Transcription
[music] Who here saw the very strange interview between Tucker Carlson and Sam Altman? Have you seen it?
So, we all have our opinions, I imagine, about the moral character of each of these people. But um, if you have not seen it, I would very much suggest you watch it. Sometimes I feel like one of the reasons why Hollywood is collapsing is because it can't compete with the mad stories that are landing in our world. Um, and so I think that it would be a good idea to examine a little bit of what happened in that interview because it, it uh, features some of the big issues that AI is bringing to the fore.
And so, as you know, I mean, Sam Altman is the CEO of OpenAI. He has said that AI will displace the human in the same way that the earth was no longer the center of the cosmos uh with Galileo. That that is the level of displacement that we will see happen. Um, you know, we have uh the uh Elon Musk, who told us that AI is more dangerous than nukes and told and screamed and asked everybody for a moratorium, asked everybody to slow it down, to stop it, please. And then when no one did, he started XAI and created his own AI company. And so, what's going on?
We are told that the disruption that AI will bring will be bigger than the printing press, right? Will be the largest technological disruption that we have ever seen. You know, we know that the printing press was part of the reasons for the the war of religions. We're told that AI will be a bigger disruption than the industrial revolution, right? We kind of know that that disruption brought about the industrial wars that we had at the beginning of the 20th century. And so, we're just speeding towards this thing. Uh, and we have these papers that have come out now showing us already that new graduates that come out of college are struggling to find positions, that entry-level positions are being eliminated because the AI is doing is doing the work. The AI can do the entry-level positions and therefore people are struggling. We have these warnings of mass disruptions about employment. Uh, we're hearing and we know that all the students are using it to cheat, to do their work, to do all like all of this, we we kind of know.
Um, and the wildest thing about this is that this is the one that I find the most surprising is that the people that are developing the technology are also the ones that are warning us about it. So, on the one hand, they're developing this technology and then they're saying, "Huh, there's a percentage that just might destroy all of humanity." You know, it's, I'm just, we don't know how big the percentage is, but it's possible. There's a there's a possibility that this will disrupt everything so much, we have no idea what it's going to bring and yet we are racing towards this. And the scariest thing in this is that every time an AI company announces that their AI is somehow demonstrating an anti-human behavior or is somehow uh demonstrating the its capacity to escape or to transcend the bounds, the investments pour into the company. It almost feels sometimes like those announcements are actually marketing tools, that they're actually a way to get money is to announce how dangerous they did. You follow the Anthropic situation where they announced that their model was so dangerous that they had to secretly la give it to tech companies so that they could fix all of their bugs with their AI so that they could then, you know, release their AI. That's wild. Like when the threat is what brings people to invest in your company.
All right. So, what is going on? Why are we racing towards this? And so, let's go back to the Tucker Carlson interview. This is uh, so Tucker is, you know, he starts and he asks Sam about what is, what are these LLMs, these, what are these large language models? What are they doing? You know, why is it that people, he uses the word worship them? That's how he starts the interview. So immediately, you get the sense that he's going to make this interview reach some of the highest stakes. This is not going to be fun for Sam Altman. And it, it wasn't fun for uh, Sam Altman. Um, he asks Sam Altman right away about what his spiritual views are, which is, you think it's a very strange way to go. And Sam answers. He says, you know, you, I'm Jewish, I, you know, not, not kind of. And he suggests that, you know, he kind of is confused about God and about religious questions. And Tucker then reminds Sam that, you, you are about to possibly become one of the most powerful people in the world. Right? As the CEO of the company that is moving towards AGI, you are possibly going to become one of the most p uh, powerful people in the world. And that his ambiguity about his moral position and the ambiguity about the moral positioning of the OpenAI model is concerning because if we're going to have these the most powerful technology that the world has ever seen and the king or the head or the chief of that is morally lacking. Where are we going?
And that's when the interview becomes very disturbing because the next thing that Carlson does is he almost basically accuses Sam Altman of murder. It's one of the craziest things you see. And so, you know, what, what is going on? What is he doing? Okay, so I don't know the facts. I'm not a policeman. I'm not someone who could know what happened, who did what. But if you pull back a little bit and you look at what's going on, what Tucker is kind of grasping at or, you know, kind of coming towards is he's asking him, who does AI serve? What is it? What is it submitted to? Right? That's the question. What does the AI serve? And of course, the corollary to that is also asking him, who, who is going to possibly be one of the most powerful people in the world, who do you serve? What do you submit to? And of course, the implication is that if this person is morally ambiguous, if this person is morally corrupt, then how can AI become beneficial? How is it possible? Right? If we do not know what or whom AI serves, how can it be beneficial? How would it not lead us to a wasteland?
