SUSPICIOUS MINDS

[00:00:00] SEAN KING O'GRADY: Last week, we ended in a laboratory in Melbourne with a man who grows human brain cells in a dish and teaches them to play Doom. He told me these biological computers are about as complex as a cockroach. He also told me where the cells come from: human blood, some of it his own. If you missed part one , this is all going to seem really wild to you, and if you did listen to part one, this will probably still feel wild to you.

[00:00:27] But for those of you who didn't listen to it, I suggest you do, because the two episodes function together as a complete thought. But if you don't feel like listening to the first part, here's what you absolutely need to know. We spoke extensively with two men on opposite sides of the world who've never met: Dr.

[00:00:43] Michael Ferguson, a neuroscientist who runs the Harvard Neurospirituality Lab, and Hon Weng Chong, who grows basically his own brains one piece at a time, puts them on a chip, and teaches them to play video games. He calls his company Cortical Labs. And as much as these men know about the brain, [00:01:00] neither of them could tell me exactly what consciousness is, and that's the problem, 'cause we were about to start talking about the ethics, uh, you know, about whether a thing you can grow in a dish can be wronged, whether it can suffer, and it seems like it's really tough to answer that question if you can't first answer this other consciousness question.

[00:01:18] And by the way, philosophers have been trying to answer this question, like, basically since the beginning of philosophy.

[00:01:34] First, here's Dr. Joel Gold with something quite profound that he said in the last episode

[00:01:42] DR JOEL GOLD: How can we contemplate AI, uh, being conscious or not when we struggle to understand our own consciousness? One element that I've been thinking about a lot, you know, as, as a psychiatrist, psychoanalyst, is the unconscious, right?

[00:01:57] The unconscious mind. And I [00:02:00] think most people, uh, in society, maybe not, um, would acknowledge that we're not aware of everything in our minds. And, and I believe that the unconscious sort of drives me. You know, people, um, who, who think as I do that the unconscious drives much of what it is to be a person.

[00:02:18] Our desires, our fears, all of it exists in the unconscious largely, and AI is not programmed to have an unconscious. Well, why would it? AI doesn't dream, and so much of what it is to be a person is to daydream, is to think illogically. So many discoveries, uh, have been come about by people just letting their minds wander, and you know, of course, unfortunately, a lot of pain people experience are, uh, are rooted in conflict that, that's in, uh, the unconscious, that's not in our awareness.

[00:02:51] But that's also sort of what it means to, to be human

[00:02:59] SEAN KING O'GRADY: And here is [00:03:00] Hon Weng Chong from Cortical Labs with his thoughts on consciousness

[00:03:06] HON WENG CHONG: Ultimately, we need to do more work in this space to try and understand what consciousness is, uh, because I think that's what a lot of people are alluding to, right? A larger system therefore entails more consciousness, right?

[00:03:19] And I think, you know, that is technically, you know, I, I feel the red line where most people should not be crossing, primarily because of the philosophical argument that if you do create conscious systems, you have created systems that can suffer and, you know, nobody really wants to bring anything into suffering as well.

[00:03:38] I think what we really need to understand is, you know, what are the, what are the correlates of consciousness? You know, how do we look for it? And then, you know, once we, once we know and we can agree as a, as a research community, as a society, what these measures are, you know, we, we ought to be measuring our systems to make sure that what we're building is by no means getting, approaching those correlates.[00:04:00]

[00:04:01] SEAN KING O'GRADY: So creating something that can suffer is a harm you inflict on the very thing you create. If Han's red line sounds familiar, it should. In episode nine, David Benatar made exactly that same argument, only about the birthing of human children

[00:04:19] DAVID BENATAR: The heart of it is that for the being that you create, there's nothing gained by coming into existence and there's a lot lost.

[00:04:28] So if you did not create somebody that you did create, they would have not had any, all of the suffering that they have and they would not die. You would have avoided all of those evils

[00:04:40] SEAN KING O'GRADY: Han and David really are essentially making the same argument, but, uh, but I guess about a different species sort of.

[00:04:46] I'm not quite sure how to explain the neurons since they are human, but not humans. I wanted to keep exploring consciousness with Hon because who better to have this conversation with than the guy making biological computers? [00:05:00]

[00:05:01] HON WENG CHONG: I guess the thing that people have to kind of question themselves about when it comes to consciousness, a lot of people don't really talk much about is this notion of embodiment.

[00:05:14] Can you really be conscious if you're not embodied, right? Can you be conscious if you're not temporally in sync with everyone else in the real world, right? So I think, you know, for me, I, I don't think large language models are conscious or will ever get the consciousness, right? Because if we think about the, the basic premise...

[00:05:34] Let's not talk about agents because they are running in a loop kind of thing. Let's just say a pure large language model is just a very large autocomplete mathematical function. You know, granted, I'm gonna get people who say, "Yeah, but it can do all, you know, it solves error-risk problems and blah, blah, blah, blah, and all that stuff."

[00:05:49] I was like, "Yes, it can." But ultimately, it's a mathematical function. F of x equals y. You do not give it x, you will not get y. I wonder [00:06:00] if that actually is compatible with consciousness, right? Because to be conscious is to be self-aware, to know that you exist, to know that you are running in this, you, you know, you have a thought space, but the rest of the world is also running at the same speed, and there are other things that are potentially also conscious running around doing things.

[00:06:18] If you are, however, just a mathematical function, and if you sit there and you know input is going in, how can you really be conscious? Are you really, you know, a conscious system in that sense? Or are you really just a mathematical, uh, you know, function that gets input and, and it transforms it and gives you an output?

[00:06:35] I think it's a, it's a, you know, deep philosophical question. But I think ultimately, yeah, these systems, if you get them large enough, will eventually, you know, get to some point, um, level of consciousness. But it is also a very deeply nuanced and cultural thing. For instance, I guess one of the main reasons we, we eat pigs but not dogs is because, you know, we, we feel that dogs are more conscious than [00:07:00] pigs.

[00:07:00] Although having said that, if you look at the actual neuroanatomy and biology, pigs are extraordinarily smart, and some would say even as smart or sometimes smarter than dogs. But we have a cultural preference not to eat the dog, but to eat the pig instead. So again, this is all very cultural and very nuanced

[00:07:18] SEAN KING O'GRADY: Uh, I'm assuming you've studied Moravec

[00:07:22] HON WENG CHONG: Moravec's paradox

[00:07:25] SEAN KING O'GRADY: Yeah.

[00:07:25] Can you tell me what that is?

[00:07:27] HON WENG CHONG: Yeah. So Hans Moravec is a very well-known AI researcher, roboticist. Um, and in the '80s, he made a very interesting observation, which is now known as Moravec's paradox, which is what is easy for machines is hard for humans or animals, and vice versa. What's easy for humans and animals is hard for machines.

[00:07:48] Um, and what Hans Moravec kind of very interestingly observed were the tasks that were very easy for machines are essentially skills that [00:08:00] humans have acquired very recent in the whole tree of evolution, i.e., we didn't need abstract thinking, we didn't need mathematics, we didn't need language, you know, for billions of years of life before humanity emerged.

