Ken Liu: Art in the Age of AI — #79
There’s no question for me that a large number of jobs in those parts of the arts can and will be replaced by machines, and humans will continue to consume them, perhaps even larger numbers. But that does not mean somehow that the kind of art that I’m talking about, where the whole point is to push beyond the cliché, to say something that does not exist before, will somehow die. I just don’t see that. And if machines are capable of doing that as well someday, then I would welcome that. I will be very interested in, as I said, learning about what the universe looks like from the perspective of a toaster.
Welcome to Manifold. My guest today is Ken Liu. Ken is a highly acclaimed writer of science fiction and fantasy. He has won the Hugo and the Nebula Award, and in fact, won both for a single short story. This interview is happening for slightly odd reasons. It was brokered or arranged by a high government official in Singapore who knows both Ken and myself and is a Manifold listener. And she sent an email saying, hey, you should really interview Ken Liu. And I said, yeah, I’ve been a fan of his for some time. And she said, oh, I can look you up. So Ken, welcome to the show.
Thank you, Steve, for having me. Yeah, it’s a pleasure to be here. I’m really glad to be on the podcast. It should be a really fun conversation.
Great. Now, Ken, just to introduce you a little bit to my audience, because not everybody who listens to Manifold is actually a science fiction fan. I myself was a rabid science fiction fan as a kid, but then as an adult have read less and less of it over time, just being too busy. Ken is a prolific short story writer. And one of his short stories, which I hope to discuss later in the discussion, is really one of the most highly acclaimed of all time. It won both the Hugo and the Nebula. But he’s also a writer of epic grand fiction. So stories that in a way you could compare to The Lord of the Rings or Game of Thrones, things that are very, very complex in their plotting and execution. And so he’s a writer of very great breadth and capability.
Also, being fully bilingual in English and Chinese can draw on both the Western repertoire of archetypes and mythologies, but also from deep histories and ideas coming from Asia as well. So, Ken, I wanted to start out by just talking about your childhood history. You were born in China but immigrated to the U.S. I think when you were like 11 years old. Maybe talk a little bit about that and how you ended up at Harvard College studying both in, I think, English literature and computer science at the same time.
Yeah, so that’s a great question. And it’s one that I think is interesting to address because I think there’s a really terrible tendency these days to sort of view someone’s biography as deterministic of whatever they end up doing, which is incredibly annoying to me. So one of the reasons why I sort of enjoy working in the genres I do is that there is a nice way to sort of not have to be forced to stay in your lane and to do things that only other people expect you to do.
So, you know, when people know that I’m an immigrant, they keep on reading my works through a Chinese lens, which is incredibly annoying to me. I don’t write in that way. I write in the way a human being would write. And so in terms of that history, I think the history is interesting because it allows me to see how I can approach that collective unconscious, which is where all stories come from. I think it’s very simplistic to say that somebody, you know, they were born in China, so their stories must have a lot of China in it, which is nonsense.
But that childhood is important to me in that it allowed me to have a perspective as most third-culture children understand. It allows you to have a perspective on the cultural environment around you. And you can remain always in this very interesting position where you see the insider’s perspective, but you also see what it’s like from outside of the system. And I think that’s one of the great advantages of growing a third culture. You have this sort of perspective on the world that’s more cosmopolitan and more, I would say, dreamlike, because you are able to discern the slippages between various systems and then see the patterns.
So to me, that’s probably the most interesting part of my childhood, which is this idea that because I sampled a bunch of different cultures, it is now possible for me to see the systems from this perspective that is not embedded in meshed, and therefore sort of all consuming and all blinding. I have a way of seeing commonalities between all human practices. And in that way, I think it’s not the China part of it that’s interesting to me. It’s the fact that it is, in fact, third culture, and therefore, cosmopolitan.
So when I was growing up, I could always adopt the viewpoint, I could simulate, say, how my immigrant mother and father would view something that I was seeing on an American playground. But I can always, of course, being born here, think of things in terms of completely the point of view of these, you know, say, white American kids in Iowa that were my friends. So obviously that, I think, for many immigrants, this makes you a very, very cute observer of both cultures throughout your whole life.
I hope that’s true. I also have to be very cautious about that because one thing I’ve learned is that you can be very empathetic without necessarily being truly understanding, if that makes sense. I think my position allows me to be very empathetic to different perspectives, but I would never claim to actually be able to understand everyone in that way. I think it’s a subtle distinction, but one that is important to me. I think the issue is, I think a lot of people who claim to be very empathetic end up claiming the power to speak for other people. I think that’s one of the great dangers in our culture, people claiming to speak for other people. I don’t think people should ever try to do that.
You can’t claim to speak for other people. All you can do is to say, I can empathize, or I think I can, but my understanding has to be cautiously understood as contingent and provisional and need to be revised when people tell me that is not correct. You can be empathetic without necessarily thinking that you actually can take on their positions.
Yeah. I mean, even if you were from a single culture, there’s so much variety in the minds that constitute that one culture. No one could claim to speak for all of those people. When you were young, did you feel alienated at all in America? Did you have difficulty fitting in when you first immigrated here?
I would imagine everybody does. I think that’s one of the reasons why I sort of avoid claiming that the immigrant experience is unique in some way. My view is that in the modern world, everyone is an immigrant. I think that is a perspective that I think I have that perhaps not everyone necessarily agrees with, but I think it’s actually a deep truth. One of the features of modernity is a deep sense of rootlessness that is imposed on a lot of people. So we’re told by our environments that you can’t stay where you are. You have to actually go somewhere else. You need to go elsewhere to pursue opportunities, to get the jobs, to learn codes, to learn to code switch.