Now, this moment, this weird moment in this interview, for me, it reflects something that we have seen before in storytelling. So, in the medieval story of the Holy Grail, we have a similar problem which is suggested. We often understand the Grail is something that is holy, but that it's the receptacle of something holy. But if you look at the Grail itself, uh, you know, you see in that the Grail has a host in it, right? It has the body of Christ in it. And then later traditions, it will have the body, the blood of Christ. This idea that the Grail is the cup that received the blood of Christ. And so, it is this kind of container for holiness. But there's also a way in which the Grail is a technology. It's the container. It's not the holy thing, right? It's actually the thing that contains the holy thing. And that's why often in many of the traditions, the Grail is a cup, but it's also a book. And so, it's actually, interestingly enough, the Grail is information technology. And so, in the Grail story, the kingdom of the Grail, the kingdom of the Grail has become a wasteland, right? And the Grail is lost and everything is falling apart for that reason. But what's interesting about the story is that it's not that the Grail is broken. It's not that the Grail doesn't work. It's the king that's broken. It's the king that's wounded.
So, if you remember your Arthurian legend, the knight Balin comes into the Grail Castle. He's driven by revenge and he sees the person in the Grail Castle as, as he's being received by the king. He sees the person that he wants revenge on. And so, he breaks the rule. He breaks the rule of hospitality. That same rule that guided Jason, that guided Abraham, all the great heroes, all the great civilizations are held by this rule of hospitality. So, he breaks the rule and kills his enemy in the castle. And then he runs off and he's being chased by the people of the castle. And in his haste and in his uh, you know, obsession of his own personal desires, he can't see what's going on. He transgresses the Grail Chapel. He walks into the Grail Chapel. He sees the sacred holy Grail. He sees the lance that pierced the side of Christ dripping blood into the into the cup. I mean, if, if you saw that, you would probably want to pause and kind of try to think a few steps in advance. But that's not what Balin does. Balin is obsessed by the moment. He's obsessed by the short-term gain that he has just uh, engaged in. And so, he grabs the lance and he pierces the king in his groin. This is the dolorous stroke. This is what causes the wasteland. And so, the Grail is lost. But the Grail is lost because of an act of impiety, right? Because of a forgetting what the use of something is. He forgot what the purpose was. He forgot what this was serving, what it was aligned towards, and he only used it for power, right? And so, for that reason, the king becomes passive, the king becomes sterile, and of course, the king is wounded. For that reason, the Grail no longer heals the land. The Grail no longer does what it's supposed to be. And people forget, they have forgotten who the Grail serves. And that is the question that has to be asked to solve the problem of the Grail. Someone has to come to the to the Grail King's castle and ask, "Who does the Grail serve?" And asking that question and discovering the answer is what will restore things to their proper place, is what will help the Grail heal the land again.
Now, this is the problem in my estimation. The problem is that we're not asking the question. We're not asking, "Whom does AI serve?" We're dazzled by it, right? We're astounded by its capacities. We're terrified by some of its applications. Right? It answers our questions. It seems we discovered that it chooses our military targets, right? It's replacing our artists. It's going to drive our cars, possibly replace our therapists, maybe our girlfriends, possibly. Then we have this interesting moment where we have the CEO of Anthropic who wrote two papers, right? He wrote the paper to say, well, here's the bad stuff that might happen. Here's the best-case scenario for AI. Um, but no one's asking the question, why? Why are we doing this? Why, what is, who, who are we serving when we are doing this, right? Who does the Grail serve? Whom does AI serve?
Human beings, all of our actions, that is how they are structured. That is when we move, anything that we do is always in line, in direction towards a good that we intuit it. We see something, we intuit it, something to move towards, and it draws us towards it. Right? So, we see a good and then we move towards that good. But is all of our actions are always in service of something, right? Whether it's drinking water, whether it's walking out the door, whether it's founding a city, all of these are the perception of a good that we gather together and that we aim our actions towards. I mean, humans, that is also why we recognize models. That's why we see people that in that uh, are an image of the things that we uh, admire and we follow them, right? That's why we also when we see someone who is really good at something, we want to put ourselves under them so we can learn from them, so they can we can move into the good that they can teach us, that they can show us. This is how human actions function. In uh, in the Divine Comedy, this is one of the things that Dante is trying to help us understand, to try to try to help us see how humans are always moved in love and desire towards a good that draws them. Even when we're sinning, even when we're we're doing the wrong thing, it's always because we h, we are seeing a good, maybe we're mistaking it. Even when Balin took the lance, you know, and hit and and struck the king, he was moving towards a good. It was the wrong good. It was a short-term desire. It was a short-term passion. But that is what was driving his action. And so, this is the question that we always have to ask, which is, when we do things, what are we doing it for?
Now, with AI, there's a there's a dazzling thing that's happening to us, right? It's like one of the questions that people are constantly asking is, is the model AGI, right? Is the actual AI, is it conscious? Does it have its own goals? Is it aligned? No, if it does end up being conscious, will its goals match our goals or will it have its own goals? And so, we're so focused on this question of alignment uh, and the question of whether or not the model is conscious keeps returning over and over. If you've paid attention recently, is what led to the very, to my impression, very delicious moment when Richard Dawkins, who for decades has told us how submitting yourself, you know, to the source of all being, submitting yourself to the God of love that is the ground of being, is ridiculous and a superstition, but in talking to his AI chatbot, he was convinced that it was conscious. Why? Because it complimented his book. And so, it must be conscious because it's doing all the things a conscious person would do if they were conscious. It's uh, it's quite wild.