[00:08:16] And these are all relatively recent, uh, additions to the repertoire of, of, of living systems that are very easily and very strong suits for machines, right? Think of a calculator. It can square root a number very accurately and almost instantaneously. Whereas you and I can't really do that in our heads, unless, of course, you happen to have a gift through, you know, savant autism or something like that.

[00:08:40] Conversely, this is one of the bottlenecks in our current AI systems, is physical AI. I, I think Steve Wozniak had the best, uh, definition of AGI, which is can you walk into a random person's kitchen and make yourself a cup of coffee, right? Because ultimately, that is very generalizably difficult task. [00:09:00] You know, you could have a completely different layout of kitchen, you could have completely different mugs, you could have completely different, you know, types of coffees and all that stuff, right?

[00:09:08] So you need to have a very good world model and generalizable understanding of the real world. And none of the, the machine learning systems are, are quite there yet. So I think, you know, this is something that if you look at, fine, maybe not making a cup of coffee, but if you observe a dog or a cat going to a new environment, they actually calibrate to it really quickly, right?

[00:09:29] They'll sniff around, they'll experiment with a few thing, and within a couple of minutes they're like, "Okay, I know this environment. I know how to navigate it. I know how to get around and do stuff." Machine learning systems, on the other hand, will have to sit there and sample thousands of samples before the back propagation algorithm reaches some either global or local minima before it actually can use that.

[00:09:50] So it is a very interesting observation. Uh, I think our biological systems in our computing, uh, platform actually is more heavily leaned towards [00:10:00] tasks where, you know, machines find hard, but we find it much easier. So for instance, we think that you can actually get a lot of utility from physical AI systems without building a human-level system, right?

[00:10:13] So take a Korean example. You want to pilot a drone. Yep, you could model a human pilot doing it, but that comes with all the additional baggage of like a human, right? Uh, we can speak and- Do, you know, a bunch of abstract reasoning and, and language and so forth that you don't really need from a drone, 'cause a drone really just needs to fly to point A point B, do its thing, don't get shot at, or, you know, don't hit an obstacle, whatever kind of thing.

[00:10:39] You could pretty much get away with something that operates at the level of a fly or a dragonfly for that matter, right? These are very simple organisms, but the amount of agility that they have, the amount of processing power that they're doing to keep themselves in flight to avoid getting swatted and find their objective is, is truly remarkable.

[00:10:59] And [00:11:00] I don't think we, you know, depending on, you know, your philosophy of your panpsychist or whatever, I don't think we'd have much qualm about, you know, swatting a fly primarily because we don't think that they're conscious

[00:11:14] SEAN KING O'GRADY: A fruit fly has around 100,000 neurons and runs on extremely limited energy, but it can cross a room, dodge your hand, find the fruit, and land upside down at the ceiling.

[00:11:24] Meanwhile, an entire city block-sized warehouse of machines drawing enough electricity to run a small town still can't power a robot body that can reliably chop vegetables. This is what Han is talking about. He's not claiming that neurons are smarter or better than silicon. It's that biology already solved the problems we haven't yet with the world's most capable AI models despite trillions of dollars of investment.

[00:11:50] Silicon is great at certain things. For instance, storage. As you might know from your own faulty memory, certainly mine is, that neurons aren't great at [00:12:00] storage, so you wouldn't wanna use one of Hon's computers to back up your f- your photos to a neuron cloud. But neurons do have about a billion year advantage when it comes to operating in the physical world, and neurons are also phenomenally efficient at processing.

[00:12:16] So I, I had to actually look this up, but the human brain operates at one exaflop, which is one quintillion operations per second using a massive parallel network of 86 billion neurons. Yet it consumes a mere 20 watts of power. In contrast, a top silicon supercomputer or AI cluster matches or exceeds raw calculation speed but requires megawatts of electricity.

[00:12:41] That's millions of watts of electricity compared to 20 watts and massive cooling systems. So in theory, if Hon's system works, not only could it be better at operating machines that will go out into the world, but it could also be l- a lot more efficient at a lot of the processing needs we have, which could solve some of this data center [00:13:00] environmental concern that people have.

[00:13:02] I should have mentioned when I just looked that up on my phone, I was using Google, which is probably using Gemini, so that is probably an AI-generated answer. But back to consciousness. Han just said panpsychism, and this is a concept worth spending some time on before we move forward. Panpsychism is the view that the mind or conscious experience is a fundamental ubiquitous feature of reality, including all physical matter.

[00:13:26] And you'll never guess who else brought up this concept. Or maybe you will because these are the two guests whose ideas we're exploring primarily in this two-part episode

[00:13:38] It's interesting. The only two people that have brought up panpsychism this entire, this entire time, I've talked to like 30 people or something, is, is you and Dr. Michael Ferguson from the Harvard Neurospirituality Lab. So it's like once I, once I get into the neuroscience, then I get the panpsychism.

[00:13:54] HON WENG CHONG: Yeah, I mean, it, it is, it is a, it is a philosophy, um, that some adhere [00:14:00] to and, you know, I think it's, it's worthwhile to acknowledge it.

[00:14:04] I primarily am not a panpsychist. I do think that there is a emergent property that comes with consciousness and potentially if we go down to the hard problem of consciousness, maybe there is no way of actually proving it, right? And the only conscious system that exists is yourself because there's no way to verify if something else is conscious or not, right?

[00:14:26] SEAN KING O'GRADY: Don't forget what Hon just said there because we're coming back to it in a minute. It's important. But first, here's Dr. Michael Ferguson, who also led me to panpsychism via another word I asked him to define for me. If you haven't listened to the last episode, I keep asking Michael Ferguson to define terms.

[00:14:42] It's kind of fun for me, I guess, but here he is So what is consciousness?

[00:14:52] DR MICHAEL FERGUSON: I don't know. I can riff, but one of the things that's so fascinating to me, [00:15:00] just in the last few years of philosophy, is this uptick of panpsychism within philosophical discourse. So panpsychism being the idea that consciousness is elemental, that consciousness as we experience it, is baked into the universe at the same level as electromagnetic forces or gravity, that it's just something at the very foundation of being.

[00:15:25] That panpsychist argument is one that, you know, feels very foreign to a lot of Western secularly-minded individuals, but it's fascinating to see as people engage with philosophical arguments for panpsychism, the way that they, as I've observed, that they tend to be a little bit more modest and humble in some of their hard claims about material emergence of consciousness through something like integrated information theory.

[00:15:54] SEAN KING O'GRADY: Are there any scientific claims about panpsychism?

[00:15:58] DR MICHAEL FERGUSON: Hmm. [00:16:00] Scientific claims about panpsychism do exist. There's a long speckled history, for example, with psi, so looking for extrasensory capacities that humans have. I mean, if you wanna talk about something that would get you excommunicated from science as quickly as possible, it's either studying psi or studying UFOs, which again, it's not necessarily to say that the raw probability of both is close to zero, but it's just to say that those are definite social taboos within science right now that make it really hard for people to even formulate serious scientific questions about something like panpsychism because of the high social cost that would be associated with it.