You know, even somebody as vice president like J.D. Vance in his original book, what was a hillbilly elity about? It was an immigrant story. It was about him moving through, migrating to different cultures. I don’t think many people read it as an immigrant story, but I did. I think that’s an example of the sort of thing that I’m talking about. In this modern world, we are constantly telling people all over the globe that they need to change who they are. They need to move outside of the worlds of their birth to change for something else, for a different career, a different city, a different education background, etc.
These are basically immigration experiences. You are learning to fit into a different system. You’re learning different rules. You’re learning to see your own sort of assumptions that you were born with in the new light. So I would argue that I’m no more alienated than just about anybody else because we’re all immigrants.
Growing up, did you have literary interests at an early age? And did those transfer from being sort of reading stuff in Chinese into reading stuff in English?
That is a really interesting question. Let me think about it. I certainly enjoy reading a lot and I read all sorts of stuff. I don’t know how you would actually compare the types of things you would read in one language versus another. I mean, I read very widely throughout and I didn’t particularly, I don’t think I changed in the sort of stories I enjoyed reading, moving between languages. So I would have to say that I’ve always enjoyed reading things that I found interesting. I don’t think the language had any particular influence one way or the other.
I see. And by the time, say, you were in high school, would you have articulated that someday you wanted to write or be a writer, or was that something that came to you later in life?
This is one of those questions where writers always make up answers for because, you know, we have a tendency to sort of assume that people’s life stories make sense. The reality is people’s life stories often are just random events, one after the other, and then we retroactively go back and try to make sense of it all. I think that’s much more honest than this idea that our life stories follow a certain kind of arc and a plot.
So what I would say is if you squint hard enough, I can sort of go back and look at my life events and say, yes, it seems like there was always this desire to tell stories, to write and to craft things out of symbols. And that I started doing so back in elementary school. And it’s just that I didn’t seriously pursue it as a thing that I might want to do as a profession until I was in college. But another way of looking at it would be that it’s one of the many things I was interested in. And there are probably many alternative universes where I ended up pursuing those other routes.
There is a certain universe in which I ended up actually as a mathematician, the way I thought I would be in high school. There’s another universe in which I ended up as a computer science researcher, which came this close to being true. But this is the particular path I ended up on, and there’s a lot of randomness to it. So I wouldn’t say that I had wanted to be a writer so much that I ended up somehow getting myself in that position.
In some sense, the career I ended up having is the result of me sort of doing things I wanted to do and trying to pursue interests I was drawn to. And then this just ended up being the result of both good luck as well as a particular kind of liking for the work I did. So that’s probably a less satisfying kind of answer to some people, but I think it’s much more realistic. I think a lot of us end up in this position partly out of desire and also partly out of luck and random chance.
You’ve had, I would say, at least three careers, right? So you worked at Microsoft in computer science related things. You then went back to Harvard Law School, and I think at some point you were a practicing lawyer. And now, are you a full-time writer now, or do you still keep up some of those other activities?
Yeah, yeah, my career is not a good model. It’s not what you want if you’re trying to teach people, you know, how to have a successful career, I suppose. I will say this. I started out as a software engineer, like you said, in Microsoft and also at a startup company. And I enjoyed working in tech. And in fact, I still think of myself as a technologist. It’s a lifelong passion and the work that I, the sort of stories that I tell, the books that I write, whether they’re fantasy or sci-fi or whatever, they’re fundamentally stories about technology in the sense of technology being a manifestation of human mental patterns.
So whatever you see as a manifestation of human mental patterns in the universe, that’s technology. And my stories are all about that. So I had a career as a technologist, which I very much enjoyed. I also ended up having a career as a corporate lawyer, which I also very much enjoyed for very different reasons. And a lot of my fiction shows this interest in the law as a way of ordering human societies. And then I also spent some time working as a litigation consultant in technology cases, sort of combining both the tech and legal backgrounds.
As of five years ago, I’m thinking five, six years ago, I went full-time as a writer. So these days I am just a spinner of tales, just a writer of words, even though I still have a deep interest and passion in technology. And some of my work in terms of the futurism work that I do is very much related to technology. I would say all of my stories are really ultimately working out, attempts at working out the implications of technology as contemporary epic poetry.
You know, I’ve heard you say that all three of those careers that we mentioned have a commonality in that you’re manipulating symbols, you’re creating things with abstract symbols. Maybe you could talk to that, speak to that a little bit.
Yeah, so the way I think about it is, when you are a programmer, right, you are constructing artifacts out of symbols. That’s really all you’re doing. Software is just, what is software other than a bunch of symbols put together to function according to the rules of computation and the programming language you happen to use? You’re constructing virtual structures and machines out of these symbols. So that’s very easy to understand.
As a writer, though, what you’re doing is not fundamentally different. You’re putting words, sentences, images, symbols together into a structure, and it’s supposed to achieve a certain effect. The difference is, of course, that your programs are executed by machines, whereas stories are interpreted by humans. And these are very different interpretive activities. In the case of machines interpreting programming code, it’s a purely communicative act. The programmer has succeeded in so far as the program passes the unit tests, right? I mean, it’s a very simple way to measure whether you succeed or not.
Literature is very different. And in fact, this is probably one of my most talked about topics on which I’ve written a lot. And it’s this idea of storytelling is not exactly a communicative act. Because a story ends up being a structure that embeds a bunch of very complicated ideas in the writer’s head. The difference between a program and a story is this. A program represents a program as an exact instantiation of the ideas in the programmer’s head. Exactly. Like everything that the programmer thought about this virtual machine is embodied in the linguistic construct called the program.
But a human writer, as Ursula K. Le Guin has said, an artist working with words is trying to say with words what cannot be said in words. So it’s a paradox. The words that we put down on the page are there to express the thing that we cannot express with words. I mean, again, I tell people, you know, you always ask me, right, interviewers often ask me, what is the message of your story? What do you want readers to take away from it? And my answer is, if I could reduce my story down to a message, I would have just written the message. The reason why I didn’t read that is because I don’t have a message, but I have a story.
