Artificial Aesthetics (with Lev Manovich)

This is AI Inside episode 53, recorded Wednesday, January 29, 2025. Artificial Aesthetics with Lev Manovich. This episode of AI Inside is made possible by our wonderful patrons at patreon.com/aiinsideshow. If you like what you hear, head on over and support us directly, and thank you for making independent podcasting possible. Hello, everybody, and welcome to another episode of AI Inside.

This is the podcast where we take a look at the AI that is layered inside of so many things in the world of technology, from creativity to productivity to everything in between. We've got a really great show, coming up for you today with a wonderful guest. Before we get there, I've gotta bring in, my co host. I'm, of course, Jason Howell, always joined by Jeff Jarvis. How you doing?

Of course you are. It's good to see you, Jeff. Good to see you, Jason. Good to have you back. Good to have the support of all of our patrons.

Thank you so much for driving much of what we do, patreon.com/aiinsideshow. Gina Jordan is one of our amazing patrons. So, Gina, thank you Thank you. For, contributing and, yeah, contributing to the health of this show on a on a weekly and monthly basis. We literally could not do it without you.

And then also just a quick reminder before we get started that aiinside.show is our website. If you go there, you can subscribe to the podcast. You can also find links to our new YouTube channel, which is @aiinsideshow on YouTube. So pretty much AIinsideshow is is kind of the thing to remember, and you can find us in all ways, shapes, and forms when you remember that. But let's, let's get right in because yet another week where we have an awesome guest joining us, and this has been a little while in the works because we talk so often.

Jeff and I talk on this show about the intersection of creativity and artificial intelligence. I've I've I've got a very creative mind. I like, I'm a musician. I'm always looking for ways in which, you know, creativity is being impacted and influenced by new technology. And I don't I mean, you can't get much more down that road than we are right now with how artificial intelligence is influencing the development of creativity and what it means to be an artist or what it means to be a musician in a day and age when we can use words to create things as opposed to, you know, having to teach ourselves these skills over decades to become an expert.

You know, the the landscape is changing a lot, and so today's guest can really talk to kind of how how society in in this creative era that we're living in is really impacted by the power and influence of artificial intelligence. Lev Manovich is a professor of computer science at the City University of New York, also author of many books, on digital culture, new media, and the theory around it. In fact, Lev, you coauthored along with Emanuele Arielli, I hope I pronounced Emmanuel's name right, artificial aesthetics, generative AI, art, and visual media, and you released it online at manovish.net. Lev, it is a pleasure to bring you on AI Inside. Thank you for being here.

Thank you for joining me. Yeah. It's it's wonderful. And this is this has been a long time coming. Like I said, Jeff, you've had, you've had the word in my ear for a while.

We've gotta get Lev. We've gotta get Lev at some point because he would be perfect to talk about this stuff. So Lev and I are actually, officially former colleagues at CUNY, though, different parts of the, of the empire there. And I had the privilege of reading Len's Lev's new book, artificial aesthetics, as it was in process because he was writing it publicly, online. And it's a really compelling, but provocative work about, AI and aesthetics.

And so, Lev, I'm not gonna try to summarize it, but, but we'll ask you to, your view on whether AI has an aesthetic. Yeah. Thank you guys for having me. So I'm not sure I can summarize the book either because half of the chapters, you know, were written by my co author, half of the chapters were written by me. Another thing which I think is interesting about this book.

So, you know, we met and we started, you know, talking, and my offer, he is a professor of aesthetics at the architecture and design school in Venice, and he lives in Berlin. And we met, like, in 2019 and started talking, realized that we have so many points are common. And then we decided to do this book, and I convinced him, you know, to release the chapters online so we can participate in this cultural moment. But I think what's interesting is we released the first chapter at the end of 2021. Even the second and third chapter with 2022, and then this explosion happened.

Right? So as we know, ChatGPT was released in the fall of 2022, but a bit early in the summer of 2022, Midjourney released the groundbreaking tool for image making. So half of a book was gonna be written almost in anticipation of what's going to come and, this explosion, this moment which really changes everything about, I think, who we are as humans, but what does it mean to produce knowledge, what does it mean to have skills, what does it mean to be creative. Even half of the book, you know, was produced after kind of trying to describe this new medium as it comes in. So to answer your question more directly, there are few pages in the book, right, about AI aesthetics.