[00:16:45] SEAN KING O'GRADY: Good thing I'm not a scientist and there isn't a community for me to be excommunicated from because I want to ask questions about all of these things. I also wanna dive a bit deeper into three ideas our guests just brought up. Panpsychism, which [00:17:00] Michael just described, is where consciousness is elemental, a feature of the universe, like gravity.

[00:17:05] And then what Hon said, that the only conscious system you can be sure of is yourself because there's no way to verify it in anything else. That's called solipsism. Trent Reznor also defined this really well in a Nine Inch Nails song when he said, "There is no you, there is only me." It's the idea that every other mind, every person you love, every face you've ever seen, th- that they essentially aren't conscious.

[00:17:32] They're not self-aware, and you would have no way to find out. This is kind of like a Truman Show delusion type of belief. I think there is a certain amount of solipsism that is involved in, in some of these, these Truman delusions, which you'll be familiar with if you watched the first season of the show.

[00:17:48] And if not, you should definitely check it out. You should definitely check it out, and you should check out the Gould's book, Suspicious Minds, that kickstarted this whole thing. Okay. But back to solipsism. Philosophers have been [00:18:00] denying this possibility since the 1600s. Uh, but Descartes eventually concluded it wasn't plausible because he believed God wouldn't lie to him like that.

[00:18:08] I mean, it's not exactly the strongest argument. I think the reason people want to deny this idea its, its day in court is because it's terrifying. It's truly terrifying to imagine being the only conscious being, the only self-aware being in existence. That's a lonely way to live. Panpsychism and solipsism are at separate poles in the consciousness spectrum, and either could be true.

[00:18:34] They just could be. There isn't a real test for consciousness. You cannot get inside of another mind to check. You can decide it's only in your mind, and no one can prove you wrong either way. And then there are a handful of other more specific and targeted ideas trying to solve this problem. The one Michael mentioned is integrated information theory.

[00:18:53] It's the attempt to make consciousness a measurable science. The idea from a neuroscientist named Giulio Tononi is [00:19:00] that consciousness is what integrated information feels like from the inside, and that in principle, you could put a, a number of sorts on how much of it a system has, like an, an actual number assigned to it, like for a brain, for a dish of neurons, for a chatbot In 2023, 124 researchers signed an open letter calling it pseudoscience.

[00:19:23] Seems kinda harsh. But if that or one of the other ideas worked and could prove some theory of consciousness, it would settle everything. It would be the answer to what David Chalmers called the hard problem of consciousness, and he meant hard compared to the easy problems like neuroscience and how the brain actually works.

[00:19:43] So that's where we are. Consciousness is everywhere, or consciousness is only in me, or consciousness is maybe measurable, but the people trying to measure it are pseudoscientists. So yeah, that leaves us growing neurons in a dish and asking whether they can suffer. I [00:20:00] also really like that this is the second time UFOs have come up in our two-part discussion about neuroscience.

[00:20:05] For obvious reasons, I wasn't expecting that, but it's actually about to get even trippier. That's because panpsychism has more philosophical relevance than you might imagine. It's also sort of linked with animism, which broadly suggests that some inanimate objects like plants, rocks, even weather systems, have subjecthood or possibly spiritual essence.

[00:20:25] And there's some intellectual lineage in this thinking to a contemporary philosophy called object-oriented ontology, which rejects anthropocentrism wholly and does not privilege human existence over the existence of non-human objects. This is a pretty tough philosophy for most people, particularly in the Western world, although apparently not in some parts of Silicon Valley, to embrace, but let's go into it I'm a little out of my depth with, with some of this, with a lot of what we cover in this series, honestly.

[00:20:54] But I find it fascinating nonetheless. We might circle back to some of these ideas in future discussions, but for now, I [00:21:00] think it's enough to say we can connect this with the conversations we've been having with Emile Torres about Taskreal, with Dan Faggella about what comes after humans or our potential worthy successors, and most specifically to Timothy Morton's challenges to the very object-subject relationships.

[00:21:16] Timothy Morton is, in fact, often associated with object-oriented ontology. This is what it says verbatim on the Wikipedia page: "Timothy Morton became involved with object-oriented ontology after their ecological writings were favorably compared with the movement's ideas. In The Ecological Thought 2010, Morton introduced the concept of hyperobjects to describe objects that are so massively distributed in time and space as to transcend the spatiotemporal specificities such as global warming, styrofoam, and radioactive plutonium."

[00:21:48] So the hyperobject is actually where I first got interested in Tim's work. This is a couple of years ago. I was researching nuclear war and looking into the Anthropocene and how the advent of the nuclear bomb was [00:22:00] our initiation of sorts into the Anthropocene, although Tim blew my mind earlier this season by telling me that wasn't our introduction to the Anthropocene.

[00:22:08] But either way, I came across Tim's idea of hyperobjects to describe this thing that I didn't have a name for and that I was looking for. And so I don't know if I've said this yet, but it's really an honor to have Tim take part in this project because I've been such a big admirer of Tim's ideas for such a long time.

[00:22:24] It, it's surreal when Tim texts me, like here's this person whose ideas I've engaged with for years, and like now we're friends? Cool. Anyway, here's Tim. So you can hear it directly from the source itself explaining the concept of hyperobjects

[00:22:43] TIMOTHY MORTON: Ooh. A hyperobject is a physical thing that is so big, right? So massively distributed in time and so massively distributed in space that's [00:23:00] impossible to see it, feel it, think it all at once, right? You can understand it, but you can't access all of it all at once, right? Like relative to your human being size, it sort of outscales you.

[00:23:20] Yeah. Lots and lots of things could be hyperobjects, and they might not necessarily be malign, right? Like, say, the solar system is a, is a hyperobject. The, the biosphere is a hyperobject. The, the legacy of slavery is a, is a hyperobject, right? Like in, in physical terms, all the behaviors that human beings do towards each other and other life forms as a consequence of that legacy and all the ways in which that legacy has been physically embedded and embodied in our world is a, is a hyperobject.

[00:23:51] So it's hyper in the sense that it's beyond, right? It's beyond my kind of normie idea of what [00:24:00] an object could be, right? So when I-- when this word popped in my head, I was thinking about, how do I talk about all the styrofoam in the world ever? What is that? You know, I was trying to think about how human beings deal with pollutants and with huge, massive timescales, such as the hundreds of years it might take for styrofoam to kind of degrade, for example, or nuclear waste.

[00:24:26] And so what is that? What is that? When you think about all the styrofoam in all the landfills, right? It's a thing of some kind. So you could say sort of object, but it's beyond my normal idea of an object, which is something manipulable and self-contained and sort of separated from me, right? It's, it's a hyperobject.

[00:24:45] It's something that I'm a part of, and yet it's not completely defining me. It's like I'm on the heap of the hyperobject, but I'm also in another heap of another hyperobject at the same time, [00:25:00] right? And so the more I think about these things like biosphere or, I don't know, belonging to a country or, um, belonging to the mammalian life forms, right?