Stories are there to say things that cannot be said in words. And so I put the words down, and now the reader comes and reads it. Some readers will read that story and then make it come alive in their heads and take away something that is very similar to the thing that I wanted to say but couldn’t put into words. You can call that a success story. But some readers will take away something that’s entirely different from the thing that I was trying to say with words but couldn’t be expressed in words. They’ll take away something entirely different. And they’re not wrong. That’s the thing. They’re not wrong. And I didn’t fail either.
It’s one of those things where I put out a piece of art, and there’s an artistic activity engaged in by the reader who takes away an entirely different kind of unverbalizable feeling or sensation or idea. And it’s something that they created with the help of my words. And in that case, it’s still a beautiful aesthetic experience in me writing it and in them reading it. But it’s not a communicative act per se in the traditional sense. And I think a lot of art is about that kind of thing. So much art is based on this kind of miscommunication, if you will.
All reading is, in some sense, misreading. It’s both a thing that is beautiful about literature, but also something that terrifies writers and annoys readers for a lot of it. It’s just that we’re sort of used to it now, and we don’t view it as strange. But it really is strange. It’s, you know, when we’re telling a story, the writer is saying something that they know cannot be said. And the reader is reading something that they know that cannot be read, which is wonderful.
So if I were to just crudely classify one of those things as evocative, so the things you, the work you put on the page evoke some concepts or feelings in the reader that maybe you didn’t have access to when you were writing it. The other case where it’s very precise, you know exactly what you’re trying to convey to the reader, and maybe you succeed for some subset of readers. Is there any balance? Like you get more pleasure when someone comes up to you at a book reading and they give you an evocation, which is totally different than what you had in mind when you wrote it, but it’s beautiful. Or do you like it when you succeeded in exactly conveying some precise thoughts or feelings to the reader? How do you balance those things?
Oh, that’s beautiful. That’s beautifully put, Steve. I really like that. I think most writers, well, I really shouldn’t say that. I think some writers would feel the greatest pleasure when they meet somebody who managed to take away exactly what they meant to evoke. The thing that they couldn’t put into words, and they can tell that that thing has successfully migrated into another person’s head and they could both see it. It’s sort of like I see a dragon and I described it and the other person managed to see the exact same dragon that I saw.
And that would be wonderful and lovely and incredible. I would be lying if I said that didn’t appeal to me, and I’ve had experiences where somebody came to me and said, in fact, this is one of the things that sort of moved me a lot. You know, they read one of my stories, they said, you know, I don’t think I had that experience at all. And I don’t think I am getting out of this what you wanted me to get out of it. Or maybe, maybe I did. But here’s what it got me to think. And then they explained to me this very, very beautiful image of what they were able to get out of it.
And I was deeply moved because now it allowed me to see my own story in a totally new light. It’s sort of like you go out there and build something and then somebody else found a different use for it. It’s sort of like the folks who invented the Internet or the web intended it to be this wonderful platform for education, for learning, etc. And now they see that people are using it to exchange selfies and to share memes and to argue about politics. I think for some of these, they would look at it and say, you know what, that’s not what I had in mind. But this is also really wonderful.
It’s not something I could have imagined that people are using it to do something that is wonderful, empowering, very human. And I love it. And that’s kind of how I feel about it. I wrote the story with a particular kind of idea in mind that I wanted to evoke. And people saw completely new things in it. And now I can see them, too. So that’s very cool. I tried to talk about one dragon, but people got to show me that there were unicorns in it, too. And that’s wonderful. You get to see that, too.
Now, independent of what is evoked in the reader’s mind or communicated to the reader’s mind, is there a sort of separate aesthetic of just the prose itself where you look at a sentence and you say, that is a perfectly constructed sentence? And I actually maybe care less about what the reader thinks about it. But I just feel aesthetically like a sculptor. I made a, I constructed something really beautiful and perfect. Is that a big part of your effort in writing?
Oh, absolutely. I mean, I sort of tell my students in the MFA program that this is ultimately what every artist is really trying to do. You know, when we talk about artists improving their craft, what are we really talking about? What we’re talking about is this idea that the artist has a vision in their head about something very beautiful that they want to capture in some sense. And they know that the ideal dragon in their head is impossible to fully capture. So what you want to do is to execute something as close to that as humanly possible.
The trick here is to execute that thing to the point where you can look at it and say, this is the best I can do to evoke that. This is me doing my very best to evoke the vision that only I got to see. Now, that involves craft. And in the case of writers, what it means is you have to actually invent your own language. I’m a big believer in this idea that anybody worth reading has to actually ultimately invent their own language. People don’t use something generic called, you know, modern, contemporary American English or whatever. That’s not what they do when they’re writing.
Stephen King has a very unique way of constructing his sentences, of using his images, of the way he uses quotations and lyrics and so on in his work. That’s his unique language. Somebody like Emily Dickinson was not writing in some generic 19th century American English. She invented her own language. These writers are so distinct to us precisely because they managed to improve their craft to the point where they actually invented their own language. Their own unique way of constructing sentences and pulling together images and contrasting things and using parataxes and other tricks of the trade to give the reader a particular, unique experience of language that they then try to interpret to see the thing that cannot be setting words.
So, I think that’s a huge part of it. I mean, I think every writer really is trying to invent their own language. This may take them a lifetime to do, but, you know, they measure their own success by the degree to which they’re able to do so and therefore execute their visions.
So, these somewhat abstract and very aspirational things that you and I are discussing now, do you feel that in the particular genre of science fiction and fantasy, maybe there’s more emphasis on creating a story with, you know, fantastical elements or whiz-bangy kind of elements and less emphasis on these really high aspirations? I feel like there’s a dichotomy there in being, I don’t know if you’d call it a genre writer, but maybe you can talk about it.