Because I think when people see the title, we think of aesthetics as maybe, like, what makes things beautiful, like, the traditional idea, or maybe we think of aesthetics in a more contemporary sense because a few years ago, people started to use the word aesthetics as a synonym for style. Right? So this is statics, this is statics. So people assume that the book is mostly about, you know, how the images or videos you make with AI tools look. You know, and I talk about it, but it's much much larger than that.

Right? So half of book, half of my half of my chapters deal with kind of history of our concepts about art, creativity, etcetera. Like for example, you know, how do we, you know, define the masterpiece? Where does the concept of creativity, you know, comes in? Like, I explained that it's a kind of a modern idea of art as this person who creates things, original things, doesn't follow rest of society.

It's a very modern idea. It's only maybe about 200 years old. And for most of human history, you know, art was about skills. Right? Right?

Ancient, you know, in ancient Greece or in middle ages, there was no concept of art or creativity in contemporary sense. Like, the artist was a craftsman, right, who had a particular skills, and, you know, Plato didn't really respect artists. So partly, the purpose of the book is to educate the people who who interested about AI, who who want to learn about AI, aesthetics, and art to kind of expand and to question our common ideas about art and aesthetics. And then another part of the book is to try to conceptualize creative AI within the history of digital media. Like like, for example, I can make my argument, but we can think of creative AI not necessarily as only something which is part of history of AI, which begins in the fifties, right, in this idea that we're going to simulate human intellectual capacities, etcetera, etcetera.

And now we're able to simulate, let's say, human, you know, aesthetic or art making capacities. But to say, actually, it's a new media which can be followed up as simply a new station with development of digital media. So things like computer graphics, which begin in the eighties and then the web and then proactive media, which begin in the nineties. And when you think about it, right, this generative AI is only possible after 30 years of web. Right?

So with humanity, it has spent 30 years Mhmm. Creating digital content, you know, images, videos, I mean, comic books, you know, texts, podcasts, and also museums and libraries and archives digitizing our cultural heritage, and all the stuff ended up online. Right? And where they can feed it to a computer, the computer can extract patterns, and it can learn to generate, right, new cultural objects, new cultural artifacts with same characteristics. Right?

So it's a very particular kind of development. And then, the last thing I would mention, in the history of AI, right, people talk about the thing which we call AI effect. So whenever the computer scientists and engineers are able to solve some problem, you know, like, automatic translation, or let's say you type in a new phone and your phone can understand what you if you hit the key b. Actually, you wanted to hit the key p. Right?

There's always kind of smart things which our devices do. And as long as as soon as some kind of problem like that is solved, it's no longer thought of as a part of AI. So historically, AI is always something which is on the frontier. AI is always something which is not working yet. And I think the same thing will apply to this moment.

So in a few years, we are not gonna be thinking about ChatGPT and Cloud AI and midjourney and stable diffusion as AI tools. We will simply become ubiquitous. We will simply become, sort of invisible, and whatever is the new challenge, which haven't solved yet, will become AI. So, so I hope I'm not frustrating you too much with my answer. But in a way in a way, like, you know, ironically, I can say, but, actually, the book is maybe about everything else and not the aesthetics about or, basically, it's about different perspectives to think about this moment.

Right? Then going back to ancient Greek and, you know, to to Renaissance and also looking at this moment in the context of, let's say, last 30 years of development of digital culture and also thinking about what will happen next after this moment. Yeah. Yeah. And, also, what what fascinated the book too is is is the, the relationship to the viewer, to the public, to, to the the the AI anticipating that and where that goes.

And and I agree with you. And I think that that inevitably with technological tools and creativity, the tools start as technology and the technology feeds into the background. And and that's where we go. I'd love it if you'd talk for a minute about your work as an artist collaborating with AI and, how you work with it, what you think about it as a collaborator, and also what reaction you get from the art community to that. Yeah.

Thanks for asking this question. So my work as a writer, a critic, and theorist is, you know, a bit different from, let's say, most of our of my colleagues because I was originally trained as an artist. Since age of 13, I was trained in visual arts. When I went to NYU film school, I worked in the computer, graphics, visual effects industry for 6 years. And when I was actually teaching visual art for 20 years, you know, and when I kind of was hired, they offered the job at the Graduate Center, and they said, well, you are, you know, we we really want you, but we don't know where to put it, because we don't have any art at the graduate center.

