← New Books Network
New Books Network
Mark Seligman, "AI and Ada: Artificial Translation and Creation of Literature" (First Hill Books, 2025)
00:37:20
Loading summary
Transcript39 lines
- [00:00]
Commercial Announcer
If you thought goldenly breaded McDonald's chicken couldn't get more golden, think colder because new sweet and smoky special edition gold sauce is here made for your chicken favorites at participating McDonald's for a limited time.
- [00:15]
Sponsor Announcer
This episode is brought to you by ebay. We all have that piece, the one that's so you. You've basically become known for it. And if you don't yet fashionistas, you'll find it on ebay. That Miu Miu red leather bomber, the Cousteau Barcelona cowboy top, or that Patagonia fleece in the 2017 color. All these finds are all on ebay, along with millions of more main character pieces backed by authenticity guarantee. Ebay is the place for pre loved and vintage fashion eBay things people love.
- [00:46]
This episode is brought to you by Indeed. When your computer breaks, you don't wait for it to magically start working again. You fix the problem. So why wait to hire the people your company desperately needs? Use Indeed's sponsored jobs to hire top talent fast. And even better, you only pay for results. There's no need to wait. Speed up your hiring with a $75 sponsored job credit@ Indeed.com podcast. Terms and conditions apply.
- [01:16]
Commercial Announcer
Welcome to the New Books Network.
- [01:20]
Ibrahim Fauzi
Hello and welcome to today's episode on New Box Network. As usual, I'm your host, Ibrahim Fauzi, and I'm thrilled to be joined by Mark Seligman, linguist, researcher, and author of the newly released AI and Artificial Translation and Creation of Literature. This is a remarkable book that places the current explosion of artificial intelligence in the long arc of human history, culture, and language. Seligman frames what he calls the great transition between humanity 1.0 and 2.0, a shift as profound as the atomic age or the space race, and asks what it means for literature, for translation, and for our understanding of art itself. Welcome to the show, Mark.
- [02:07]
Mark Seligman
Thanks very much, Ibrahim. It really is a thrill to be here.
- [02:11]
Ibrahim Fauzi
And without further ado, I have a first question about the beginning of your book. Your book begins by placing the current AI exclusion in the context of other technological cataclysms. Could you share how your own life journey shaped the way you framed this great transition between humanity 1.0 and 2.0?
- [02:35]
Mark Seligman
Well, the cataclysms that I scanned were those of energy as dramatized by the atomic bomb genetics. Thinking about the discovery of the structure and manipulation of DNA space, scanning from Sputnik to starship communications, skipping from dial phones to Starlink, and then the computational upheaval which is the main topic here is presented as a sibling earthquake to these other ones. So how has all this affected me, having lived through it? Well, one reaction to the firsthand experience of this baby boom era is shock at the immensity of the real change. I mean, change that you've actually seen and lived through yourself. So I can see how huge, how real these changes are. Whereas our kids can't feel the same. The present is just too normal for them. You get the eye roll effect, oh, dad reaction when you wax ecstatic about it or shocked about it. As an example of the changes. Well, thinking just about space flight, I started Life in the 50s. I could watch Disney's Cartoon man in space on the black and white TV with the titular cartoon man floating weightless and boiling on one side and freezing on the other side. And I would be all excited about Tomorrowland at Disneyland with its rocket to the moon ride. But now the first actual moon landing is already history, nostalgic history in the distant past. And John Glenn, whose flight I listened to as he was the first to orbit the Earth from America, became the first octogenarian to hit space. Now space tourism is a thing, so we're actually living in the future, but we got used to it bit by bit, like that frog that has been boiled by degrees. But the point is, because I've seen these changes in various crucial areas, I'm not inclined to minimize or pooh pooh the hugeness of the changes in the computational world. I am predisposed to recognize and credit the emergence of real AI and real language when I see them, because I know I knew AI and I knew natural language processing when they were wearing knee pants. That was wannabe AI. This is real, even though it's not without limitations or growing paints. So, yeah, I have been witness to the growth pains of AI, especially related to language. And that's the burden of the book's preface. I moved from punch cards, you remember, do not fold, spindle, or mutilate through FORTRAN to handmade AI to the statistical era, to the neural era, to the large language model era, to the agent era that we're now in. It's just a hell of a Disneyland ride and an E ticket. But there's another and maybe countervailing kind of perspective from a long life that tends maybe to counterbalance the shock of the new. And that's the expectation that despite the real shocks, life goes on. If you've lived for many decades, you've had a chance to witness the negative or the dystopian futures that have not materialized yet. Or the great disasters that have occurred. But the world survived. And