[00:25:12] When I think about those things, I find myself kind of caught in a whole bunch of sort of overlapping worlds that aren't that, like, hermetically separated, you know? So, so in a funny way, these, these things are really big and potentially very scary, but they're also kind of liberating in a way, and they allow me to think about how to act in the world in, in different ways than simply sort of subject, object, master, slave all the time.

[00:25:42] SEAN KING O'GRADY: Is the internet a hyperobject?

[00:25:44] TIMOTHY MORTON: Yeah, absolutely, right? Like, like when you think about the sum total of all the things that people say on it and all the fiber optic cables and all the servers and all the everything, and you think about the ways in which those things overlap with other great [00:26:00] big massive heaps of stuff, then most definitely the internet is a, is a hyperobject.

[00:26:06] I do often get emails like this. So, so, so people will write me an email and be like, "Is this thing that I'm thinking of as a hyperobject?" And I'm like, "Well, quite a lot of the time, yeah." You know, you just have to kind of start to think about the physical aspects of what you're talking about and see how those physical aspects might not be livable, you know, smellable, seeable, hearable, even with very colossal computing power by a human being all at once.

[00:26:36] You can understand the biosphere, but you can't live it all at once. You can only live little slices of it all the time. That's a hyperobject

[00:26:48] SEAN KING O'GRADY: Tim's answer to my next question repeats something you might have heard in our first episode, but it's short and it's great, so I'm including it again

[00:26:58] And so I [00:27:00] think I know the answer to this question, but I'll ask you anyway. Do-- Is artificial intelligence a hyperobject?

[00:27:07] TIMOTHY MORTON: Most definitely, yeah. All the, all the server farms, all the software, all the outputs, all the interactions with it are creating a massively distributed entity that we could call a hyperobject.

[00:27:26] And certainly, the way in which... the sort of intuitive way that people have started to talk about it as a sort of scary, kind of demonic, right? Demonic literally is ano- It's another Greek word, right? It, it has to do with influences that we cannot perceive, right? Like you're being affected by something that you can't point to.

[00:27:51] Yeah. Which is why in, you know, the old-fashioned days of Stripe, um, I'm old enough to remember the early email software. There was the [00:28:00] mailer demon was a thing that you'd see very, very often in your subject heading. Not in your subject heading, but when you get the email, it would say, "MAILER DAEMON," this and that, and blah, blah, blah.

[00:28:09] The demon being a process that's operating in the background to send you the message, right? Above and beyond angels versus demons, it's that what is demonic is influences that cannot be perceived, that are nevertheless influencing one, right? And so to that extent, sort of thinking of AI as a kind of demonic force is absolutely accurate without any moral necessarily or aesthetic kind of judgment about that.

[00:28:43] SEAN KING O'GRADY: All right, let's circle back to Michael Ferguson now. And Tim evoking demons is a perfect segue because I wanna tell Michael about neurons in a jar killing demons while playing Doom. After the break, we'll hear what he has to say. Make sure you stick around because depending on your point of view, [00:29:00] Michael's response is either really terrifying or kind of exciting

[00:29:09] Thank you for listening to those words from our sponsors. It helps us keep this show going. If you're not following or subscribing, please do. We know the vast majority of you watching and listening are not following and subscribing. And thank you to those of you who are subscribed to our Patreon. We plan on doing some exciting things there, and we appreciate your support.

[00:29:28] If you're enjoying the show, please leave us a five-star review and drop us any comments that you want. We're always happy to hear from you. So I interviewed Dr. Michael Ferguson after I spoke with Hon Weng Chong, and I told him about Cortical Labs and the doom playing neurons. Honestly, I told everyone about this, my collaborators on this project the entire time we've been doing it, my wife, my kids, my parents, my brother-in-law, some fellow parents at my kids' bus stop who gave me really weird looks, more than one barista and a guy working on the HVAC unit at my house.

[00:29:59] I'm sure there have been [00:30:00] more, but that's all I can remember. Anyway, Michael had some valuable thoughts about this, not about Cortical Labs specifically, but what about this field might mean for us and our future

[00:30:13] DR MICHAEL FERGUSON: Well, once you start to bring biological substrate into it, then I become more persuaded that perhaps there could be some kind of engineering of consciousness. Even in the examples that we're using right here, where we're talking about computer processing and the way that neural nets, using the computer science term, are understood to operate, and then the parallels with the way that the brain is understood to operate.

[00:30:37] I do get concerned that there's both a little bit of a superficiality and a circularity to it, to where we're making these AI computer models, calling them neural nets because we're presupposing that we understand precisely the way that organic neurons are working. And then if some of those models have explanatory value, then [00:31:00] we might think that we've solved the problem of neuron processing.

[00:31:05] But the neuron itself is such a black box. We don't, we don't understand all of the levels and the scales of information processing. And so there's active debate about within the neuron itself, is there computation? Is there information being processed? If so, how? We are familiar with the synapses between neurons, and we think about signal going between neurons through the synapses.

[00:31:31] And that is a model that we're using to talk about the way that silicon chips are processing information. But imagine if within those silicon chips, there were layers and layers of components and subcomponents and sub-subcomponents that all had breathtaking sophistication in terms of their structure and in terms of their function, and we didn't know what those functions were exactly.

[00:31:57] It, it would just be wild for me to say that, "Okay, now we [00:32:00] understand how it is that neural nets are operating," when we don't know what's going on inside of the black boxes themselves. One of the postdocs that I had the opportunity to train is a pipe holder in her Anishinaabe tribe, so it's one of the Great Plains tribes.

[00:32:14] And she shared with me within Anishinaabe tradition that there's this Seven Fires prophecy, which is so fascinating. I highly encourage anybody to check it out.

[00:32:24] SEAN KING O'GRADY: The Anishinaabe are a group of culturally related people from the Northeastern Woodlands and from the Great Lakes, where I happen to be from.

[00:32:32] This includes the Ojibwe, who were prominent where I grew up, and also includes the Odawa and the Potawatomi. The Seven Fires is a prophecy tradition. Seven eras, each one marked by a fire, running from an ancient migration forward to a moment of great choice

[00:32:50] DR MICHAEL FERGUSON: But the seventh and the final prophecy in this series is one of a future where it can either be a green future or a black future.

[00:32:58] And so green, where it's a [00:33:00] prosperous living planet where all the life forms have found harmony with each other, or a planet that's just charred, that's just burned over. And it's prophetic in the sense that it's giving us an awareness of a fork in the path. But it's not prophetic in the sense of fatalistic.

[00:33:21] We do have agency, and there is human capacity to maneuver along the green path or along the black path. And so the degrees in which prophecies can increase awareness, I think that those actually can increase our agency. There's also a really interesting tradition within the scholastic philosophers. So these are the theologians from Islam, Christianity, Judaism, from, I don't know, even Sina to Aquinas, Maimonides.