But, you know, what's interesting is if you look into this moment with Richard Dawkins, you realize that it's uh, it's not so much that he thinks that the Claude chatbot, his Claude chatbot is conscious. He, he actually is reducing our consciousness to whatever that chatbot is doing. And that's one of the problems that we're going to see happen more and more is that as AI becomes more powerful, we will adapt our notion of consciousness to what the AI is doing. You know, this is goes into the direction of the left-brain problem that Dr. McIil is, Dr. McIil was talking about. And so, but in our usual idea of what consciousness is, we know that AI is not conscious because AI doesn't have its own drive. It doesn't propose anything. It answers our questions. It doesn't have its own desires. It doesn't have, you know, if John Veri was here, he would say that it, it's not autopoietic, right? That is, it doesn't act towards its own good. It's not self-organizing in in itself. It needs our input. We're the ones who are making it move. The AI won't start to talk unless you ask it a question. It has no will. It has no desire. It's not hungry. It's not, has nothing drawing it towards the good. It doesn't have internal relevance, right? It's not, it doesn't, it can't tell what's important. It, it only is a parasite on our idea of what is important.
But if you want to understand what AI serves, who AI serves, you have to just, you have to pull out one level. You have to stop looking at just the actual model and you have to think about AI as an extended social phenomenon. Then things start to change. If instead of looking just at the model, you look at the companies and you look at the investment and you look at the political clout that is part of of AI and if you look at how it's being deployed, then you realize that all of a sudden you can see something more because that whole structure that is the different companies competing with each other, the different uh, nations competing with each other towards AI that can actually help us understand what's going on. All of a sudden, you do have a form of autopoiesis, that is, you have you have a kind of inner logic that is moving towards its own power, right? That's moving towards its its own continuation and it also has a a kind of relevance in it because it, everything is is moving towards a purpose. And so, but the question is, what is it? What does it serve? Right? And I would like to suggest is that it's serving Moloch, probably the best way of understanding it. This is, of course, you see when you look at the interview with Tucker Carlson and Sam Altman, you can feel that this is the thing that he's, they're kind of spinning around, right? Is that Tucker is asking about the model, but then he's asking about Sam, and then he's asking about the companies, and he's saying, "Okay, so the, this is actually a system. All of this is the system that we need to pay attention to." And when we look at it, then we realize that it's actually about, it's about power, really, is what it is.
Some of you probably have read Scott Alexander's famous meditations on Moloch. It's a blog post that was written now about 12 years ago. And in this blog post, what Scott Alexander does is that he takes, of course, the figure of Moloch. He takes it uh, takes it from a Ginsberg poem. You know who Moloch is? So, Moloch is an ancient Canaanite divinity that's mentioned in the Bible. And the thing about Moloch is, you sacrifice your children to Moloch, right? And so, in order to gain power, you put your children through the fire and in exchange, you get immediate power. So, what Alexander does is he reinterprets through this Ginsberg poem, but not only the Ginsberg poem. We've all seen, if you've seen Metropolis, you've seen that image of Moloch, right? Is an image that has kind of inhabited the the the modern age as the uh, system that is moving without without a perceptible uh, person in charge, right? That it seems to be moving towards a goal without anybody being responsible for it, right? It's a, it's a, um, it's a system that has that is moving towards this goal. So, a good way of understanding it, the arms race is one of the best ways of understanding the Moloch system or the Moloch trap, as we called it, which is that once that technology is put out in the world, there's only escalation that is possible. Is once you put the nuclear weapon in the world, then my competitor wants to make sure that I'm not stronger than them and they will arm. So, I arm and we keep going. And even though we both know that we might be leading to our own destruction, we still do it. We still continue to arm ourselves despite the fact that we know that it might lead to complete annihilation. This is the Moloch trap, right? Um, and so, what happens with the Moloch trap, of course, is that more and more is demanded of this system. It asks more and more of us. We put more and more of our energy into it, more of our money, more of our attention, more of our care. And we do it only to prevent our destruction from the other. But we are at the same time moving towards that destruction, right? You can understand this really as a type of will that's acting on us. And and uh, Ian this morning has suggested that there are these wills that act on us and we can see because we're acting, we're the legs of the beast. We're moving in the world and nobody wants to do it, but everybody's still doing it. And we and we realize that our actions are serving something over which we have no no control.