Yeah, I know what you mean. I think some readers certainly expect that of genre fiction, and I think some writers, to some extent, expect that of genre fiction as well. But again, I go back to Le Guin, who was, you know, one of the greatest critics, and she basically had very little patience for this sort of thing. She wrote this beautiful collection of essays, criticism of sci-fi and fantasy, called Language of the Night.
In it, she, I’m not giving you an exact quote, I’m paraphrasing, but she says something to the effect of, in art, the highest is the standard. You can’t aim for trash. And she had very little patience for genre readers who expect trash. She says, you know, some genre readers say, we like the trash. We don’t want you to bring in your literary ideas and muck up our trash. We want the trash. That’s what we want. And she says, that’s ridiculous. They don’t actually want that. This is a kind of defense mechanism, because, you know, people don’t take genre literature seriously, so they act as though they don’t care either, and they want to defend their turf that way.
This is nonsense. People would say the ideas matter more than craft, than the literary craft. That’s nonsense. It’s ridiculous. If your ideas are good, you cannot convey those ideas unless you actually have the craft to execute them. That’s the whole point. So I 100% disagree with the notion that somehow in these genres, your attention to craft is somehow not as important. Now, I will also say that these genre definitions are very silly to me. I’m on record as having very little patience for genre definitions because people often argue over what exactly what I’m writing, you know?
So one of the stories that I wrote, the one that you were referring to, The Paper Menagerie, when sometimes people come to me and they’re like, I don’t understand how the story won any awards because it’s not sci-fi. And I said, of course, it’s not sci-fi. What made you think it’s sci-fi? And they’re like, well, isn’t it a beautiful award for sci-fi? No. No, it’s not. You can go look at it. It’s not. But the fact that people feel so defensive about genre boundaries has always mystified me. Who cares?
This is a very weird thing to get hung up about. If you don’t want to think of what I do as sci-fi, I don’t care. It’s not like I actually want to belong to your definition of sci-fi anyway. And if you don’t think what I do is epic fantasy, I also don’t care. It’s not as if I wanted to fit into your definition either. As I mentioned earlier, my aesthetic idea is to express and to engage with this idea of technology. Technology defined as all tangible manifestations of mental, of human mental patterns.
That to me is the most interesting thing. The way humans as a species are capable of imagining things and then instantiating those ideas in the real world. That to me, no matter how many times I think about it, study it, practice it, it never fails to take my breath away. It is the closest that we can approach to being divine. To imagine something and then to make it come true. The universe had no airplanes. And then some people imagined things called airplanes. And then some people actually made them real. And then between having the very first powered flight and humans setting foot on the moon, barely six decades passed.
No matter how cynical you are about technology, you cannot help but just wonder at the fact that our species was capable of doing this. That by itself is the greatest epic poem that has ever been written. Nothing will come close to that, that humans achieve this. All of my works, whether you call it fantasy or sci-fi or magic realism or what have you, are obsessed with this idea of technology as manifested human mental patterns. So, you know, if anything, if I want to coin a genre for what I do, it’s tech-fi. Whatever I do, ultimately it’s about technology in that sense. This broad sense of the realization of human imagination.
Very nice. This might be a good point to segue into one of the topics we said we would discuss, which is artificial intelligence. I’m curious what you think of human activities to create artificial intelligence today, but also its impact on art and storytelling. So, wherever you want to take it.
Oh, wow. Yeah, this is something that I’m absolutely obsessed with. So, I think, you know, a lot of the contemporary debate about artificial intelligence and arts, to me, sort of misses the point. I don’t think we need to argue over whether what we have now is intelligence or not. These are not interesting debates to me. Defining words has never been all that interesting to me. So, what exactly is interesting to me? Well, what’s interesting to me are some basic assumptions and some implications to be drawn from those assumptions and how reality of what we’re actually witnessing.
So, one of these fundamental points I want to make is that I don’t think I’m a materialist. I don’t think there’s anything mystical or non-materialistic about human cognition or intelligence. So, the fact that we are conscious and we think is the result of materiality and there’s a real material substrate for all of this and cannot be divorced from it. So, given that consciousness emerges out of materiality, there is no, I can see no obstacle to the idea that we will eventually create true artificial intelligence, whether it’s by imitation of humans or simply by finding, discovering another way in which intelligence can be materially embodied.
So, I don’t see any theoretical objection to the idea that artificial intelligence is possible and can one day be achieved. Having said all of that, I do not believe large language models are somehow the way to do that. I think this is an approach that has a certain place where it will plateau. And I think that the idea that we will reduce human intelligence to just some version of a large language model is very flawed. I think this is so different from all prior attempts to map the human mind to the closest technological thing that we’ve invented. You know, when we had telegraph networks, we thought human minds were like telegraph networks.
When we had the earliest analog computers, we thought the human mind is probably like that. And then people later on started talking about how because we solve problems using computers by using this set of rule-based symbol processing, then the human mind must work exactly the same way. I think this is all a very common pattern. And I don’t think we need to be deceived into thinking that this very fancy, shiny toy we have is the same thing as a human mind. So obviously, it’s not a human mind. I also think it’s important to sort of understand that we don’t know everything there is to know. I mean, when I was in college, right, this current approach of building AI by throwing a huge amount of data into neural networks was laughable. Nobody thought that was going to succeed.