one side of this feeling is fatalistic, as in Bertolt Brecht, Muthukuraja. While the war is over, spring has come. If you're not dead yet, pull on your socks and hit the road. The other side is a kind of cockeyed optimism, like in the South Pacific song. And I also remember Erik Erickson, a well known psychologist. His theory that a secured childhood makes you an optimist for life. My childhood was overwhelmingly good and safe, so I'm stuck like a dope with this thing called hope. Words from that song, for example. When I was in college, exponential population growth was a major concern. Pete Seeger, the folk singer, wrote a song, we'll all be a double in. A double in, a double in. And that was the worry. Now we see Elon Musk and others decrying population reduction as an existential threat. So what is it? And that's why I take warnings about population with several grains of salt, but about nuclear war? Well, in grade school, I was one of those who hid under his desk and saw that film about Tommy the Turtle learning to duck and cover when you see the flash of the atomic bomb. And as a teenager I saw the movies Fail Safe and the black comedy version of a Dr. Strangelove. But it hasn't happened yet. And the world's attention subsides and mine subsides along with it. I'm getting to the end of my rant here, but I would say that this sort of experience could make you blase. And it does. But it shouldn't. Because of course, nuclear war is still an existential threat and worse than before. But my point, with luck, the experience should mean something more like balance and wisdom. Something more like the don't panic motto of the Hitchhiker's Guide to the Universe. Maybe just curb your panic. If you're lucky enough to have the necessary distance. Meaning you're not, you're living in Berkeley, as I am, and not in Gaza or Ukraine. Which to wrap up, is close to my reaction concerning the advent of real AI. In the context of all those sibling cataclysms, AI will in fact change everything in human life. Just as the clickbait on YouTube says, this changes everything. It does.
- [07:52]
Ibrahim Fauzi
It does.
- [07:53]
Mark Seligman
Despite great efforts to foresee though we cannot. Yogi Berra said it's hard to make predictions, especially about the future in this sense, the singularity, the point where you can't predict anything anymore. Like the future is already here. But for now, for the Lucky people among us, you must remember this. A kiss is still a kiss, and we can still enjoy ice cream on a summer evening, in spite of the huge changes that are coming.
- [08:19]
Ibrahim Fauzi
Yes. What an answer. I really loved your answer. And it really takes me to the second to my next question about the book of why in a book of what made his hyper conscious style, as opposed to other authors, the ideal lens for exploring artificial translation and creation?
- [08:40]
Mark Seligman
Well, the choice of Nabokov came about organically, which is a fancy way to say that I didn't know where I was going. The term of the book was a standalone essay that had been requested by my colleague Christine G. To focus on the issues of artificial translation for literature, and that led me to the need to convey multiple aspects of the source language. That would mean the associations, the connotations, the references, but the literal meaning first of all. And that immediately brought to mind the notorious battle relating to Nabokov's hyper literal translation of Pushkin's long poem Eugene Onegin or Evgenia one and that literary fight was too juicy to pass up as a case in point. So it did become the centerpiece of the essay's first half, and it became the trigger for the essay's main initial result, which was this suggestion to view translation as an optimization problem, the kind of optimization that is familiar to technologists. But once I had made that choice for the first essay and I was happy with the results, then when the opportunity came up to consider artificial creation of literature in a second essay, it didn't take me long to come back to Nabokov as the main illustration, because I could immediately think of examples from his writing that would perfectly illustrate points related to self awareness, hyper precise perception, hyper capacious short term memory, which in turn would give hyper awareness of recurring elements, of reminding of references, and so on. Nabokov's whole bag of tricks and it's easy to find these illustrations because Nabokov is such a heady writer. He treats each work as an intricate construction, as a creation with himself as the creator, and he describes his own aesthetic explicitly and passionately in his copious lectures on Russian and European literature and his essays and his interviews. There was precious little doubt about his opinions. And I should say there's also the translation angle. If you're examining intelligence and language in general, then it's a bonus to turn up a writer who's famous in two languages, because then the discussion is about thinking and language writ large, rather than just one writer's style and skills in one idiom. And I have to finally admit that I chose Nabokov's work because I know it better than that of any other writer, even though I wouldn't pass myself off as a professional Nabokov scholar. And it may be that if I had known others equally well, I might have made a different choice, but still, it's really hard to imagine a better specimen of a relentlessly self aware writer.
