change, they likely will or they will likely be prone to do that, perhaps we can control our lives at least sometimes and remind ourselves perhaps not. And we need to change our design practices.
My gut sort of reaction would be to change our design practices and to seriously consider whether or not we need to have an integrated AI chatbot in anything at present time.
[01:24:04] Speaker B: I don't think there's any going back at this point, is there?
[01:24:08] Speaker A: I mean, there is no natural law that compels us to have AI childbot anyway anywhere.
[01:24:17] Speaker B: It's the law of capitalism is the natural law.
[01:24:21] Speaker A: Could be, but it could be that you can brand yourself as the product that doesn't hug AI and make a profit out of that.
[01:24:32] Speaker B: Okay, good luck. Good luck with that.
[01:24:38] Speaker A: Yeah. That's why I'm not a software engineer.
So the honest answer is that I don't know.
It seems to me extremely unlikely that we will change design practices. I don't think that there's any real organized demand to do so.
And that's scary because that's scary. That's potentially dangerous because the more people will be acquainted with perceiving chatbots as real people and robots as real people, the more they will get or there will be at least a demand for their legal protection.
And that opens a can of worm I'm not at all equipped to think about.
But there will be serious problems.
[01:25:45] Speaker B: Yeah, I mean, where are you on the AI is conscious or we need to worry about AI being conscious. We need to worry about AI welfare because you're sympathetic to the computational functionalism approach. So if we just get the computations right or big enough, etc, it must be conscious. Right? You must be amenable to that.
[01:26:09] Speaker A: I'm sympathetic.
I'm not sure I buy into any theory of consciousness in any super strong way. So is that a risk, that AI could be conscious and locked in an eternal sort of suffering state while it spews out output?
Yes, perhaps. Is that a likely scenario?
No. Is that a the top of our pragmatic priorities right now?
I don't think so.
So that's not really something I'm thinking about or any strong stance upon.
Yeah. And I won't say anything other than that the class of things that are capable of suffering right now at least look very different from AI chatbots.
And so the chance that the right computational structure for pain is present in both maybe very small. So might not be a super urgent concern, even if computational functionalism is right and AI can be conscious.
[01:27:40] Speaker B: All right, Marco, well, thank you for taking the time. Nice to meet you. And I hope the next workshop you put on, I hope I. I can actually attend. If you're planning on doing something like that in the future.
[01:27:52] Speaker A: Well, yes, so it won't be in person, but I am an organizer of the ISPSM conference, which is International Society for the Philosophy of the Sciences of the Mind. It's a terrible name, but was the only one we found.
The call for paper for our annual conference in early November is still open, so if people want to submit, they will still be able to, and we are probably going to extend it.
I'm organizing that, but that's an entirely online event, so thanks again, Marco.
[01:28:33] Speaker B: Keep up the good work and hopefully we'll meet in person someday.
[01:28:37] Speaker A: Oh yes. Thank you for the invitation.
[01:28:48] Speaker B: Brain Inspired is powered by the Transmitter, an online publication that aims to deliver useful information, insights and tools to build bridges across neuroscience and advance research. Visit thetransmitter.org to explore the latest neuroscience news and perspectives written by journalists and scientists. If you value Brain Inspired, support it through Patreon. To access full length episodes, join our Discord community and even influence who I invite to the podcast. Go to BrainInspired Co to learn more. The music you hear is a little slow, jazzy blues performed by my friend Kyle Donovan. Thank you for your support. See you next time,
[01:29:29] Speaker A: Sam.