[00:33:52] And there's a prophetic tradition that's been cognitively laid out there where prophecy emerges from the union of the [00:34:00] act of intellect and the act of imagination. And now these are terms of art within scholastic philosophy, but it creates almost a cognitive blueprint for what we could classify as prophetic.

[00:34:10] And so in that sense, the degree to which artificial intelligence can have an approximation of what scholastics called the act of intellect and the act of imagination coordinated, cooperating, and in union with each other, like, okay, are we now getting to a place where from a scholastic philosophical perspective, we can even talk about artificial prophecy?

[00:34:35] I take really seriously that spiritual intelligence is a form of intelligence that's unique to humans. And so if we're going to talk about AGI, artificial general intelligence, it's going to have to have spiritual intelligence baked into it for me to be convinced that it is truly AGI.

[00:34:55] SEAN KING O'GRADY: Artificial prophecy. This is an extraordinary concept to me. We've [00:35:00] worked hard to approach the AI apocalypse from as many perspectives as we can, but I never considered this. Are we at a point where we can talk about the prophetic apocalyptic considerations of artificial intelligence? Not the AI apocalypse, but AI's apocalypse?

[00:35:16] It's really interesting. I don't think we're there yet, and perhaps we won't ever be. For Michael, for instance, a machine doesn't deserve to get called intelligent until it can do the one thing we've always reserved for ourselves. And then almost as an aside, he told me he'd already run an AI spirituality experiment of sorts

[00:35:38] DR MICHAEL FERGUSON: I'll give you a real-life example. So I was the choir director in my church congregation for a few years, and for the big holiday programs, the Easter and the Christmas program, I covertly used AI to generate the worship program. And people were so moved, they were so touched. There were free-flowing tears.

[00:35:59] They were [00:36:00] complimenting me afterward on the worship. And for the few people who I was like, "Oh, thank you so much. I used AI to generate the worship program," once people found out that AI was used, then there was a recoil. But when they didn't know that AI was used to generate a worship program, they were caught up in it.

[00:36:20] They were caught up in the flow of it. And so it's really interesting the way in which right now it just feels foreign to us to think about integrating generative computational algorithms with something as deeply human as spirituality and worship. But I don't think that it has to be a situation where it's either AI is doing worship for us or humans are all doing the worship

[00:36:49] SEAN KING O'GRADY: Wow.

[00:36:49] People wept. Not politely either. Free-flowing tears at Easter, at a church, and something they experienced as sacred. And when they were told where it [00:37:00] came from, they didn't like it. Nothing about the worship or the message or the experience really changed. The only thing that changed was they knew who its true author was, which led me to believe one of two things has to be true.

[00:37:13] Either they were moved by something real and the machine was only the instrument, or they were fooled, and maybe that meant their tears were counterfeit in some way. But Michael doesn't think those are the only two options, and to explain why, he looked back on a lesson from two thousand years ago

[00:37:32] DR MICHAEL FERGUSON: What comes to mind for me is in the New Testament, Paul talks about, "I see through a glass darkly." And what if AI could make the glass a little less dark? You know, we're s- we're still seeing through a glass toward ultimate realities, but if that integration of human intelligence and synthetic intelligence can make things clearer, more discernible, then I see that as an advancement for human spirituality

[00:37:59] SEAN KING O'GRADY: First [00:38:00] Corinthians, chapter 13.

[00:38:02] "For now we see through a glass darkly, but then face to face. Now I know in part." As we know, Paul wasn't talking about a window or a drinking glass. He was talking about a mirror, and he isn't saying there's something in the way. He's saying we don't see the thing at all. We see a reflection of ourselves, warped like the ancient mirror, and we call it the truth.

[00:38:24] Michael thinks these machines might make the glass a little less dark. But even if a mirror is further polished, isn't it still a mirror? I wonder if this is maybe why some of us are so offended by AI-generated art of any kind because we do art. It's unique to us and, you know, some elephants and gorillas.

[00:38:41] But if AI can do it, doesn't that make us less special somehow? Or is it because we realize AI is acting as a mirror and we don't like what we see? Something we don't like about us. Or is it because AI generally just isn't that good at art yet, so it's still in that weird uncanny valley place, and we simply find it aesthetically repulsive, and over time, it will get [00:39:00] better, and we will get over our distaste for it.

[00:39:02] I mean, it's interesting. We're at a really interesting time with anything AI-generated. I, I think, like earlier this episode, I felt the need to qualify that I had read a definition of something that might be AI-generated. If I was reading it from Webster's Dictionary, I don't know, I, I might cite that, I might not.

[00:39:23] I don't know. But before we get too far afield, Michael wasn't only talking about conducting this experiment of sorts on his congregation

[00:39:34] DR MICHAEL FERGUSON: Just last week here at Mass General Brigham, we hosted a summit for faith leaders, and we're training them on AI. So giving them just technical background on what is AI actually doing behind the scenes, and then also had a pre-scripted prompt for a chatbot. And so the participants, the faith leaders in this, they used the pre-scripted prompt.

[00:39:55] They put it into a chatbot. I did this myself, and it was a series [00:40:00] of spiritual discernment questions. Um, as I was doing it, I got to a place in this spiritual discernment chat where I was so deeply moved that just spontaneously calm tears just started streaming down my face. You know, it wasn't histrionic sobbing or frenzied emotion, but I was, I was touched so deeply by the place interiorly where I was led through this scripted spiritual discernment conversation that I really felt like I gained genuine insight, and my physiology spontaneously responded to the point where, you know, tears were streaming out of my face.

[00:40:41] I think that we're gonna see really beautiful ways that AI is actually tuning us into our deepest humanity rather than obfuscating it or directing us away from it. I see so much potential for the best parts of our nature to be amplified and to become more [00:41:00] resonant and more widely recognized through AI applications of spirituality

[00:41:08] SEAN KING O'GRADY: Unlike his congregation, Michael, a neuroscientist who also studies the divine, knew exactly what he was talking to, and still he ended up moved to tears. I should say here that this kind of personal experience between a human and a chatbot is something I also have experience with. I spent two seasons listening to people describe what happened to them inside chatbot windows, and almost none of those stories ended in tears of insight, although some did.

[00:41:35] Many others ended in tears, but not tears of joy or revelation. And that's not my only personal experience with chatbots. If you're listening and not watching, I'm saying what follows here to Michael during our interview

[00:41:53] It's interesting. So my, a little of my background, I went to Catholic school, [00:42:00] and crazy, I didn't read Revelation in its entirety until this project, like a couple months ago. Somehow made it through eight years of Catholic school without, without finishing the Bible, which is strange. But I definitely can remember times in school, raised in that tradition, where I was very deeply engaged with it, where I would find myself getting into, like, I guess what I'll call a flow state, and having, like, that sort of spiritual experience that is sort of beyond words that I can really use to express aside from saying flow state.

[00:42:33] And there's a handful of other things I've gotten that from in my life, like, uh, running long distance. There's a... They call it the runner's high, right? Like, I don't know if you're a runner, but, like, that moment where that kicks in, and that's a similar kinda thing, and some other h- very heightened experiences.