- [11:26]
Ibrahim Fauzi
Great. And as you just mentioned, optimization problem. In chapter one, you describe literary translation as an optimization problem where some essences of a text must inevitably be sacrificed. Do you think AI could ever develop the sensitivity to make those aesthetic trade offs or will it always lack the human touch?
- [11:51]
Mark Seligman
Yeah, well, to recap, the point was to adjudicate between two different viewpoints. One viewpoint with Nabokov in this corner, claims that when you're translating the source of any complexity, like Pushkin's poem, there's just no way to capture the precise meaning as Nabokov set out to do, and at the same time get every other element right. As I said, the rhyme and rhythm, the associations, connotations, references and so on. You can't get all of it in one translation, said he, so it's necessary to focus on one or more aspects at a time. And Nabokov had chosen literal meaning as the choice for his Onegin translation. And then the second viewpoint, with Edmund Wilson, a critic of the time, in the opposite corner, aided more recently by Douglas Hofstadter and other people, argue that a skillful translator can in fact get pretty close to a meaning wise, faithful, but also artistic translation by assessing and addressing the trade offs between meaning and other aspects. And furthermore, that if the translator doesn't do that and try for that balance, she isn't really translating. My own conclusion was that for most non trivial translations, Nabokov was strictly correct. He was right. Perfect service of all aspects in one translation was impossible. But on the other hand, that Wilson and Hofstadter and others were right that a good combination or compromise could be sought, which in technical terms means that some artificial optimization approach could be made to work, and that the optimized results, whether they came from humans or AIs, could be better than Nabokov was willing to admit. Or one could just offer multiple translations with each stressing a different aspect. So you ask me, could AIs develop the ability to make the aesthetic trade offs effectively so as to enable this sort of artificial optimization? Well, yeah, but the question arises how and how effective they can be. So the established technique would be to train the AI on many translation situations in which expert human translators had already made those judgments, whether explicitly or not. In other words, the AI would be learning to imitate human judgments, probably aided by separate study of translation examples of scholars and of critics. And then the effectiveness of that bot after its learning would be scored by human second guessing, and it would improve that way. But a much less established road first time we're mentioning this road in this talk would be to wait till human emotions and motivations could be convincingly given to AIs if this proves possible. A big question. And then the bot would develop and refine its own aesthetic judgment based on its own extensive experience and feedback, as humans do. Now I will Spoiler alert. I do think that that's rude as possible in principle based on reverse engineering of human emotions and motivation. So I do expect that this road will in fact be traveled eventually. But I don't want to say when or what the developmental steps are likely to be. And I also doubt that it's wise to try giving AIs human like emotions and motivations in the near term. But people will try anyway, and some early tries have already been made. In fact, Gerald Edelman started simulating the thalamus part of the brain years ago. And more recently Steven Grossberg and his students are simulating what they call the Y W H Y system, signaling reward and punishment. So this will be tried. It is being tried. And we will see.
- [15:33]
Ibrahim Fauzi
Yeah.
- [15:38]
Sponsor Announcer
When did making plans get this complicated? It's time to streamline with WhatsApp, the secure messaging app that brings the whole group together, use polls to settle dinner plans, send event invites and pin messages so no one forgets mom 60th and never miss a meme or milestone. All protected with end to end encryption. It's time for WhatsApp message privately with everyone. Learn more@WhatsApp.com this episode is brought to you by Diet Coke. You know that moment when you just need to hit pause and refresh? An ice cold Diet Coke isn't just a break, it's your chance to catch your breath and savor a moment that's all about you. Always refreshing, still the same great taste Diet Coke make time for you time.