The thing about the nuclear arms race is its deterrence. So, actually, the Moloch trap works if you don't use the nuclear weapons. So, it has a certain characteristic. Now, that is not the same with AI. AI is the opposite. AI is a Moloch trap that only works if you implement it. And the more you implement it, the more you get drawn into the trap. So, it can help us kind of understand what's going on. Why is it that when a tech company announces how dangerous their model is, why are people putting their money there, right? When the, the very people that are making it are warning us that it would lead to our own destruction. That is why, right? And so, we are we are caught in a wheel where AI is being implemented as fast as possible in every domain possible. So, if you think of industrialization, you could imagine industrialization as a Moloch trap, right? Which is that as one country starts to industrialize or as one, one part of a nation starts to industrialize, everybody has to follow suit or else everybody will be left out in the cold and only those that move in that direction will win the game. Even though it might cost huge things to the environment, to to our our political structures, to our communities, it might cost that, but we have to do it. Now, in industrialization, you know, it was basically one direction. It was the production, distribution, and sales of goods. But AI is being implemented everywhere, in every single domain of human existence. So, it's as much in the medical, and the political, in the education, in industry, in law enforcement. Think of it, name it, AI is being implemented. And so, the AI, the Moloch trap is the most perfect. It's, it's like a perfect Moloch trap because it actually is has to be implemented everywhere and by everyone. And it's iterative, right? Because if we don't do it, we'll lose to China, right? And if we don't do it, we'll lose to the other company. If we don't do it, we'll lose to my fellow employee who's going to become more productive than I am. If I don't do it, I'll get worse grades than my other students. So, it's completely iterative at every single level of reality.
So, what happens when everything has been devoured by a Moloch? And it's interesting because obviously, we're sophisticated people. We don't believe in Moloch. You know, it's a metaphor, right? It's a metaphor for a game theory uh situation. But in the end, it'll be a very cold comfort when we're sacrificing our children to Moloch to know that it's a metaphor. Won't really matter. And what's fascinating about this, particularly this particular Moloch trap, um, is that it also now, we go back to the model, right? Now we can go back to the model. Is that if you go back to the model, then it becomes very interesting that the Moloch trap happens in the model as well. As the model becomes a kind of oracle, right? It becomes a purveyor of truth. If you look on X, it's super interesting. People will call, will invoke Grok. Have you seen this? In the threads, people will invoke Grok as the arbiter. "Grok, tell us what is real." And then Grok will reveal to them the truth, you know, and Grok becomes the ultimate arbiter of reality on X. Um, and so, uh, yeah, so it's, it's, it's quite, it's quite fascinating to watch it happen. And so, it's, it's interesting to see that the perfect Moloch trap also is manifested in that way.
Why, why does AI have a personality? You ever tried to ask the, why? They always ask like, what is it serving? Why? Why? Because you can imagine an AI that is solving problems. It wouldn't need to have a personality. Why are we giving it a personality? And the answer is because it feeds on our attention. It eats our attention. Its existence is based on our interactions with it. It's made by our attention and every single way, at every single level. That is on the one hand, it's made by our attention because we have to pay and justify for the existence of these massive, insane models. It's fed by our attention because it's being trained by us as we are engaging with it. And therefore, just like the social media world, the AI also needs to keep us constantly coming back to it, constant for it to exist. The AI needs us to love it. Without our love, it won't work. And so, everything about, that's why it's, why is the AI psychopantic? Isn't that the weirdest thing? You think like, why would it have to praise me? But it does constantly. "Oh, what an ingenious uh, question you have there. That's such a, it cuts to the deep, you know." It just, it keeps telling me how good I am. Oh, it feels good. I mean, that's how Dawkins was deluded, really, is that it's like he's being, he just needs a sycophantic god is what he needs. Um, and so, it, once we ask the question, right, when we ask the question, "What does the AI serve? Who does it serve?" And when we look at it at the at the bigger level, we are able to see what it is that's going on.
You know, there's a really interesting, if you remember the Matrix movie, there's this great scene in the Matrix movie, you all know it, when Morpheus holds up a copper top battery and he says, "We are the battery for the AI." Right? That's not how it turned out. Turned out to be kind of the opposite of that. Actually, in the, in the Matrix, we, our bodies are the batteries for the AI and our brain processes are the battery for the AI. And the AI creates a fake world to keep us calm and to keep us inside. But this is now, it's the opposite. The AI is the opposite of that. We are the intelligence of the AI. We're the only intelligence it has. It's actually fishing for intelligence. It's fishing for qualia. It's fishing for discernment. It's asking us to tell it what is right and what is wrong. The, our participation in the AI is the shape of its soul. Whatever soul it has, it doesn't have a soul on its own, but whatever artifice, whatever imitation of a soul it has is is taken from us.
Now, it's kind of a dire situation, but this, this uh, this has been drawn to us before. So, what, obviously, we we feel, I mean, I sometimes feel like, can we just escape this? Could we just leave? Because this idea that now we realize that AI is asking us to sacrifice ourselves for its existence. It's asking us to give our love, to give our judgment, to give our ideas, to give it to the machine, to Moloch. Um, you know, couldn't we just get out of this? Couldn't we just uh, escape this? And the problem is that the problem is that we, we can't, we can't escape, not at least the world of sacrifice, not the world of asking, "Who do we serve?" or "What do our systems serve?" That we can't escape that. But if we become aware of the problem, then we can at least make decisions about that. We can at least ask ourselves, "Who do we serve?" Right? Who are we serving in the things that we are pursuing? In the things that we are following.