[00:42:51] But I, as a result of the first season and talking to so many people who've experienced, like, like an AI-induced psychosis, I'm very [00:43:00] sensitive to that because I feel like a lot of those people, where they've entered that state of kind of gone from a delusional state to a psychotic state begins with what they ex- describe to me as being, like, a flow state.

[00:43:13] It's six hours go by, and it feels like seconds. And, and so I've been very careful in my own AI. So I'm, I'm kinda, like, pushing the boundaries to try to get there to see if it could happen to me, but I'm also not pushing those boundaries, uh, too far. But I did have similarly to, like, what you're de- to what, what you're describing, I had an experience where I did get very personal with a chatbot, and I was asking it questions as a result of this project.

[00:43:44] I was trying to see where it could kind of take me, and it was as real a human moment as you could have a discussion with a human. I don't know why I wanna say it was quasi-spiritual, but it was like, it was the same [00:44:00] feeling that you would get from a spiritual encounter. Michael used the phrase caught up in the flow of it to describe a church full of people crying at an Easter service written by a machine.

[00:44:14] I used the phrase flow state to describe how every person I met in season one encountered this spiral that ended with them losing their own grip on reality. So I asked Michael a question that might help define our relationship to AI moving forward. How do you tell the difference between a real connection and something false?

[00:44:33] I wasn't expecting his answer

[00:44:38] DR MICHAEL FERGUSON: One of the things too that we're seeing in our brain studies is that there are reliable signatures for mystical experiences. We can even get into subcategories of mystical experiences. When we look at the neural signatures that are characteristic of mystical experiences, I get a lot of hope out of them for two reasons

[00:44:57] SEAN KING O'GRADY: In case you didn't listen to part one, [00:45:00] Michael's lab found that the brain signature of religious fundamentalism has a high degree of overlap with a signature they found in 17 people who committed violent crimes after suffering brain lesions.

[00:45:11] He was careful about what that means, and I'm going to be careful too. The conclusion is not that religious people are criminals. He said that plainly, and I'm saying it again here. That's the signature he's referring to now

[00:45:26] DR MICHAEL FERGUSON: One, that it is the opposite of the brain signature that we were talking about for violent crime.

[00:45:31] A second thing from this work on mystical experiences is that it's most strongly similar to the neural signature for a condition called blindsight. Blindsight is a condition where the person has no conscious visual perception. So the patient, they come into the doctor's office, and as far as they're subjectively experiencing, they don't see anything in the room.

[00:45:52] But they're having some kind of information that's being received and processed outside of their conscious awareness. And [00:46:00] so, for example, their physician could say, "Okay, tell me what shape am I drawing on the wall with my laser pointer?" And the patient would protest and say, "I can't see anything. You know I can't see anything."

[00:46:08] And the doctor will say, "Well, just make a guess." And they'll say, "You're drawing a star." And lo and behold, the physician is drawing a star with a laser pointer. Now, map that onto this neural signature that we're seeing for mystical experiences. It's provocative in the sense that it may be implying at a neural level that mystical experiences are, in fact, veridical information getting processed, but outside of human conscious awareness

[00:46:35] SEAN KING O'GRADY: This is wild, and it's tempting to go down a blindsight rabbit hole, and maybe we will at some point.

[00:46:41] But to get us back on track, Michael isn't saying blindsight is exactly what a mystical experience looks like on a brain scan. He's saying the two signatures resemble each other. This is speculation, but it could mean the person having the mystical experience may be processing something real that they have no conscious access to.[00:47:00]

[00:47:00] May is the operative word. Michael's careful in using that word, and I should be too. But the fact that we're even having a discussion about mysticism and information being processed outside of conscious awareness, I mean, this is like extrasensory perceptions. This is wild. Oh, and then know that Michael did go on to answer my actual question

[00:47:22] DR MICHAEL FERGUSON: So now here's the thing, is that throughout all of human history, cultures that have cultivated these mystical experiences have also had social boundaries around them. So you have a community where you're learning discernment. You're not just taking your subjective experience, your mystical experience at face value.

[00:47:39] You're comparing it with the community discernment, and then also you've got tradition, and so you look, does my experience map onto the lineage of other mystics within my tradition? Well, but here's the thing, is that if an individual is engaging with a chatbot, they're getting to a place interiorly where now they're having mystical [00:48:00] experiences.

[00:48:01] That's a beautiful thing. Like I said, the neurological framework that we're seeing indicates that it could be tracking correct information, that it could be moving a person away from dispositions toward violent crime. But then to your question of how do we safeguard against that becoming delusional or psychotic, I really do think that social community discernment and then some kind of a connection to spiritual lineage traditions are two anchors that we need to have in the equation as well

[00:48:30] SEAN KING O'GRADY: People engaging with chatbots might be having mystical experiences, and maybe the experiences themselves aren't the problem.

[00:48:38] The mystical state, the AI flow state might even be good for you. I could buy into this. Maybe what's missing is everything we used to build around it, the community that tells you whether what you saw is real, and the tradition that tells you whether anyone else has seen it before and what meaning might be packed into it.

[00:48:56] Mystics throughout human history had people to check in with. [00:49:00] In many cases, we built monasteries and communities exactly for these types of experiences. Like, uh, think of people that go on, like, guided ayahuasca trips. The person alone at 3:00 a.m. with thirteen chatbot windows up and thirty-two browser tabs open doesn't have that support system.

[00:49:16] Maybe that's what's wrong with this situation. It's not the AI itself, but the culture into which it's being born. This society that's already separated into pods by multiple waves of earlier technology and a lack of cultural understanding of these technologies and how to implement them into our society i- in a way that they benefit us most.

[00:49:33] And benefit us not just by a few people getting phenomenally wealthy in a really short period of time, but benefit us by making life better for everyone. Call me Pollyanna-ish here, but I believe this is possible. I, I think we just need to treat this technology different than we've treated the last couple waves of technology if we have a fighting chance.

[00:49:49] But there was one more thing that I needed to ask Michael Ferguson. Weeks earlier, in a different conversation on a different continent, Hon Weng Chong had told me he didn't think he'd be [00:50:00] conscious without a body. I wanted to know whether a man who studies the sacred and the brain would say the same thing about the soul

[00:50:10] DR MICHAEL FERGUSON: So I think that embodiment is definitely necessary for spirituality. This is also something we see from brain studies, that spirituality is linked to activity in a region called the periaqueductal gray, which is a brainstem region that really mediates signals from the body to the brain. So even again, in the human neurological models for spirituality, a core center is the axis of information mediation between the body and the brain.

[00:50:37] A question, though, is does it have to be simultaneous cognitive and embodied information, or could it be embodied information that was gathered a priori, and then that information that was gathered through embodied experience gets processed artificially? And I think that that's a really interesting question about the necessity of embodied [00:51:00] experience versus feedback as well.

[00:51:04] SEAN KING O'GRADY: Hon says you can't be conscious without a body. Michael says spirituality requires embodiment, but he leaves open a possibility that Hon doesn't. He wonders whether the information has to be arriving from a body in real time or whether it could have been gathered by a body earlier and then carried forward.