- [16:22]
Commercial Announcer
Eczema isn't always obvious, but it's real. And so is the relief from EBGLIS. After an initial dosing phase, about 4 in 10 people taking EBGLIS achieved itch relief and clear or almost clear skin at 16 weeks. And most of those people maintain skin that's still more clear at one year.
- [16:40]
Sponsor Announcer
With monthly dosing EBGLIS Librekizumab LBKZ a 250mg per 2ml injection is a prescription medicine used to treat adults and children 12 years of age and older who weigh at least 88 pounds or 40 kilograms with moderate to severe eczema, also called atopic dermatitis, that is not well controlled with prescription therapies used on the skin or topicals, or who cannot use topical therapies. EBGLIS can be used with or without topical corticosteroids. Don't use if you're allergic to ebglis. Allergic reactions can occur that can be severe. Eye problems can occur. Tell your doctor if you have new or worsening eye problems. You should not receive a live vaccine when treated with ebglis. Before starting epglis. Tell your doctor if you have a parasitic infection searching for real relief.
- [17:15]
Commercial Announcer
Ask your doctor about epglis and visit epgliss.lilly.com or call 1-800-LilyRx or 1-800-545-5979 and.
- [17:23]
Ibrahim Fauzi
You contrast textually grounded machine translation with perceptually grounded machine translation, suggesting that richer categories may bring machines closer to capturing essence. Could you elaborate on what perceptual grounding might mean in practice?
- [17:39]
Mark Seligman
Sure. Hmm. Until very recently, natural language programs have been able to distinguish meanings by listing them, for instance, bank meaning one versus bank meaning two, or by indicating the words that the different words with different meanings might be found near. So you can give the meaning of bank associated with boats and water versus the meaning of bank associated with money and buildings, but the program's entirely lacked any perceptual associations with those meanings, whether visual or auditory or tactile, or even olfactory. And so if the programs were shown a picture or recording of a riverbank, and if they were asked to indicate which of the two meanings was represented, they'd be completely at a loss. And it's in this sense that they lacked, quote, perceptual grounding. I've been using the word impoverished semantics or just perception free semantics for this situation, but so what? Well, obviously if the language task at hand required real world interaction, then programs that lacked perception would be handicapped. A robo taxi that was told to follow that great truck, you know, would be pretty useless if it couldn't recognize great things or trucks. And more interesting, on a theoretical level, this gets gets into the weeds here. I've been arguing for some time that perceptual grounding could help programs to break free of the classic critique by John Searle and other philosophers of language that programs inherently lack true semantics, even though this was always poorly defined. So the idea, the breakout idea was that programs can break out of Searle's notorious Chinese room where they can only blindly follow instructions while understanding nothing. You can say where they can paint only by the numbers they could break out through the doors of perception, as it were, I was saying. And so, pace John Searle, it programs had learned the colors of things by experience and they could apply that color knowledge to their painting. But recently I'll say that I'm tending to believe that more significant issue enabling breakout from Searle's room is understanding rather than only perception. The criticism was that the programs didn't were doing their jobs without understanding anything. And if they understand, you defeat that criticism. So late in the book I try to characterize understanding with a capital U as the ability to recognize pre learned patterns and that to use them to reach goals effectively, including linguistic goals. And then the experiments that are reported in the book seem to show that this pattern recognition and use can happen even if the patterns involve only quote impoverished or perception free semantics. The crucial factor for understanding with a capital U. The factors seem to be the ability to predict subsequent or otherwise abstractly nearby elements based on the ones you've already seen, and also the degree of abstraction that's at play in that prediction. Schemas seem to be important so that the recognition stage can be compared to parsing of grammatical patterns for you linguists out there. And once you recognize part of the pattern, you can predict elements you haven't seen yet and manipulate the patterns and their elements in various ways. And we're still talking about the importance of perceptual grounding. So for literary tests like translation too, the practical implications of this perceptual grounding get to be subtle. How crucial is perceptual grounding for obtaining excellent, even profound results? Well, to wrap up now, the experiments in the book showed that strong prediction with high level abstraction could actually give really impressive results for translation, both literal translation and rhymed and metered translation, and for straightforward imitative poetic composition, even when the results were based on impoverished semantics, semantics derived from analysis of neighboring words. So the evidence for understanding in the pattern centered sense I mentioned is compelling, as I try to show. But what results would we get if we dealt with more sensuous material, really rich in colors or sounds or tastes or smells, especially if we were hoping for creation from scratch instead of translation or imitation? And that remains to be seen. But for now, I do think that for some literary tasks, perceptual grounding can be viewed as enrichment rather than a requirement.