So we have to attach it to some program. I said, so what are my choices? And they actually said, well, we can put you into our history, or we can put you in computer science, because we think you're gonna belong to both places. And I thought, okay. Well, I'll choose computer science.

But just remember that I only take I only took one computer science class in my life. Right? And now I'm now I'm teaching I'm teaching p yeah. I'm teaching PhD students in computer science, and I do know more than them. So working on this book, right, and coming up with all these ideas kind of went hand in hand with almost daily using this, AI image making tools.

And, so the book is also about these experiences, reflecting about them. And I will just say a couple of things. So on the one hand, for me as a artist, right, who used all kinds of media in his life, you know, starting with pen and oil paints and then, you know, computers in the ages. This new development is absolutely amazing. It is as mind boggling.

It is as exciting, and I think it is as consequential both for my own practice, but I think also for everybody else. At least as we develop it to web and the 90s, and probably at least as big as we develop it to photography and, like, to send you printing press, etcetera. So it's huge. So what does it mean, right, to be creating a a UI? Well, so one very practical result is that you can basically try and realize many more ideas, right, in the same amount of time.

So it's it's like it's more of having thousands of very skilled assistants. Right? And you could try this thing and that thing and that thing. So let's say if I make some realistic drawing with lots of detail, maybe it takes me 50 hours to make these drawings. Now I can make, you know, 5,000 of these drawings, you know, in a few minutes.

So that's one side of it. But then there's another side. So let's say you have all these assistants, but they have their own mind. Right? In a way, they're not like the traditional tools, where the tool just kinda follows your hand.

Right? Like, you take a pencil, and as long as you have skills, the pencil is not going to interject its own ideas, right, its own aesthetics. It kinda follows, you know, your mind. But with AI systems, and again, using one AI tool, it's like because it can simulate different media and can simulate the styles of different artists. It's like having a studio of tens of thousands of people.

But often or most of the time, they actually don't you don't quite know what they're going to do. Right? And you don't quite know, like, how we're going to interpret your idea. So you type a prompt and you have this, right, image in your mind. And sometimes you get something which is closer to what you imagine.

And most of the time, you get something completely else. You know, when you look at what you get, you can say, oh, that's very interesting. So AI suggested something completely different than what I imagined. And most of the time, you say it's not interesting. And once in a while, you're like, oh my god.

This is amazing. And then you cannot refine your prompt and generate more images, you know, in this particular part of a creative space. So the way I think of, working, right, creatively with AI, sometimes AI works with your collaborator, and other times, it's like a war. Because it it it right? It's ejecting its own ideas.

It's misinterpreting you. And, actually, these are the moments as a artist I find most interesting. Right? So, you know, I've done lots of different image series, and we have lots of ideas. And for whatever reason, maybe I'm just, like, too stubborn as a person.

When AI simply executes almost mechanically what I want, to me, this is boring. So my favorite results happen when, let's say, maybe I mistype a prompt or maybe I didn't explain something correctly or simply in a way I told it something I could not imagine. And, I kind of like these mistakes. I like this, I like this, sort of randomness. And, you know, creative century artists, lots of artists explore, right, randomness, random numbers, and computer graphics.

Like, serialists use different techniques, like, you know, excuse the corpse, etcetera. So AI introduces, like, a high level of randomness. Right? The high levels of unpredictability. You know?

So, basically, it has this ability to really surprise you, you know, and and being stubborn. So I like and and I really enjoy this moment. Right? Basically, what I enjoy is, like, using these tools. It's sometimes enjoyable, and sometimes it's really frustrating.

And sometimes you simply feel that you're at war with some alien creative intelligence. Yeah. And that's exciting in its own in own right, and and interesting in its own right. I think what, what comes to mind for me, from hearing you kinda talk about that and then squaring that up with the, like, the title of the of the book, Artificial Aesthetics, is the fact that when we look at art generated by AI right now, like, it has a certain quality to it. It has this like glassy surrealistic quality that if you know what you're looking for, you can look at it and be like, okay, that was definitely not generated by a human or if it was, it was a human imitating AI art.