The fact that these LLMs have succeeded beyond our wildest dreams is shocking. And even now, we don’t fully understand why they work. I mean, there are literally, I mean, it’s sort of, it’s really comical. I mean, Apple Research put out this paper explaining basically that they don’t believe LLMs do reasoning in the sense that humans reason. And the paper ended up causing a huge amount of controversy because there are some people who say, this is so utterly obvious and banal, why is this even a paper? And then other people who are like, this is completely nonsense. People don’t understand what reasoning is. Humans reason exactly the same way as LLMs. The fact that people who are actual experts in the field cannot agree on something so fundamental tells us that there’s a lot of things going on here that we don’t understand.
Then I will say that that’s actually a fairly common pattern where the technology has outstripped the science. I think one of the things that people who don’t work in tech often sort of misunderstand is the idea that they think science always guides technology, that science is always ahead. In some cases, that is true, but not always. There are plenty of instances where the technology has outstripped the science, meaning we’re capable of doing something and building something and it works. But we don’t understand why. In fact, I would say historically that has been more common than the other way around. Technology has a much older history than science. And for much of human history, we did things and got them to work without understanding why.
That’s where we are now with this current instantiation of AI. It’s a lot of interesting things are happening, but we don’t fully understand why. Okay. So what does this have to do with the arts? Well, because I don’t think current AI is anywhere near achieving the state of separate consciousness or independent consciousness or, you know, just having a real kind of sentience or sapience in the universe. I don’t think of them as separate alien intelligences. I will say that when we achieve that state, obviously, I will be super, super fascinated in learning about their art.
You know, the idea of a toaster being capable of writing a novel would be incredibly interesting to me. I would love to read a novel written from that perspective and giving me a sense of what it’s like to experience the universe through that kind of perspective. But we’re not there. And I don’t think we’re going to get there anytime soon. So where are we then? Well, right now, because AI is not a separate intelligence, it is just a tool. It’s no different from any other tool we have invented in our long, illustrious history. And we’ve always invented new art forms based on new tools.
I mean, people don’t sort of think about this, which is that the arts have evolved technologically tremendously. We don’t enjoy the same art forms that people used to enjoy even a couple hundred years ago. I mean, think of how many art forms have been born in the interim and died out. You know, player pianos were a thing. They no longer are a thing. People used to go and watch silent films and have a pianist play the accompaniment in front of them. We don’t do that anymore. People once enjoyed things like Tableau Vivant. We don’t do that anymore. People used to listen to actual bards recite long poems. And we don’t do that anymore.
Technology moves and new art forms are invented based on the technology and old art forms die out. There’s no reason to think that the novel, which has just a few hundred years of history, is somehow a permanent art form. It won’t die out, I’m sure. There’s no reason to think that cinema, which has, you know, just about a hundred years of history, will be there forever. Which is why I get so frustrated when we’re talking about art and AI and all people can talk about all the examples, all the so-called exciting examples that Silicon Valley brings out are just basically machines imitating humans doing what humans already do.
Who cares? This is as boring as the earliest films, which are just, you know, pointing a camera at a stage and having people act out a stage play. That is not interesting. It only became interesting when cinema became its own language and discovered its own way of storytelling independent of the stage. Right now, all this idea about AI writing novels and creating movies, I find these incredibly boring and not interesting at all. And I believe this is a transitional stage and nobody will care about this sort of thing.
I certainly see no reason to be worried about it. What will be much more interesting, though, is when the AI equivalent of cinema comes about. The moment when filmmakers realize that a movie is not merely a stage play that happens in front of a camera is the moment when something truly interesting has happened. I think we’re not anywhere near there. The moment when we realize AI can craft entirely new narrative forms, new visual forms, new audiovisual forms, new interactive forms, new forms of art that are not mere imitations of what we already have, but actually new things that cannot exist without AI. That’s the moment, I think, when something truly beautiful will take place.
In terms of AI being employed as a tool for people who are writers, imagine you had a student in your MFA class and that student was secretly, every assignment that you gave her, she was actually going back and making a pretty good prompt, but then just literally taking the output of a really good LLM and turning it in. Could you imagine that you would be impressed and say, hey, this is one of my top students, she’s really got talent, she could go somewhere? Is that a plausible scenario for you or is that implausible?
Oh, it’s entirely plausible, but it’s also meaningless. That’s not something I worry about. Because when you’re trying to get a degree as an MFA student, what you’re trying to do is to discover your own language to practice that craft. If you end up fooling people by using AI, then you will have done nothing because there is a fundamental disconnect here. Number one, I don’t really think that the sort of thing that comes out of large language models on its own is interesting. For them to be interesting, they have to be guided by a human consciousness. A human has to say, this is the thing that I’m trying to do, and now I am using the tool called AI and shaping it until it gets to the result that I envisioned.
That, to me, is interesting. So if you think of the analogy of a photographer, right? If a photographer is just putting the camera there and pushing a button and then saying, this is a painting that I did, nobody would be impressed by that. It’s not interesting. This is just you trying to lie and fool people, which is, by the way, why I’m a big proponent of clearly labeling things. And then I would go much further than some other people are suggesting. I don’t think you should just say, this work used some AI. That is meaningless. You have to actually specifically explain what was used.
This work was spell-checked by AI, which I think every work that we have now has that. So I think everybody should assume that’s true. But this work had sentences adjusted by AI to fit a certain style. That would be an interesting disclosure, and that should be disclosed. This work, here are the passages that AI wrote, which I rewrote. That would be an interesting disclosure. But I want people to make very, very detailed, long-form disclosures about exactly where the AI is used. I think that’s interesting, and I think that’s helpful, and I think that’s very, very useful for people who are trying to decide if they want to consume a piece of work or not.
They need to know exactly how much of the work they’re looking at is representative of a human consciousness engaging with the universe and trying to craft their own language to describe the dragon. And how much of it is somebody with a vision, but without the skill to execute it, and therefore relying on the machine as a shortcut. And how much of it is just somebody trying to make a quick buck with no vision whatsoever and using AI as a way to generate garbage. I think people need to know that and then be able to make their own decisions.