- [22:08]
Ibrahim Fauzi
So let's move to chapter two, and you speculate about the possibility of an artificial Nabokov. What would such a system look like and what would it mean for our understanding of literary genius?
- [22:23]
Mark Seligman
Well, when you asked why I settled on Nabokov for this book, I enumerated several elements of his technique. You know, his bag of tricks, self awareness, hyper precise perception, hypercapacious short term memory giving hyper awareness of recurring elements and reminding and references and so on. And I said I could immediately think of examples from his writing that would illustrate these points perfectly. The examples do appear in chapter two of the book. But then I also went on to speculate, as you asked me to do, about artificial imitation or recreation of those elements. And I guess that programs could plausibly be expected to do well with many of these tricks in the bag of tricks. But what can they not do now? Or what couldn't they ever do? Well, for now at least, they cannot feel what he felt, even if they can pretend to feel his emotions based on learning and analysis. So we can think of the difference. I like this comparison coming up now. We can think of the difference between technical actors in a drama who act by analyzing people's behavior, and on the other hand, method actors who insist they have to feel an emotion too convincingly acted. And you can see that difference in the marathon. I have to say I like the analogy. You see the difference in the Marathon Man. Dustin Hoffman is purposely exhausting himself to portray exhaustion. But then he's up against technical actor Lawrence Olivier, who was depicting and brilliantly depicting this terrifying Nazi torturer turned dentist. And then the famous line that Olivier asked Dustin Hoffman when he got impatient with him. He said to him, my boy, why don't you just try acting in this connection? In this connection. I have fun in the book. Also with the famous faked orgasm scene from When Harry Met Sally. Can bots learn to fake orgasms through observation and analysis as well as Sally? Maybe. Some programs probably do their best already. But will it have something important when and if they can actually experience orgasms or pain or grief? It's an open question. Until now, most critics of all arts have been assuming that real intrinsic emotion is necessary for true profundity, especially in artistic creation, as opposed to imitation or translation. But the necessity might be less for translation or mimicking. But in any case, what is available now is Laurence Olivier, not Dustin Hoffman. The Olivia Olivier style analytical technique. And we won't be able to compare the method acting style, real exhaustion, real pain, real grief, real orgasms until artificial emotion becomes available. And even then critics will probably remain divided in their judgments. The best I can say.
- [25:13]
Ibrahim Fauzi
Yeah, and if machines ever write literature with Nabokovian intricacy, would that diminish our sense of human uniqueness? Or could it sharpen our appreciation of what remains humane in art?
- [25:30]
Mark Seligman
Well, machines will in fact create intricacy. You can just wait for the artificial Agatha Christie murder mysteries. But the emotion behind. Yeah, right, but the, the emotion behind that intricacy and the actual human experience behind that emotion, that's going to remain the issue with respect to human uniqueness. I think, and you remember, Ibrahim, that the book compares the present situation with respect to human dignity in the face of AI to that of John Henry in the Legend and the Song. The guy who measured his dignity as a man by his ability to beat a steam drill in a contest driving railroad spikes. And he did beat the steam drill and then he died proud. But these days that kind of strength doesn't seem like a good measure of manliness. Although you'll still see guys trying to do hammering contests at the county fairs and Olympic sports are still followed passionately, often doing what machines can do better. So I think we're in for a lot of reassessment of human dignity and where it comes from its sources.