Like, it has that kind of AI gloss. And on one hand, like, that's something that's interesting and appealing to me just from the sense that it has its own aesthetic, and I think that's that's kinda cool that there's these different types of art that elicit different types of approaches. And this is very much the approach of the AI generation, of art aesthetic, but there are a lot of people that see that and immediately go, oh, well, slop. It's it's trash. It's trash because it was created by AI, and any of the kind of unpredictability of the image is a tell that this is garbage and we shouldn't pay attention to this.

Like, I I and I guess at the root of this is this idea of, like, what is true art versus not true art? And I think a lot of people look at AI generated art and don't see it as art. I do, but I think a lot of people don't. I'm curious to know what you Yeah. Kinda where you land on that.

Yeah. So so a couple of things here. So one thing is that you're right with lots of images with you online, which people label or, like, you know, maybe Facebook labels it as producer vi, do have this recognizable, kind of very glossy, kind of clean, somewhat generic, and sometimes, like, serial quality. And I would say it's a reflection of partly the aesthetics and the taste of, like, millions of users who use these tools, who, actually haven't, you know, went to art and design schools. And also it's a reflection of, like like, what tools like Midjourney call their house style.

Right? So it's like with default. You know, think about, like, when you're using Jet gpt. Right? You're typing a prompt, it's going to write something back.

It also has this default voice, which, you know, you may find enjoyable or not. But, you know, if you just spend, like, a tiny bit of effort and if you, you know, add, let's say you can add image references to control to control more aesthetics of the results to a prompt, you can add, you know, say in the style of its artist or, you know, add our specifications, like, you know, line drawings, etcetera, it changes. Right? So if you actually look at my images, you know, each series has somewhat different style, and none of them has this generic style. Right?

But I think, you know, because let's say what is according to the web statistics, which I know, 15 or 20,000,000 of kind of users using these tools and 99.90% of them are not professional designers, artists, architects, you know, we just kind of go over with how style and you do get this, yeah, you know, AI aesthetics, which, I think maybe people have the right to criticize because simply these users have not spent enough time. To me, AI aesthetics is actually about something else. Right? What's kind of amazing about, this development of generative AI is that you have a single tool which if you tell it to can simulate to certain extent with styles of tens of thousands of architects, designers, you know, fashion people, street artists, etcetera. I mean, maybe you can simulate, like, perfectly or not.

That's a separate question. Turns out there are limits to it, but on the surface, it can. And to me, this is very interesting. Right? Because in a modern period, we think of a artist or as a offer.

Right? We think of a creative person as having a particular voice, a particular ways of expression. Maybe it is, like, Jeff writing, you know, his his, you know, his, his blog, like, to me writing or, I don't know, Marcel Proust, etcetera, with a particular recognizable voice and the vocabulary of his voice. Right? And let's say, even if I'm professional, you know, well trained artist, I can't you know, or a writer, it would be very difficult for me to start writing tomorrow in a different voice just as it would be very difficult for me to start working, like, with different work or, you know, only but only professional actors can, like, imitate, right, with different face expressions, different voices, and it takes big effort.

So so every human has a particular voice, particular way of talking, right, particular way of expressing, you know, and some people will really cherish how they express themselves. So now we're in this amazing new situation where I can use this tool, which allows me to separate my idea, my concept from the aesthetics, right, from the style of expression because I can basically render, right, the same idea in thousands of different aesthetics. Right? So what does it mean for our concept of art? Right?

What does it mean for my for our concept of, like, artistic identity as art? Right? I mean, how do we think about artists where does it mean that all art becomes conceptual and then this ideal voice is no longer important, because voice is something or your particular signature is something which is simulated by the machine? I I don't think me or anybody else has a answer, but that's one of the very interesting questions. But in a way, AI static simply means that, you know, AI can simulate all types of aesthetics.

And then what does it do to your identity as artist or creator? Still love I'm I'm now that I've left CUNY, I'm associated I'm a visiting professor at Stony Brook University, and I have a course, on the schedule in the fall, on AI and creativity as an undergraduate level course. And I'm fascinated to go off of what you were just talking about in terms of voice. I argued when ChatGPT came out early on that that one of the options for it could have been, could could be to think that people could extend literacy. Those who were frightened of writing can now tell their stories.

And then some of my students in the executive program I started and said, woah, there, you might lose the distinct voices of people to the homogenous voice of all that came before that, that trained AI from those who had the privilege to publish. And similarly with aesthetics and with and with design. And so I'm curious what lessons students should be getting about their own creativity and voice. And, you're an artist, an established artist, and you're using this tool. You're getting into war with it as you said, which I love because you're stubborn and you have a vision of what you wanna create.