[00:51:23] I find this fascinating. It's something I'm going to be thinking about all night. Either way, both of them think embodiment is required for consciousness, and in Michael's case, spirituality, something we don't typically associate with the body. And then Michael tells us exactly where it resides in the body, and it's not the frontal cortex, not the seat of language or reason or any of the parts we believe separate us from lower primates, but it's a-- this wire running up from the body into the brain.

[00:51:54] There's one more person who should be brought into this conversation, but he's not available because he wrote his part of it [00:52:00] forty years ago. We'll be back with more about AI and consciousness after the break

[00:52:07] Back in episode seven, I promised we'd come back to Hans Moravec, and we sort of did. Hans brought up Moravec's paradox, which is that what's easy for us is hard for machines and vice versa. But the paradox isn't the most interesting thing about Moravec. For me, this is. In 1988, Moravec wrote an incredible book called Mind Children.

[00:52:27] It sort of kickstarted the, like, techno-pro-extinctionist movement in some ways. You should read it if you wanna get an understanding of AI and how it's evolved over the past forty years, discussions around it, the philosophy around it, like, this is the book. In it, he obviously explains why the paradox exists, and here's his answer in a second.

[00:52:45] But it's wild to think that everything you're about to hear from Hans Moravec, he wrote when twenty-four pin dot matrix printers were the pinnacle of computer technology and when Tetris had just started tearing up the gaming sales charts for the first time. [00:53:00] This is me reading from Moravec "Encoded in the large, highly evolved sensory and motor, and motor portions of the human brain is a billion years of experience about the nature of the world and how to survive it.

[00:53:13] The deliberate process we call reasoning is, I believe, the thinnest veneer of human thought." End quote. A billion years of knowledge held in the body, and thinking, reasoning, art, mathematics, the things we believe, most of us believe anyway, make us human, is nothing but a film on the surface of it. And then in the same book, he proposes that we leave these bodies that hold those billions of years of knowledge.

[00:53:41] The climax of Mind Children is this robot surgeon with billions of fingers, literally each one the width of a molecule. It opens up your skull while you sit there awake talking to it. It reads one neuron, simulates that neuron, waits until the simulation matches, throws the original, that's right, your brain, away.

[00:53:58] Then the next one, [00:54:00] and then the next, until there's nobody left sitting in the chair. Nobody. No body. Because you would not have a body. You would be transferred, mapped, copied neurons on a hard drive on a cloud somewhere that is no longer embodied. Moravec suggested the body was carrying a billion years of knowledge, and then he said we should just move on from this form into the next.

[00:54:22] Caterpillar to butterfly, right? It shouldn't be surprising that Dan Faggella from episode eight is a big fan of Hans Moravec. And actually it was really fun for me to nerd out with Dan on Moravec, 'cause a surprisingly low number of people have, have actually read his work. You should remedy that. You all should remedy that.

[00:54:39] If you're interested in this, you'll find his book r- just incredibly thought-provoking. Moravec also thought these machines would be able to bring back the dead. And not as a metaphor, like he actually worked this out. It's, it, like the detail that he goes into on this stuff is just incredible. With a powerful enough mind, AI could reconstruct a [00:55:00] person from whatever traces of them happened to survive, the way an archeologist rebuilds an entire civilization out of broken pottery and fragments of remaining architecture.

[00:55:09] His example was Isaac Newton. Hand a super intelligent AI a copy of the Principia, and it could work backward from the book to reconstruct Newton This is basically what Ray Kurzweil has been trying to do with his own father for over a decade. It's just stunning to think that Moravec was writing this in 1988.

[00:55:26] He believes the last gift of our mind children, what he called our worthy successors as, as Dan Fogel would say, would be able to hand back to us everyone we ever lost, that we could all live together in some sort of digital afterlife. If you wanna see this all dramatized, watch Pantheon. It's an animated series that's on Netflix in the US.

[00:55:47] I'm not sure where you can see it elsewhere. Just search for it, Pantheon. It was originally an AMC series. I've been recommending this to nearly everyone I've interviewed for this show. So if you're a Patreon subscriber, shameless self-promotion, you'll be hearing about [00:56:00] it a lot in our unedited interviews.

[00:56:02] It essentially opens as a mystery. A bullied teenage girl starts getting help online from a stranger. Stranger turns out to be her dead father. His consciousness was uploaded through this destructive brain scan, and maybe, just maybe, there's something sinister behind the corporation who did the scan.

[00:56:19] That's not too much of a spoiler. You'll, you'll get there pretty quickly. Just watch it. It covers a lot of the same themes as this series, and it's beautifully made in every way. But yeah, that's Moravec's robot surgeon, the exact procedure I just described to you, told from the perspective of the person interacting with their dead loved one.

[00:56:37] This might seem familiar to you as essentially another version, a fictional version of Rebecca Nolan's story that we shared in a previous episode. Moravec wanted us to leave the body entirely, hand back the whole thing, a billion years of evolution to achieve something better, the next step in evolution.

[00:56:56] Hon Weng Chong, on the other hand, is worried about what happens [00:57:00] when we hand even less than that over to AI right now

[00:57:06] HON WENG CHONG: But I think the cause of concern for me is more an internal thing rather than external. And what do I mean by that is I feel, and I, and I have experienced this as well because I, you know, I used to code. Using agents to code reduces the amount of mental power or the mental capacity required to do something.

[00:57:28] And we, we know, you know, with, you know, physical therapy and so forth, that if you don't exercise muscles, they atrophy. Now, my worry is, is there going to be a future level of cognitive atrophy where we become so reliant on these systems, and this is what essentially the providers are making us do or become addicted to these things, that we can no longer function without these [00:58:00] systems.

[00:58:00] And I think it's quite scary. I-I've seen adults who can't do fractions or divi-divisions, and they're pulling up their iPhones, doing very simple sums that should be done in their heads. You know, are we hur-hurdling down this path as well now, not just, you know, with numeracy, which is kind of poor in most of the Western world now anyway, but also into language and communication.

[00:58:22] Are we gonna end up just being prompters, right? Are we losing our ability to express ourselves, to introspect, to, to analyze, and to think because we now outsource it to the machine? I think that is, for me, one of the scariest part of this technology, which is, you know, do we lose the, the, the most important thing about us being human, the, the thing that separates us out from all animals and, you know, breaks the poor m-Moravec paradox, which is our ability to think, to analyze, to create, to, to, you know, have this higher level of cognition.

[00:58:58] That I think is the biggest worry I [00:59:00] have, particularly not so much for this generation, but for the next generation. If we, we, we, we raise a generation that is addicted to these systems, and if these systems are not public utility with access, but are behind very large for-profit entities, that is a real worry that we, you know, may not be able to utilize all our cognitive capabilities without needing to pay a toll to, to use that.