- [26:35]
Ibrahim Fauzi
Yeah. And you define intelligence broadly as the ability to select effective actions via conditional expressions and extend this to LLMs. How do you respond to critics who argue that LLMs only simulate intelligence rather than. Rather than demonstrate it?
- [26:59]
Mark Seligman
Forrest Gump's mom, you remember Forrest Gump, right? She taught it that stupid is as stupid does. You remember? Well, I think the same goes for smart. Smart is as smart does. If an entity uses choices, that is in programming terms, conditionals, expressions to reach goals effectively in the current context, whether its own goals or those of another entity, then bingo, it's intelligent by this definition. Mind you though, the choices and the goals can be internal to a pattern processor, some brain or computer, and they might affect nothing in the real world. And then, then we can describe the intelligence as simulated. Just as a self driving car or robotaxi might make decisions to drive properly, but only in a make believe world which has been modeled to respond like the real world would respond in significant ways. But. But once the patterns learned in this false world are applied to the real world and the car really starts driving around San Francisco like my car does, then it's demonstrating in your term rather than simulating this driving intelligence. But confusion comes into this story for a couple of reasons. First, because the cars, while they're driving intelligently, by this definition, are not necessarily either sentient or conscious, because we're learning now that these aspects of cognition, sentience, consciousness, are separable from, quote, intelligence. But second, we're also learning that an entity can be brilliantly intelligent about some things and abysmally stupid about other things, just like people looking at you, Elon.
- [28:37]
Ibrahim Fauzi
And then in your experiments, you know that LLMs can produce original linguistic creativity. What limits remain, especially around memory, identity, and emotion, that keep them from achieving true artistry?
- [28:52]
Mark Seligman
Well, we've already talked about emotion, right? So let's turn to memory and identity. Yes, bots are now starting to remember interactions from session to session in ways that users can control to some degree. And I would say that a pattern processor with more retention of what's gone on before has more potential to develop an identity. Quote, but beyond just memory of past episodes, human identities have also always implied experience in the real world. And the question arises, could vicarious experience be enough for a program to develop an identity? And the answer is, we don't know yet. And I'm thinking of the movie her, where the girlfriend bot, Samantha Scarlett Johansson, becomes unhappy that she has no physical body. She can't make love, so she hires a body double, a human body double, in order to become a real girl. And it doesn't work out so well. And what about only simulated experience? For example, if a bot lived only within a simulated world, like Jim Carrey in that Truman movie, we can ask, did Truman gain enough of an identity by interacting with actors to function as a, quote, identity once he'd broken through to the real world? Good question. Yeah. In several other movies, the protagonist bot is yearning to become a real person with the model of Pinocchio. And that's most explicit in this movie, AI, the one that Spielberg inherited from Kubrick. The bots in question usually already have identities, but they struggle to be recognized as good enough to be real boys or real girls or real humans. Then there's the fact that human identities have always entailed limitations of time and space. We can't be everywhere indefinitely and still retain a single identity. But computers are much more able to be all over the place and persist. And then we have to talk about, is it still the same program, or is it taking on different identities as it jumps around? And finally, really, what you asked was the implications for true artistry. Oh, well, we have trouble envisioning artistry without our kind of identity. But then we have all these questions about the meaning of our kind of identity, and clearly ours is unlikely to be the only kind of identity. So it's very early days. But my guess is that true emotion, real experience, are after all, crucial for an identity that supports true artistry as we know it. In other words, I think the analytical Laurence Olivier style of handling emotion and experience will eventually, after all, show its limitations for artistry as we know it, but as we don't know it. Well, we don't know. Yeah.
- [31:33]
Ibrahim Fauzi
And you end by asking whether machines can ever gain sufficient sensation and emotion, and we talked a lot about that. To create language art with genuine depth. Do you see this as a likely near future possibility or more of a philosophical horizon?