And maybe it collaborates. Maybe it it it doesn't. But for a student who doesn't know their voice yet, or doesn't think they know their voice, or they're or they're timid about it, how should AI be used as an educational tool with creativity to enhance their creativity in their voices, not to homogenize them? Well, I think that's a question which, you know, millions of educators around the world, right, are struggling with. Like, for example, many of my colleagues say, you know, we no longer can assign essays because the students can turn in these essays written by AI.

In my case, like, you know, I teach computer science students, so we build the right essays, we do our things. But let's say I started getting emails from colleagues, maybe somebody wants to interview me or somebody wants to ask me a question, you know, and I get these emails from these countries where you don't expect people to write in perfect English. And now suddenly I started getting these emails From me. Which I read in perfect English, you know, and then, like, you wonder how to think about it because you know that the person wrote this using AI. And, hopefully, when the person wrote it using AI, you know, or maybe a person, like, wrote it in her language and AI just translated this, Hopefully, the content of this email, the content of this interview questions is exactly what this person wanted to write or maybe not.

Right? Mhmm. But, ultimately, I think it's a good thing because it does allow more people to express themselves. I think that, like, all I mean, I believe that, like, all technology, all digital media is particularly good for people maybe who haven't who don't have privilege of going to top schools, or, you know, people, like, in like, various developing countries because it levels a little bit more playing field. So I do think with tools as you're having extremely positive effect, in the same time, right, I think the way, you know, everything in society.

Right? You know, you often you you can't just get something and only good effects. Right? So many things, you know, come to us and with both positive and negative effects. So the one thing I see is incredibly democratizing effects of AI technology, but in the same time, right, where's the danger where you start speaking or expressing yourself, right, and making music or making videos like everybody else, but there is a kind of particular kind of loss of individuality.

Right? So it's like we enable you, but as a part of this enabling, maybe, you know, you can you can ask new questions, you can you can email to Jeff and Lev Manovich, but suddenly the questions you're asking are maybe less original than you would want before. And you would have to decide, right, as society, as people, how do we deal with this kind of, you know, this tension, right, between these two different effects. But one more thing I want to say, it may change years from now, but I think my experience, right, you know, I use AI, you know, to make images, I use AI to edit my books, you know, etcetera, etcetera. Right?

Basically, in lots of studies which have been published in top scientific journals over the last years, like, Science and Nature, you know, they would give some tasks to doctors, even compare doctors' performance like making diagnosis to AI, AI, or we compare AI performance to CEOs, or we will compare the AI performance to, like, writers. You know? And all the studies results are more or less similar, where we say, in this experiment, we are is better than 95 to 97% of professionals. So this is with with horrifying news. Right?

With the eyes probably better than your doctor. Right. So when the conclusion which people draw from this is, like, an optimistic, they say, okay. But the best humans, the best doctors, the best CEOs, the best artists are still better. Okay.

But here's my experience. So let's say I'm editing my text. Right? So I can ask AI, okay. Here's a sentence.

How do you improve a sentence? Yeah. I said, okay. Here's 10 different ways to write the sentence, and here's what differences are. I never ever in my whole life and I published hundreds of articles or I can lots of books like you.

Yeah. There's probably this amazing editor sitting somewhere I don't know, 100 year old editor sitting somewhere in some publishing house or sitting in New York Times who is better than AI. But I've never met this person in my life. Right? So we kind of so we sort of kind of we're we're horrifying conclusion is realize, yes, there is this 3 or 5% of of human experts who are better, but the typical doctor, the typical editor, you know, not to talk not talking about and also talking about the students.

Right? No student in my class, no student at I mean, last week, I had a conversation with students at Harvard. They had nothing to ask me, which is silent, which is sitting where eating great pizza. Whereas the AI will basically generate a list of 10 questions. So we have to kind of, like, accept this horrifying conclusion that we create a technology which is better than most of us, you know, and then, you know, what does it do to our dignity.

Right? What does it do to our kind of Yep. Sense of self worth? But but but in the same time, what I want to say is at the moment, like, AI, like, you asked something to do something, it's good, good, good, even bad. So what realized, like, if I ask AI to make some theories like me, you know, look at the results, I have to know as much as Lev Manovich to evaluate the results.