Again, the idea that, you know, somebody won a literary award by submitting AI-ridden work, people say that as though that’s supposed to impress me. I’m not impressed. You just successfully lied. Why is that impressive? People lie all the time. We don’t reward them for that. That doesn’t prove anything.
Anyway, so I do have students who are very interested in using AI as part of their work. And some of them have done stuff that’s absolutely wonderful. I mean, this is not actually submitted as actual work for credit. This is just something they did on their own. But they managed to basically train two separate language models on the output transcripts of CNN News versus Fox News, two separate models. And then afterwards, they were able to trace out a path between the words in these two separate embeddings. So you can take a path from the word Mexico to the word America through the CNN corpus and hit a bunch of different words along the way. And you could trace the words from Mexico to America through a set of different words in the Fox News embedding.
And it’s fascinating to compare these embeddings. That, to me, is artistically really interesting. This is using AI as a tool to see the universe in a new way, to allow us to imagine and reimagine, revision our language in an entirely new way. I found that deeply evocative and a powerful piece of art. There are other students who are very interested in using AI as a revision tool and to clearly label all the parts and to show that a work, a draft that they wrote evolved through this sort of dialogue with AI into something else. And the entire thing we published along with the transcript. That, to me, is also interesting because this is one of those things where you can now examine the process of an author working with the machine, trying to figure out if it is possible to discover your own language and to craft your own unique language by using the machine as a way, as an interlocutor, as a kind of beta reader.
These are efforts that, to me, are interesting because they’re sort of going beyond merely imitating, but they’re actually using the tool to stretch our art in new directions. I think you wrote a novel that’s set in the Star Wars universe, is that right? That’s right. Could you imagine that, like, so that’s, I assume that all that IP is owned by, like, Disney now or something, right? And imagine Disney employing a bunch of AIs to just keep expanding that canon, that universe of the Star Wars history.
And could you imagine, you know, a case where an editor only gave kind of rough guidelines and the pre-existing canon to the AI, but then the AI created some very interesting character, like some early renegade Sith Lord or something, that you, maybe not knowing that it was made by an AI, you read it and you said, hey, this is an interesting addition to that universe. Could you imagine that?
Oh, absolutely. And I wouldn’t have a problem with that. And the way that I would ask listeners to think about this is to think of two analogies, right? One of them is this. It is entirely possible. And in fact, there are many cultural traditions, including multiple traditions in East Asia, that encourages and contemplates and believes in aesthetic experiences evoked by natural objects. So, for example, in East Asian countries, a long tradition of contemplation of scholar rocks is a thing. You contemplate these natural rocks that have been worn down by erosion and they form interesting patterns. It’s entirely possible to have a very deep aesthetic experience of that.
So, if you can have an aesthetic experience with a natural object, why is it any more insane that you can have an aesthetic experience with something output by a large language model? You know, inherently, the idea that you can view something as a piece of art, even though it has no human author, is not that strange. I don’t think that’s particularly problematic. So, that’s not an issue. Now, the other part, which is, what about the fact that, you know, I keep on talking about how human art is about creating unique visions, etc., etc., but this is not that. This is the machine basically iterating on what humans have done and putting on new variations.
Well, that just means that, historically, when we talk about art, we sort of conflate two very separate practices. One of them is art in the sense that we’re talking about here, you and me, saying it’s about discovering a new way of looking at the world, it’s about this new vision, new language, etc. But that’s not all of art. Quite a bit of art is about providing comfort to people and giving people exactly what they want. So, for example, when you’re buying Christmas decorations, you’re buying something that you know exactly what it is, and you’re buying a cliché, right?
So, the distinction is, I have been speaking about art as though it’s about not doing clichés, crafting something entirely new. But a great deal of art is, in fact, about giving people clichés. So, for example, I used to read a lot, a lot of Dungeons & Dragons books. And many of them are not very good. They are formulaic. Basically, you were given, basically, a set of characters and a set of things, and you have to reproduce those clichés. You know exactly what you’re supposed to do. And a lot of what I call C-plus romances fall into the same pattern.
These are romances that are written according to a very strict formula. The publisher tells you exactly what it is you’re supposed to do when you execute it. People do have a craving for clichés. This is also a type of art. Now, I personally think that the cliché-ridden kind of art, the art form, the arts where you’re trying to reproduce clichés, those are arts where machines will ultimately be way more successful than humans, right? In the same way that painters stopped basically painting certain kinds of portraits and certain kinds of still lives and etc., because machines, the camera can do a much better job.
It’s basically why monks stopped copying manuscripts because the printing press can do a much better job. In these art forms where the ideal is the instantiation or the reproduction of some original, machines will always do it faster, better, and they can do it more accurately, faithfully, whatever that means. So, there’s no question for me that a large number of jobs in those parts of the arts can and will be replaced by machines, and humans will continue to consume them, perhaps even larger-over numbers. But that does not mean somehow that the kind of art that I’m talking about, where the whole point is to push beyond the cliché, to say something that does not exist before will somehow die. I just don’t see that. And if machines are capable of doing that as well someday, then I would welcome that. I would be very interested in, as I said, learning about what the universe looks like from the perspective of a toaster. That, to me, would be fascinating. So, to me, there’s no way to lose here. It’s win-win.
Now, of course, you go back to the question of what about people’s jobs? What about people’s jobs? This is a separate question, right? It’s our choice as a society to put the arts into the market, to support artists, basically, via capitalism. If we do it that way, then, of course, we’re going to end up with a bunch of people who are going to depend for their living upon the production of clichés. And if automation removes that, the solution is to figure out how to give people the income they need to pursue their fundamental task of being human.