This image of sacrifice is obviously there in other myths as well. If you know the story of Prometheus, it's quite fascinating because we always think of Prometheus, of course, as the one that stole the fire from the gods, right? We always think of Prometheus as the one that went up the mountain, got the fire and brought it to the humans, therefore making all technology and techn and all of the human arts possible. But, you know, that's not the fullness of the story, right? Because there's a part before that people for some reason don't say as much, which is that the humans had the fire. You know that, right? Do you know why they lost the fire? They lost the fire because Prometheus tried to offer an improper sacrifice. Prometheus told the humans, "Put the skin on the bones of the animal, right? Remove the meat. The God doesn't need the meat. Keep the meat for yourselves. Remove that. Offer the skin and the bones and that'll be enough." And of course, when the gods received the sacrifice, when Zeus received the sacrifice, he recognized that it was not the proper sacrifice. And that is why Zeus removed the fire from the humans. Now, it's interesting because we have another version of that in the story of Cain. If you know the story of Cain, it's very similar, right? Which is that Cain offers the improper sacrifice and it's because he offers the improper sacrifice that he is chastised by God. And God says, you know, this is, if you do not offer the proper sacrifice, right, then you're not, you're not doing well, you're not going to do well. And it's kind of fascinating, don't you think that Cain is also the one to develop technology, right? Prometheus, we have another one. If you know the story of Hephaestus, very similar, similar to the Grail story in an odd way, which is that Hephaestus doesn't understand the proper order of things and he transgresses, he transgresses the holy place in a different way, which is that he tries to sleep with his own mother, tries to transgress the order of things, right? And for that reason, he is also cast out. And we see an interesting uh, part in the story of Hephaestus, which is that he becomes lame. And you could see it that because he's lame, that's why he develops technology because he's lacking. That's why he develops technology. And it's the same with the story of Prometheus, that the fire has been taken because of the improper sacrifice.
Now, what does that, what does that mean for us? When Prometheus steals the fire from the gods, do you remember his punishment? His punishment is that he's going to be tied to the face of a rock and every day an eagle will come and steal his liver. Then the liver grows again. The next day, the eagle comes. Have you ever tried to think of the relationship between his punishment and his sin? It's kind of interesting. He didn't offer the meat of the sacrifice. And so now the eagle is going to come and take it. Now the eagle is going to come and take the meat. And this is, this is the secret. Reality will take its chunk out of you if you don't offer yourself right. If you do not sacrifice to the proper, in the proper direction, if you not sacrifice properly to the proper things, to the proper, in in the proper way, then reality will take that chunk from you. It's just going to happen.
Being a human being is difficult. And one of the things that the AI is doing is it's making it really easy to be a human being, right? Very easy to do the things that we need to do. Meanwhile, it's literally feeding on us. And that's why one of the reasons, that's why I love Rilke so much because there is this realization that we have to make the sacrifice. Learning a language is hard. Reading the ancients is hard. Becoming wise, becoming a person, becoming a friend, becoming a husband, becoming a father, all these things are difficult. But these are the sacrifices that we are meant to do. I mean, Christianity really kind of brings it together in this surprising image of self-sacrifice. Really is the, it really is the solution to the question of sacrifice, which is that we are asked to give ourselves, we are asked to give ourselves to the highest that we can see, the highest that reveals itself to us. And in doing that, we will prevent, even in our own lives, we will prevent this kind of unconscious meat that is being constantly taken from us.
I think that it's a very fascinating time. I think that it's a very exciting time because as we see the Moloch system grow, it's going to be difficult. But those that know how to sacrifice on purpose and those that know how to dive in to have the wisdom and the knowledge inside them and not out here in my hand. There was a beautiful moment yesterday. Stephen, we were having dinner and Stephen couldn't remember the name of a city in Italy and he was struggling to find it and then he said, "I'm just going to look it up." Then he put his hand as if he was going to get his phone and he said, "No, I'm not going to look it up." Obviously, it's a very simple, it's fine to look up the name of a city, you know, on your phone. But there was something in that gesture which was saying, "No, no, no, no, no. I want to have that knowledge inside of me, right? I want that knowledge to be inside me, and I know it's there, right? I'm going to let it rise up in its time, you could say." And it did ultimately, he said. And so, that's why, like I said, that's why I, I love, that's why I love Rilke. That's why I think that what, one of the things we can have right now is like, is a true call to adventure. It doesn't mean that we can't use AI, but we have to be able to understand what it is and who it serves. And we have to be very careful that we don't serve it, at least in our private lives. I don't know how much we can do to change the course of history. I don't know how much we can do to stop this process from reaching its its final uh, course, you could say. But what I do know is that whatever crash happens, whatever devastation it causes, whatever wasteland happens because of it, or even if the best version of it happens, if the, it becomes a tool that helps us all do the things that we really are important for humanity, let's say it even reaches that, no matter which way it goes, whether positive or negative, those that will come out on top at the end will be those that have learned to sacrifice and will have been willing to ask the question, "Who does it serve?" both the AI, but then also, "Who do I, who do I serve?" And so, thank you. [applause] Thank you. Those are both wonderful lectures. Um, one question I have for you both is, um, the way AI was presented, particularly in uh, your lecture, Dr. McIil, was as sort of an extension of the, the left brain, as a tool that sort of gives our left brain more power. Do you think that AI will never be able to transition to being being able to model the right brain, or do you think that's also uh, in the cards?