Right? I have to have this 50 years of art experience to, evaluate AI image output and to say, good, good, bad. In other words, I have to I still have to be the top expert or narrow expert in a particular field to evaluate AI outputs. Right? Which means that the students still have to learn.

In fact, they have to learn harder. And the best students, the best minds would use AI, I think, in a way, you know, adventurous way, I mean, advantageous way. And the rest of people would use it in a very complacent way. So in a way, like, AI would kinda make the best people better, and the rest of people, it may make them, like, a bit even more lazy. But maybe in the future, people will invent, like, a new level of AI where there's gonna be no AI, which will evaluate this AI and say, actually, this is good result or not.

But for now, it makes, right, sufficient number of mistakes, and these mistakes matter. So when we say AI is wonderful, I think we maybe don't we forget with, yes, but AI wrote this text for me, but I corrected the sentence. And if I didn't correct this one sentence, the result actually would be terrible. So, so in a way, education matters. And I think what we can also say, what people always say, the society is going to change, the technology is going to change, and the people have to learn how to learn.

So, like, in a way, we're we're more flexible you are, the wider your knowledge base, the more skills you have, as long as you will be able to evolve, right, in future, you'll do okay. Right? If you just can learn one little thing and get stuck, then I think, like, AI will replace you. Yeah. It seems like what you're what what I'm hearing you stress the importance of is the difference between people who use AI wholesale to do the thing.

I'm just gonna use this thing to create that thing, and, yay, my work is done here. Because in a in a you create And I'm done. I'm done. Right? Versus the collaborative approach, which is, in my experience, been the real powerful kind of, insight into how I use these tools is when I stop looking at it as a replacement for my effort and start looking at it literally as, like, even though it's a machine, you know, a a a human like machine sitting next to me that I can bounce ideas off of and get to places that I couldn't have gotten before, or it would have taken me a lot longer or a lot more effort to get there before it kind of accelerates through that collaborative experience.

Or what you said, Lev, it's not just a collaborate. Well, I'm thinking about collaboration. What I love about my takeaway from our conversation is it's more the confrontation. That that's where when it challenges you to be better or it challenges you with something that it fights you on, that's where the interesting stuff is. Not that it helps you do what you already wanted to do, but gets you to do things you may not have known you wanted to do.

Is that fair? Absolutely. And maybe if I have to have, like, 30 more seconds, I would then one more thing. Oh, yeah. I think I think with big limitations of generative AI right now, like, the way it is making images, videos, music, or, like, editing, writing text, right, it basically learns patterns in the training data.

In the training data, which is the web, right, it's not, you know, there are certain things which are represented, right, more frequently, certain images, certain ideas, right, certain theories, and there are other things which are more rare. Which means when we're using generative AI, it tends to give you answers, write text, and make images, which often tend to be on the generic side. Right? In other words, it is less likely, and also that's just how technology works. Right?

It basically tries to predict, you know, what is probable. Right? When you train the CI systems, you know, you take a sentence, you, mask part of a sentence, and when you train the machine to predict the rest of a sentence, so it will predict the most common sense sentence. So, this battle, right, against the AI systems which I'm describing, it's constantly looking at the output and saying, okay, this is good, but here is something generic. This is not my idea.

This is not original enough, which is not specific enough, so I have to rewrite it. And I think it takes lots of education, lots of experience to be able to recognize that this is a generic result, this is stereotypical result, and this is a unique result. Yeah. And I don't think this problem will be solved tomorrow because this is the fundamental to how statistics, data science, and AI works, which means that, it's in a way, it's a struggle against, Vekuman with 3 atypical Vekanaal. You know, and that's, you know, it's an interesting struggle, and I think we have to somehow educate our students, right, to recognize this and to strive against simply replicating, simply copping, you know, the most common answer.

To care enough to have something they want to say, that that we've helped that's what that's what education is. Right? We help them bring out what they want to do, and what they need to say. But I think also encouraging one, which is I think a real challenge in our environment. Right?

Because some people say that the cultural developments of last 10 years, right, created this atmosphere where maybe people are afraid to have real dialogue, people are afraid to have, like, real struggle of ideas because nobody wants to offend everybody. And that's wonderful. Our society could become very sensitive. But as a result, like, you know, maybe people, afraid to express sometimes in public ideas which are not popular. Well, this is this is actually a terrible thing.