The question, the answer should not be, let’s keep the machines out so that humans can continue to act like machines. That is a terrible answer. The idea that automation is to be avoided because we want humans to continue to act as machines is a horrible, horrible answer. And I think it’s deeply dehumanizing and incredibly short-sighted and just inhumane, honestly. That’s fundamentally what it is.
Now, there is an objection to my argument here, which is for a lot of people who are doing the kind of jobs that I’m talking about here, it’s not just about money. It’s about a sense of satisfaction, a sense of purpose in doing that work. Because work is not just about making money; work is about sense of purpose, meaning. Again, my answer is, then our society has failed. If our society says that the way for a lot of people to feel meaning and purpose in their lives is to engage in repetitive tasks and reproducing clichés, then again, we have failed.
You’re basically saying that the best way for humans to be humans is to act as cogs in a capitalist machine to reproduce clichés. I say to you, that cannot possibly be the right answer. We need to work harder at giving people purpose and meaning and helping them figure out how to be human.
I think in interviews of yours, I’ve heard over the years, sometimes the interviewer is asking you about some project that you’re in the middle of or that you’re contemplating. And I can see that really a lot of what you’re doing is expressing some ideas that you yourself want to work through that you find deeply interesting and you want to conduct that project. And that will always be there. It’ll always be an expression of yourself regardless of what AIs or other machines are doing.
I agree. And I feel absolutely very privileged that I get to do this. I mean, you know, if you had asked me as a kid, you know, what your life will be like in 40 years or whatever, and then, you know, the revelation is that I get to sit around making up stuff and that I actually get paid for this and that I get to travel around and attend conferences to listen to scientists talk and share their research and I get to teach other people who want to do the same thing. I would be like, this is some sort of weird utopia you’ve invented that doesn’t exist. There’s no job like that. And I’m very lucky that I actually get to do this.
But my point is, though, that I think all of us should get to do this. It should not be just one or two people or 1% or 2% or whatever have you to do this. Why do so many people feel like they have to do something repetitive, cliché-ridden, and not meaningful to basically put their lives on hold as the show Severance shows? You have to sort of live this life in which you’ve separated out your working selves from your real selves so much so that maybe you can have an operation so that you don’t even suffer the pain of having to endure that. That, to me, is a terrible way to live.
Our ancestors, as hunter-gatherers, did not live that way. So why did we put ourselves in this state? We should try as hard as we can to get ourselves out of it. We shouldn’t be trying to preserve jobs and resist automation under the misguided idea that somehow having people stay in their jobs and derive meaning and money from working like machines is somehow an ideal way of living. It’s not.
Yeah, you know, it’s very mysterious to me. Keynes, a long time ago, just said, oh, given that we have economic growth, eventually people will work just a few hours a day or a few hours a week because we’ll have plenty and people could do that. But somehow the way the social dynamics works, only a few people like yourself or myself are willing to allow their sort of inner desire to express themselves and do the projects they want overwhelm, you know, these other things, social convention or trying to maximize the amount of money they make, etc. So it’s all very mysterious to me.
I mean, we’re obviously very privileged. I mean, you know, I don’t want our listeners to think that somehow we don’t understand that we are very privileged. But the fact is we should strive to make our society to grant that kind of comfort and that kind of freedom and that sense of purpose to everyone and not act as though there’s no solution for it. It’s just not. I mean, we humans had, like I said, we are capable of walking on the moon and we can’t figure out a way to give everyone the sense of purpose and to keep everyone fed and healthy. Come on.
You know, my daughter is in college right now and she wants to be a film director. She loves film. She’s a cinephile and she’s, you know, she’s pretty well read as well. And, you know, on the one hand, I think the standard Asian parent playbook would be, oh my God, make sure your daughter goes to med school first before she thinks of…I mean, you’re laughing, but you went to Harvard and Harvard Law School. I’m laughing because this is actually literally me. I can’t tell you how long it was that my mother kept on sending me these articles about how important it is to be a doctor and to go to medical school long after that path was forever closed to me. So, I laugh only because it’s so real.
Now my, you know, I grew up here. My wife is originally from Taiwan. My wife is a professor of literature and film. We’re both encouraging her. You know, I’m just, I tell her like, if you love that, you’re already ahead of everybody else that you have something you love and you want to actually try to do, go try to do it. If it doesn’t work out, we’ll pick up the pieces later, but just don’t deny yourself, go do it.
So, I don’t know, maybe it’s a mistake, but no, no, I, I agree with you a hundred percent. I agree with you a hundred percent. Also, you know, what does success mean, right? Either she will succeed in the sense of achieving exactly what she set out to do and also achieve greater conventional success along the way, or she will have learned some incredibly valuable lessons. So, no matter what, you know, I don’t think she could ever lose by going down this path. There’s no bad outcome.
I agree. In terms of personal growth, there’s no bad outcome. Maybe in terms of, you know, integrated dollars made during her lifetime, sure, but hey, let’s not focus on that as the objective. Oh my God, integrated dollars made, I love that. That is, I’m so sick and tired of the way we evaluate everything by integrated dollars made. I mean, the idea that, you know, you evaluate your education or the degree that you get by that metric to me is just nonsense. I can’t believe we do it that way.
Yep. So, so Ken, we’re an hour in, I want to be conscious of your time, but I can’t resist asking you a little bit about your tastes in science fiction and fantasy. So if you wouldn’t mind just sharing the authors that you admire most, the works that, you know, you most enjoyed immersing yourself in, not necessarily maybe now, but even when you were a kid growing up.
Oh yeah, absolutely. I’d love to learn a bit more about you. There, there’s, there’s just so much great stuff out there. I mean, you know, this is probably the most dangerous question to ask a writer. You ask them what books they like to read and you can expect to be stuck there for hours as they go on and on. So I’ll try to be very brief on this. I like, I like a lot of classical literature because I think what’s interesting is the classical literature, this is the sort of thing that we think we know, like before we ever read Hamlet, we already know what the story is about. And so we think we know what it is, but then you actually read it and you realize it’s nothing like what you expected.