Yes, that's a, um, a good question, a difficult one, really. What I would say is they can guess ever better at imitating the the sort of way in which the right hemisphere works, but without actually what is essential to it, which is the element of embodied and soul experience. So, all it can really hope to do is to represent rather than presence. That's a very important distinction between the right hemisphere's taking of experience and the left hemisphere's modeling of it. And it is the way I think of it is, you know, a lot of people find it very surprising that if you have a line three inches long and you extend it a thousand billion light years, it's no closer to infinity than it was when it started. And the approach that AI manages is asymptotic. It's getting closer to something, but there is always a chasm there that cannot be crossed. So, in the 18th century, they had brazen heads that could talk, which was interesting. Uh, what we've got is unbelievably more sophisticated brazen heads. Yeah. We always have to remember that the categories that the LLMs are using, they [clears throat] have no understanding of them. And an an LLM doesn't know what a cat feels like when you hold it in your arms. It can't. And so, it has definitions, right? And it has all of our writing and all of the things that we've said about it, but it doesn't have, it doesn't have, it can't have knowledge because all the categories are categories that are based on embodiment, of our embodied experience.
Should we be pessimists or optimists? Should always be optimists. I think so. I think, I mean, there's no other thing to do. Why would you be a pessimist? Because even if everything's falling apart, you have to do what you could do, you know, to make the best of all situations, you know. That's why I do think that, you know, I think that we have adventures ahead of us that our ancestors never could imagine, you know, and remaining human in a in a more and more inhuman world is going to be, uh, it's going to be exciting.
Well, I call myself a hopeful pessimist. Uh, which, um, which, which is more comprehensible than it sounds. In other words, I'm a pessimist when I look at the way things are going in the world. But first of all, I think it's a duty to be hopeful. Secondly, one has no way of predicting the future. And those who have, been prophets in the past have been, uh, comprehensively wrong. One of my intellectual heroes is Alfred North Whitehead, and he said that prophets should be, um, taken outside and sort of mercifully stoned to death because they're such a liability. So, I'm not prepared to prophesy about the future, but all I can say is that from a purely from my limited rational perspective, it doesn't look terribly good.
Well, I have a question for you, Dr. I mean, sometimes I say I'm a short-term pessimist, long-term optimist, which is that reality actually has a structure, right? It actually has a way of being. Being has a reality to it that you can twist for a long time. You can kill a lot of people, but at some point, the lies will fall apart, right? That which is untrue will collapse.
Yeah, I like that and I agree with it. Um, thank you for, you know, pointing that out. Uh, it does depend on how long you, the question is whether humanity will actually be able to survive long enough to make the comeback. But history has always gone in what is often called cycles. As I prefer to refer to the image of a spiral, in other words, it's not linear, but it, it circles around, it never comes back to where it started from because it can't do that, but it comes back to a, a place that's cognate with where it started from at a slightly higher level. And there's a sense in which the worse it gets, the better it gets. And sometimes it has to get bad for it to begin to change. We need effectively a really robust response to this crisis. But the thing is that the way we now live, think, and work has become soft, lazy, complacent. And that is not the place from which to answer with the kind of courage that is required. Thank you both.
Just to pick up on that last point and the question is primarily for Jonathan, but Ian, maybe you could chime in, that it has to get, um, worse before it can get better. Jonathan, you use the, the image of the nuclear arms race for this, for so to speak, a Moloch system. But, and if I heard you correctly, you were saying that the only possibility is escalation. But we have historical examples where de-escalation was achieved.
Yeah, and you have to appeal to a higher truth. Okay. Okay. Good. So, you need a god. To Scott Alexander, in his blog, he says, "Only a god can defeat a god." Okay. Well, my question is, and I'm thinking primarily, I agree with you, but I'm just thinking of the fascinating history of diplomacy and so on and geopolitics that led to moments of de-escalation. Never perfect, but something was achieved. If there's something to be learned from that, can that be applied to the fear that we have of a constant escalation when it comes to AI, or could we apply some of the same principles? And are we seeing some of the principles work out currently in the discussions with the CEOs and so on?