Right? Because that's what AI, the way it's designed to do. Right? AI is designed in a way to, express, you know, to give you most common ideas. So we have to fight for independent thinking.

We have to fight for diversity of ideas. Yeah. I had a conversation yesterday. I know I'm dragging you on a little longer than than we planned, but this is fascinating. I had a conversation yesterday with a journalist in Denmark.

And and we were talking about how people they do this with search engines. People are are franker with a search engine. They'll ask a question of a search engine. They won't ask them a friend. Right?

They may end up in the dialogue with AI that they wouldn't end up in public. And so it's interesting too. And and when I wrote the book The Gutenberg Parenthesis, I was fascinated by the impact of silent reading. Silent reading led to things that you wouldn't do in public like pornography. Right.

Silent reading is all in writing. So I almost think But but but but if information because people before, like, end of 8th century, people would, like, read loudly, but you're referring to it. Right? Exactly. End of 8th century, people started to read silently, which to us, we're taking for granted, but was actually cultural revolution at the time.

Right. Right. And it led to it led to heresy because you could have a forbidden thought, that you could read or write. Right? So as you're talking, I almost envision, a potential of private art with AI.

That as you just talked about about the the sensitivity to not say things in public because you can't, well you can enter into a dialogue with AI in text or in images, or in sound, that is private, which is fascinating. So sorry. I didn't that was just a little No. No. It's it's Rat holes.

Perfect. That's perfect. Yeah. Yeah. I mean and, you know, and I think there's something encouraging here because I'm fascinated by the genetic ability to be, like, willing to click.

So when you ask a question about something, right, you say, okay. Can you give me some option, some ideas? It will say, here's 10 ideas. Yes. Here's 10 options.

Right? So in a way, like, we humans, we get obsessed about particular idea, and we can go down this rabbit hole, and then the best of us develop this really original, you know, visions, etcetera. But our ability in a way, our lack of ability to be very, like, objective and to be very, it also what hinders us. So I think if we can somehow learn, right, to combine our obsession with kind of going with being, like, very deep and narrow with the AI, like, Wikipedia, like, and ability to look at all the options. And then in a way, we can have a new kind of symbiosis, you know, and I think a better humanity.

So kind of using it, like right? Not trying to not trying to kinda make AI exactly like us, not trying to replicate the negative parts of evolution, but using AI as a kind of, yeah, as to supplement what we what we can do as well, for example Yep. In thinking. Yeah. 100%.

So cool. Thank you so much. Thank you so much, love. Yeah. I really appreciate you, taking the time to to join us to talk about this.

I realize I realize we we hardly talk about the aesthetics. We talk about reveals, but I think that's what, I think that's what in the way, I think it reflects you know, it represents the book pretty well. Right? Yeah. It does.

It does. Indeed. Yeah. Like, what about this moment? And, you know, how does AI change us also knowledge, thinking, creativity, education, so art is just maybe one place where we see the things more clearly, but it basically applies across the board.

And our relationship to art. Yes. Sure. And the fact that all new technologies require new skills, and we're kind of in the point right now of learning those new skills to work with it in the best possible way. And that's a pretty like, I I have an optimistic outlook on that.

I think, ultimately, we're gonna get to a good place, and it sounds like you do too. Lev Manovich, of course, is a professor of computer science at the City University of New York, coauthor of Artificial Aesthetics, and, just really appreciate your time, Lev. Thank you so much for hopping on with us today. Thanks for having me, and thanks for having me this wonderful podcast. Oh, thank you, sir.

Appreciate that. And we'll, yeah, we'll reach back out. Head back again sometime down the line. Thank you again, Lev. Oh, my gosh.

We'll talk to you soon. Yay. We made it happen. Yes. So happy.

Love it. Good stuff. Anyone who's who's watching or listening to the podcast, if you have thoughts about what we talked about with Lev, contact@aiinside.show. There we go.

I've got too many shows. contact@aiinside.show. Let us know what you think, and we could read it in a future episode. We appreciate your feedback. Let's take a quick break, and then when we come back, I'll give you 3 guesses what we're gonna talk about because if you've been following AI news, you already know it's DeepSeek.

That's coming up here in a second. Alright. DeepSeek. You know what, Jeff? We talked about DeepSeek before.