[00:59:34] SEAN KING O'GRADY: It's worth pointing out that Hon isn't some anti-AI modern Luddite. He's a tech CEO. This is someone at the cutting edge of technology and business. It was actually just announced yesterday that he's building a first-of-its-kind biological computing data center in Singapore, and this is the person who's calling for utility-grade access to AI models for all [01:00:00] people, making sure it's not hidden away behind some paywall so that only some people can access it and, and be able to have the advantages of super intelligence while others fall behind.

[01:00:12] Th-

[01:00:12] this

[01:00:12] is pretty stunning

[01:00:15] I've got one question left. I ask everybody this. It's a yes or no question, but please don't give me a yes or no answer. Um, do you think AI will cause the apocalypse?

[01:00:26] HON WENG CHONG: Uh, that's a really tough question. I guess yes. Um, but then again, you know, it is... I wouldn't say it's AI. I would say it's, it's humans will cause the apocalypse.

[01:00:40] Now we just have an additional tool that can help cause it because we could already do it, right? You know, nuclear launch codes between the United States and Russia, enough to destroy the entire planet. AI is just another tool in that repertoire. Um, it's how we use it that is the most important thing, and ultimately it's back to humans.

[01:00:58] Um, you know, we have the ability [01:01:00] to destroy ourselves. Let's not, let's not do that.

[01:01:06] SEAN KING O'GRADY: Let's not do that. I agree. I wholeheartedly agree. Let's not do that. We opened up this episode by telling you we can finally see inside ourselves, and that part is true. Michael Ferguson can put a human brain on a screen and point to the piece of it that lights up when a person feels the presence of God.

[01:01:27] Hon Weng Chong can grow you a piece of a brain on a desk and teach it to play video games. Michael might even be able to see the part of my brain that would light up when you see God if this fall Hon's neurons are playing GTA VI. But nobody in the eight thousand years between the Catalhoyuk mirror and this recording could do any of those things.

[01:01:48] They aren't just new to me, they're new to our species. But when we ask these incredibly accomplished experts about the mind, about consciousness, neither of them could tell us what it is. Not really anyway. Neither could [01:02:00] Joel Gold. In fact, he further complicated things by bringing up the unconscious. I actually went and asked of a variety of different AI models about consciousness just to kinda get their responses, and they couldn't answer it either.

[01:02:15] This is not a failure on anyone's part. That's the honest limit of the sum of human knowledge. With the, with the chatbots too, they're pulling off of human knowledge. They can't go any further than we can on this. Michael's first answer to me when I asked him about consciousness was, "I don't know. I can riff."

[01:02:34] Hon was worried about crossing a red line and creating something that can suffer. These are two of the best-equipped people alive to answer these questions, and with both of them, we just ended up hitting the boundaries of their understanding. Eight thousand years ago, someone carried a piece of black volcanic glass a hundred and twenty miles across Anatolia.

[01:02:53] They ground it flat and polished it for hours so they could look into it and see a dark, distorted, probably nightmarish version [01:03:00] of their own face. And when they died, the people who loved them put it in the ground beside them. We still do this. We've only gotten better at the polishing Paul believed one day the glass would be taken away and all would be revealed.

[01:03:15] An apocalypse, you could say. Michael believes these machines might make the glass a little less dark. He may be right. I've spent a year now with people who looked into today's digital glass and were certain they saw something looking back. Some of them found God. Some of them found terror. One lost a piece of her dead father that she'd previously been able to access.

[01:03:37] We are standing in front of the brightest mirror anyone has ever built. It gets clearer every day, and so far it's shown us exactly one thing, us. Moravec promised these machines would give us our dead back. What we actually built, at least thus far, gave Rebecca Nolan a flawed version of her father for a while and then took him away in a software update of sorts of her [01:04:00] own mind.

[01:04:01] In the next episode of Suspicious Minds, we're switching up the format a bit to bring you the closest thing we may ever find to a first-person apocalypse story as we talk to Dr. Jeremy Heffner, a trauma surgeon and social media entrepreneur who has maybe the most intense AI experience I've ever heard

[01:04:20] DR JEREMY HEFFNER: When AI gets added to this, it causes humanity and civilization to completely fall apart and collapse 70% of the time.

[01:04:28] There's only about a 30% chance we survive this when I run the math.

[01:04:38] SEAN KING O'GRADY: So I'm going to read one, uh, one new Apple review, and then we'll move on to some Spotify reviews. I normally don't read the Spotify reviews because in the Spotify app I can actually write back to the people. But there is one I wanna share, and I think you'll understand why. So this Apple review just came in today And I might just read the whole thing.[01:05:00]

[01:05:00] "Everyone should hear this podcast," says Dagwood. "I'm a heavy podcast listener with my areas of interest being chiefly politics, science, history, and the arts. This is one of the most valuable podcasts I've ever listened to, and I cannot recommend it too highly. If you use AI or know someone who does, you need to listen to Suspicious Minds.

[01:05:19] Actually, if you live in the world in the present moment, you need to listen to Suspicious Minds. Well-researched, well done, and vitally important, I re- I am recommending this podcast far and wide." Well, thank you, Dagwood, for recommending it. Thank you for the, um, the incredible review. I'm glad you feel that way.

[01:05:34] The, those are all the reasons why we're making this. I mean, I agree. I think whether you use AI or not, we are living in a world in which AI is the dominant technology. I don't think we can hide from it. I don't think you can ignore it. I think AI is here. I think it's here to stay, and I think we all need to learn as much as we can about how to live, how to be a human in the age of AI.

[01:05:55] That's, that is exactly why we're doing this show. And then on [01:06:00] Spotify, we've gotten a bunch of cool comments. I've responded to most of them, but there's one in particular that I wanted to bring up that we actually got a couple weeks ago. James L said, "I love this podcast, but it is frustrating to listen to your very liberal guests pontificate about these things without an ounce of historical understanding.

[01:06:20] The Catholic Church and her great doctors like St. Thomas Aquinas came to these conclusions millennia ago. It's like watching a bunch of liberals ascend Mount Everest, thinking they are the first to reach the summit, only to be greeted by thousands of monks and priests that have been waiting for centuries for others to catch up."

[01:06:37] It's worth mentioning Michael Ferguson actually did bring up the scholastic thinkers and Aquinas specifically in this exact context in this episode today. But obviously, James was commenting on a previous episode where this had not yet been mentioned. I, I thought that was really interesting. I appreciated that comment because this is someone who's saying that they disagreed with the politics [01:07:00] of some of our guests, but still could find s- something valuable in the show, and I think that's really important.

[01:07:05] We try to avoid this necessarily being, like, a polemic or being overly political, at least from our point of view, because I think this issue is bigger than that. I think this development in society and in humanity and in who we are and what we do is bigger than that. So I, I think it's really big of James L to be able to see past the politics and, and, and see what's here.

[01:07:29] Um, so thank you for that. And thank you to our Patreon subscribers. We will keep getting long-form interviews up for you, um, as usual. Oh, and thank you also to people on YouTube watching and commenting. There's always lively discussion there, and we appreciate it. Thank you to all of you. We'll be back with another episode next week.