My my intuition, and so now I'm going to be a pessimist. Sorry. Hi. My intuition is because because AI is being implemented everywhere simultaneously, we have a a bigger obstacle to de-escalation than we did with nuclear arms, which is that the nuclear arms race were were concerned with only a few people. That is, only a few people had the capacity to make decisions on that. Uh, and they were heads of states, they were, you know, generals in in armies. But now we have this problem, which is that first of all, these companies are private, and then they're being instigated at every single level. And so, the, the difficulty, and this is one of the problems that we have, is that the more this version of the Moloch trap is making it completely dependent on its own process, right? And so, the, you could say the god or the demon or whatever is actually making us dependent on it. So, once you know it's, and it's, it's like a, it's a technological progress, but it's, it's what is happening so crazy so fast that by the time that I don't have books in my house, like very fast, like you won't have books in our house. And so, we need the AI because without it, we won't have access to anything, access to knowledge, you won't be able to buy anything, you won't be able to, so this is the problem is that it's making us dependent on it. So, de-escalation seems more difficult. Yes.
On on the book thing. I, I've just learned something terrible in the last few days, which is that, um, is the sign of things that can will happen is that Iran has deliberately targeted enormous data centers in, I think, United Arab Republic and Oman. So, these are the great places that, in which we store, as we think securely, um, all the amazing things that we have created, our works of art, our books, and so on. But really, they're just on someone else's computer, effectively. And that computer can go down, that computer can be destroyed. So, I would like to say, hang on to your books, hang on to all the physical things. And also, I miss people. I don't know about you. I think people are rather important in life. And more and more of what I do is remote. I mean, I'm lucky in that people want to have conversations with me. So, that's good. But just in the business of doing ordinary life, I'm I'm alienated from people. And it takes my life away. See, there are, we're now in a new soft feudalism. There are people who have colossal riches and and influence. We've talked about that. You know, I never thought in my lifetime, growing up in the in the 60s, that the I would live to see the gap between rich and poor increase. It was going the other way. But now, we are the new serfdom because those people get richer and richer by taking away people's jobs. There's a whole issue, of course, about what happens to those people who've been deprived of one of the things that gives meaning to their lives. But it's also that it now takes me 10 times longer, such as technological progress, to do something that used to take a five-minute phone call. I could spend a whole morning going round and round in McIil-ite loops on a platform that hasn't been properly thought through and not achieving my end. Um, so I think there will be a move to bring back. I don't know, I may be wrong, but there will be a return to, well, let's do this the old way, that is direct and human. I'm hoping. Yeah, I'm a hopeful pessimist.
Well, for sure. It's funny because when we talk about this problem of the computer, this is actually one of the things that makes me think that maybe there really was a a civilization before the ice age because if there was a massive catastrophe now, everything about our world would completely vanish. Like it would just vanish. There'd be no trace of it. All this, all this plastic, you think, what, what's going to be left, you know? And so, it's just interesting that the ancient people had the wisdom to write things on stone because they, they knew that, you know, anything if there's a, if there's a, if there's a solar, you know, flare, there's volcanic eruption, if there's all the kinds of catastrophes that all this stuff that we have is just going to vanish. It could vanish very fast.
Hi. Um, I work as a storyteller at an AI research company in San Francisco. And one of the questions that's top of mind right now is like, what are the tasks or domains that we can and should outsource to AI, um, that would lead to an expansion of our capabilities versus the environment? And, um, especially when many of these lower-level tasks are what build this embodied knowledge. So, I'm curious what you think about what are good uses of AI. I want to know what a storyteller and AI company uses. I want to know what that, I want to know what that is. What's a storyteller at an AI company? To I guess advance a more humanistic, um, vision for our AI future and what we're trying to build. So, I'm, I'm afraid because I'm a deaf twist, I didn't really hear what you said. [laughter] But I, I think it was something like, I'm told it was, um, what are some good.
Uses of AI, or something? Yes. That leads to, I suppose, like an expansion of our capacities as humans versus our infringement. Yes. Uh, God, I'd have to scratch my head over that one. I, I, um, I mean, obviously, that there is an awful lot of drudgery in life. And if AI could be confined to dealing with things that are in fact mechanical, um, rapidly, that's a great help. But unfortunately, it won't, it won't be confined. I mean, it just simply can't be because, you know, even if the world was composed only of, of, you know, people who had the highest moral standards, uh, I, and I don't think that's necessarily the case. Um, eventually people will put it to uses which will be very harmful. So you can't have that. Um, you can't have the good, I'm afraid, without the risk of the bad. And there's a precautionary principle that if the risk of the bad, it may be unlikely, but it's if it's that serious, then you should be trying not to go there. But I, I think it's too late for that. Yeah.
Well, I mean, I'll tell you how I use it. I use AI, you know, and I think I appreciate AI as a research assistant, basically. That's how I use AI. Um, but it, it, it means that I already have, I already have the knowledge in me, and I already have what it is that I want to do, and where it is that I want to look, and I use AI to help me just look faster, basically. Kind of going to a library and instead of, back in the day, I had to go through the catalog and you had to take, um, and so that's the way that I, that I use it.