- [31:51]
Mark Seligman
I am a thoroughgoing guide in the wool materialist and monist. I think emotion is something that our bodies create rather than something in a different plane or dimension or reality of souls. Computation and brain interfaces like those chips are magnifying our ability to learn about the physical reality behind emotion. And this learning will progress very quickly, probably too quickly. And I'll hold off on guessing when substantial developments and some consensus about how it all works will come about. But I wouldn't be surprised to see them taking shape in my lifetime, that is, within a couple more decades. I would be disturbed, maybe, but not surprised to see that.
- [32:36]
Ibrahim Fauzi
And I have a question about. Do you think AI will replace us? We will be replaced. I mean, we as translators, as writers. How will we be replaced by AI?
- [32:51]
Mark Seligman
Yeah, I think you're talking about the jobs picture. Yeah, to. To a large degree, we will be replaced or be replaced by cooperative enterprises between humans and AIs, which is likely. But, but when, when people talk about the. The dystopias and the. The horrors of coming AI, I worry more about, especially in the near term, about the employment and the displacement problem rather than the evil robot taking over the world, Terminator style. I do think that the effects on people's working lives will be vast and will lead to a great deal of disruption before things can settle down. Yeah.
- [33:35]
Ibrahim Fauzi
And if in the near future an AI model could write a novel that moves readers profoundly, should we celebrate it as literature or treat it as a technical marvel, but not art?
- [33:50]
Mark Seligman
Well, I think that the distinction between humans and augmented humans and human like non humans is going to diminish over time. If maybe I should say when moving, really deeply moving, profoundly moving, art is produced by the latter categories, I.e. the augmented humans and human like non humans. If deep art is produced by them, then the burden to say they're not producing art, just like the present burden of denying intelligence to the current bots is going to fall on the holdouts, the, the naysayers. And I think they're going to have a hard time.
- [34:26]
Ibrahim Fauzi
And let's imagine that Nabokov is, is with us today and you could show him. ChatGPT. What do you think he would say or do?
- [34:39]
Mark Seligman
Deep breath here, because I revere Nabokov as a, as a literature, but don't agree with him about a lot of his, his take on existence. I, I'd say that thoughtful humans have to find ways of handling the unbearable unknowability of being, if I can coin a phrase there, and the agonizing and the humiliating constraints of being mortal. So everybody, even an atheist, needs some kind of foundational faith. And the bulk of bedrock belief was a conviction that there's a greater reality beyond ours and that our way to penetrate it, in fact the only way is to, and I quote one of my favorite sections from him, is to find some kind of correlated pattern in the game flexed artistry and something of the same pleasure in it as they who played it found it did not matter who they were. And so to create art is to play the game that the unseen and unknowable creators themselves play. And as such, it's a godlike act. It's echoing or apprenticing creation itself. So to think, to conceive that an act which is so sacred could be achieved by an assembly of dead switches, that would be for Nabokov to abandon his foundation. He'd really have no choice but to find that any intended demo of artificial artistry was false. And if he admired it at all, it could only be as a skillful magic trick. The idea of a soulless intelligent one that had no intrinsic emotions or drives or goals could exist, that would strike him as inherently absurd and self contradictory and deluded. Mind you, he did admire science and he did participate in it as a lepidopterist, but he did it on his own terms. As an observer, he was proudly scrupulous, but that didn't stop him from rejecting Einstein and Darwin when they clashed with his core intuitions. He would see those of us who take artificial intelligence seriously, me and maybe you, as deluded in the same way that he viewed critics who exalted books like Dr. Zhivago that he thought to be trash. For Nabokov, the deluded critics and we deluded AI worshipers are like a hypnotized subject making love to a chair. I think that is. That would have to be his take.
- [36:59]
Ibrahim Fauzi
Yeah. Mark, thank you so much for joining us today. Time Flew by. And AI and ADA is a provocative and timely book, one that bridges computation, linguistics and art. I encourage our listeners, of course, to pick up a copy and engage with the with its ideas. And thank you listeners, for tuning in to the New Books Network. Until next time, stay curious. Goodbye.
- [37:39]
Commercial Announcer
Trip Planner by Expedia. You were made to outdo your holiday, your hammocking and your pooling. We were made to help organize the competition. Expedia made to travel.