I’m constantly surprised by how there’s just nothing like what the stereotype or the popular image would make you think. Like one of my favorite books, absolute favorite books of all time is Moby Dick. It’s absolutely wonderful. It’s one of those books that was never assigned to me in school. So I decided to read it on my own and I, I had this small copy paperback that I carry with me everywhere. And it took me a year to read it because I just read it very slowly. I enjoyed it so much. I wanted to savor it. And it is the funniest book ever. It’s digressive. It goes all over the place. It’s full of absolutely over the top scenes of satire and just high comedy. One of the funniest books I ever read.
So I don’t think I ever heard people describe it that way before I read it. So I love classical literature precisely for that reason because you end up discovering how different they are from what you expect them to be. So that’s one. In terms of more contemporary stuff, I… Really, really enjoy the stuff that’s not absolutely contemporary. But as I mentioned earlier, Le Guin is one of my favorite authors of all time. Le Guin and who else? Philip K. Dick, PKD. PKD and Le Guin are my two favorite authors of all time in terms of classical world, mainly because again, they write sci-fi and fantasy. That’s very hard to classify.
I mean, now we accept them as core sci-fi, but I think at the time people doubted whether they really were because there’s so much sort of what I call dream-like mysticism in them, if you will. They really go into the collective unconscious and excavate mythological tropes and ideas. And I think they’re very beautiful precisely because they don’t take the idea that science fiction and fantasy are very different. That’s what I love about it. The fact that Le Guin says sci-fi is really just the newest province of that ancient empire called fantasy. I think that’s incredibly potent and very true. Sci-fi is, in fact, fantasy using a different set of metaphors and images. And I like that sort of work.
In terms of my peers, so one author I admired way deal is Tate Elliott, who writes some of the most intricately built worlds ever. I love intricate, beautiful, complicated worlds that are not just easy analogs of the world that we have. They allow you to see the world that we live in, in a new light, precisely because they do things differently. They follow their own logic, they’re consistent, they show what’s this other possibility reality that we just didn’t happen to live through, what that would be like.
Cade has written a great many epic fantasy series. I mean, I’ve written only one, and that took me 12 years to do. Cade writes like a dozen of them. And so she is like 10 times the world builder I am. And I admire her work so much. They’re so intricate, so thought out, and so empathetic. They show humanity in this really complicated, beautiful, nuanced light. I really love that.
In terms of other works that I like, honestly, I could go on forever. I had the opportunity to read a whole bunch of contemporary novels published last year because I served on the Le Guin jury, the jury for the Le Guin Literature Prize. And I got to read some just amazing, beautiful books. And it really makes me realize just what a golden age of the arts that we’re living through. The number of people who are writing, the quality of the works they’re putting out, the diversity of the ideas that are being engaged with.
I mean, you can go onto Amazon and find genres, niche situations that you never could have imagined convincing a traditional publisher to take on, and they’re all out there. I mean, this is such a beautiful moment if you’re a reader. So I’m just drowning in good stuff, which is a great problem to have.
That’s fantastic. Can I just close out? Maybe I want to ask you a question about PKD because, well, I’m a huge fan of his too. And I agree he’s an incredible genius and mining the collective unconscious, I think you said, is one of his amazing gifts. Does the fact that he wrote a lot of his stuff while on speed compromise the quality of the prose itself? What do you think about that?
Oh, that’s a great question. Okay, so I actually think PKD is a great stylist and some of the sentences in Do Android Dream of Electric Sheep are seared into my brain forever. I mean, I still assign the opening chapter of Do Android Dream of Electric Sheep to a lot of my students because I think it’s just an incredible, incredible example of how much world building you can fit into a short scene and how you can craft every sentence to do so much more. Him and Orwell are both excellent stylists. Orwell does something very similar in 1984. Just a huge amount of world building crammed into a very small amount of space. Very dense, beautiful prose.
So, in terms of explorations of alternative modes of consciousness, I honestly think that this is one area where the research is lacking. I think we should try to be a little bit less puritanical about it and try to explore more of the possibilities of how we can, in fact, allow our consciousness to engage in these altered states. Again, it’s a fascinating fact of life that these plants have somehow evolved to have these chemicals that affect our consciousness in such fascinating ways. How did that happen? What is the purpose?
You know, there are theories that perhaps human consciousness actually is rooted in psychedelic substances. And, you know, it’s the whole theory. And I think there is just a lack of understanding, research, acceptance, and exploration into the area. I think this is deeply fascinating, the idea that we can, in fact, access these different modes of consciousness. Again, Le Guin was someone who wrote about that quite a bit. She was writing in the context of cultural practices that value dreaming as a way of knowing, and she wrote extensively about that.
It turns out that some of the anthropological literature she was relying on turned out to be very unreliable. But the fundamental idea is good, that we need to engage with these other states of consciousness and to really know the universe in its full panoply of modes.
All right, last question. The novel, Do Androids Dream of Electric Sheep, and the Ridley Scott movie. Which is more dear to you?
Oh, by far the book. That’s not even a comparison. I mean, I think Do Androids Dream of Electric Sheep is a much more nuanced and interesting work. It’s much weirder, it’s much more deep, it’s much more unsettling. Blade Runner is fine as a movie, but I don’t really consider it to be anywhere near the same level of interest. I’m not nearly as moved by it as I am by Do Androids Dream of Electric Sheep.
Great. Ken, I thank you for your time. It’s been a great conversation, and I hope we get to do this again sometime.
Absolutely, Steve, thank you so much for having me. This was, I had a blast. Great. Thank you.