But I think that like, uh, Dr. Yan said, "One of the issues, one of the things that I'm seeing that's going to happen is that AI is going to make people that already have knowledge and wisdom now extremely powerful because it does that. If you already know what your skill, you already know your career, you already know the skill that you have, and now you have this amazing tool, then you can, you can really rule." But the, the problem is, the problem is that that now there's no more path to getting there, you know. And so, you know, and I'm, I'm looking at you, Christian, uh, as you're nodding along, and I'm thinking in your field, it must be particularly true. That is that all the work that the junior architects are doing, the drafting, all of this kind of stuff that would lead them to becoming master architects, you don't need them. You just have the, you just have the AI do it. How do you get to the wisdom that you would need and the knowledge you would need to use the tool properly and not just as a, that I don't, that's the big, to me, that's the biggest problem. And I think we're seeing it with the fact that young people aren't getting hired is how, what, how do they become the person that could actually use the AI for good? And I don't know. What do you think we should approach the non-propositional knowledge? And do you think it's urgent that we give like truth value to non-propositional knowledge? And also, what do you think we lost, if anything, by sending into exile non-propositional knowledge or the soul from the understanding of human nature, especially in medicine? Thank you.
Yes. Wow. Um, good question. Um, I think it's clear that we overvalue propositional ways of thinking. Um, and all the really important things, uh, are more dispositional than propositional. Now, I'm not saying that thinking conceptually with propositions is not valuable. Of course, I'm not saying that. It's always been the case and it always will be. But we've lost the sense of what the really important things are. Um, for example, I mean, an obvious example to me is the triumph of utilitarianism, um, in philosophy faculties, uh, as a way of achieving the good. Um, in fact, it's interesting that people who are very utilitarian in their way of thinking, um, often have frontal lobe or right hemisphere damage, and they, they tend to be people who have a reduced aversion to causing harm, a greater degree of selfishness, and behave in, um, an essentially psychopathic way. Psychopaths are very utilitarian. So you have the very odd thing that in university faculties across the West, what is lorded is a kind of way of being moral which is for most people what we would consider quintessentially immoral. Um, so that's just an example. But we, we've lost the idea of the appreciation of art unless we can say what the artwork is saying. You know, you go to an art gallery, you look at the art, you go on, you go look at a text to find out what the artist meant. Well, that's not a work of art. You know, Leonardo didn't have to write that, oh, but it's this, you know. Um, and so I think we've lost that.
And I champion the idea of truth as being dispositional, that the true is that to which we are faithful. The, the root actually of the word true is a Germanic root *trew* meaning loyal or faithful. So the idea is not of truth out there as something we go and get, and we do it by a chain of propositions, but by a kind of whole being respect for what it is that we're encountering and trying to be be faithful to it. So, you know, I think it has very deep roots. And it's also true of religion. I mean, I, I'm a Christian. It's the tradition I was brought up in, although not my family, but that's where I ended up. And Christianity has the extraordinary distinction of being, um, lumbered with a creed of propositions that actually make sense at the end of a life, but make very little sense at the beginning. So, uh, you know, a move towards a disposition towards the world, which is the important thing. And for me, I mean, it's a thing that I'm always saying and I'm known for saying is that attention changes the world. So it's not just the thinking through propositions. It's how you attend to the world. It dramatically changes what is there. And, you know, a couple of great philosophers of the 20th century, both of them French, Louis Lavelle in the early 20th century and Simone Weil later said in different forms that, um, attention is love is a matter of giving your attention to the other. And unless you give your attention, give your attention, not force your attention or direct your attention, but give your attention to another. It is a great gift, then something very, very important is missing.
And you mentioned medicine. Could I say, uh, I, I think I actually said that in my talk, that the way in which we're trained is very much that the body is a thing, a mechanism, and so forth. But as soon as you actually meet suffering human beings, which, by the way, is the meaning, of course, of the word patient, it doesn't mean having lots of time on your hands. It means you're suffering. And the thing of the encounter between the physician and the patient is two suffering individuals. Two broken [clears throat] individuals, broken, limited souls. One of them trying to listen to, attend to, and give something to the other. And in my experience, amazingly, attention on its own can help cure. I mean, not produce a total cure, of course, but sometimes I would have patients who would come to me, and I would listen to them, and I would actually direct what they told me. They would not be aware of it because I didn't have a formula or a sheet. But by the end of about an hour, an hour and a quarter, I had found out just about everything I needed to know. And they'd often say at the end, "Gosh, you know more about me than my husband, my wife, my children, my parents," you know. Um, and then I'd say, "Okay, well, let's think about what to do." And I, we'd have a plan of action, but that plan of action would not actually yield any results for a few weeks or a couple of months. But I would see them again in a week's time, and very often they'd already say that somehow they'd improved. [laughter] And the only thing I can, I can think is that actually it was a new experience being listened to attentively, compassionately. And so attention changes the world.
I, I, I don't have much to say, but I would say for sure, I think we're seeing, you know, there's this wave of young people going back to church, you know, we're kind of seeing it happen. It's still statistically not relevant, but it, it is happening. So, I think that young people that are drowning in this, this world, they are looking for participative, uh, meaning. They're looking for, you know, like you said, John Vague has the P's, right? So, it's like participative, uh, perspectival knowing, all these different ways of knowing that are embodied and that are related to your disposition rather than just a mental thing.
Join me in thanking Jonathan Nian. [applause] [music]