It was really, really cool. Yeah. We were cool. Benedict I I give credit to Benedict Evans who I quote often who at the end of the year said, oh, deep seek snuck in with the big news of the year. And I thought, what the heck was deep seek?

And then we looked it up and we played with it and talked about it. And, and so, yeah, we're way ahead. We should have bought some sold some we should have shorted NVIDIA, out of that. Yeah. Right.

If we had if we had if we had only known and if I've ever like, I've never had a good financial sense of anything along those lines, so that would never be the thing that comes top of mind for me. But boy has DeepSeek shaken so much of the conversation around artificial intelligence, around these, you know, companies that for the last couple of years, we've been watching invest massive amounts of capital into their models because that's how you do it. And, you know, and throw that in air quotes and NVIDIA creating these, you know, massively capable GPUs that are incredibly expensive and super powerful in order to drive all of this. And then a Chinese company called DeepSeek comes along and does it for a fraction of the cost in a short amount of time, and they're able to create a product that, is really going toe to toe, with all of those systems. And so, you know, everything that I've read from people who have been analyzing this, they're just basically saying, woah.

The goal posts have changed. Things have shifted because of what DeepSeek has proven here. And just real quick before we get to the actual news, I think you'll appreciate this, Jeff. I had a DeepSeek dream a couple of nights ago. That's how You're doing too much of this, Jason.

Apparently, I am. But in my dream, I was having a conversation with my wife, and I was, like, telling her about the fact that you and I were gonna talk about deep seek on the podcast and kind of explaining it all to her. And in the dream, it's I sounded like I really knew what I was talking about. So so so this news is fascinating because I I'm not really sure why it hit the way it did in the last week. DC came out in late December.

It's different in a couple of ways. We played with it on the show. It it it it's supposedly it has two things, a search and a and a and a and a rational sense, a thinking sense. Right. And if you ask it a question, it will show you supposedly show you its thoughts.

Now I don't know if that's we talked about this in the show. I don't know if that's real or if that's made up to make you think that it's thinking. But anyway, it's a it's interesting. It's a bit different, but you're right, Jason. People are impressed.

But what was clear is that they did it, and we talked about this in the show last week, they did it with fewer expensive chips because they can't get them in China Yeah. Because we're cutting them off. So, you know, maybe Within their limitations. Right. And often, so often, really great things come, you know, the, like, shifts in thinking come from working within limitations.

And this Iambic pentameter, mate. Right? Haiku. Right? The limitations create creativity for our last discussion too.

Mhmm. And so, obviously, what happened was there there was a there was a moment where people woke up and said, oh, crap. We're spending, $500,000,000,000 on Stargate. We're we're buying NVIDIA chips up the yin yang. NVIDIA's doomed suddenly, which was stupid.

It wasn't doomed. Maybe the markets a bit smaller. Sell off of, of stocks 17%. Went down 17%. And then yesterday, the next day, I saw stories on this today in the Wall Street Journal.

Retail investors, individuals pumped it back up because they bought on the dip that they said, this is ridiculous. 70% for something that was so rising so high. So it rose again. I think yesterday it was about 8.5%. Last I looked, it was down 4% from that line today.

So it's gonna be a roller coaster, as technology in a way should be. And there's discussion about whether deep seek cheated a little bit, whether it it used, OpenAI to train its thing. Even so, it created an amazing model without, a a data center the size of Manhattan, without, a gazillion NVIDIA chips. This says something. The other thing about it, and and I don't know if this is in the rundown or not.

I forget. But we the other thing I talk about in the show a lot, and I think this is a bit of a victory for open source too. Oh, yeah. Because they used meta meta's, llama. They, built on something because it was open.

And again, it's not fully open source. It's open, stipulated around it. Open ish. But I think that, that it's important for all of these reasons. Do we need to have this huge investment?

Is it possible for other developers to do things at a smaller and doable scale? Is the expense thus a lot less? Is the environmental impact potentially less? Is the competition greater? That's what I think deep seek represents.

Whether it itself is the future, whether it's anything. No. But it opens up the discussion, which I think is really healthy in this gestational industry that we cover. Yeah. And especially at this moment, you know, the the next story that we were gonna talk about, has to do with, you know, kind of the government.

Actually, actually, no. I'm realizing I took that out because I think we talked about it last week or it happened before last week's show, but, president Trump's, rescinding the Biden era Yep. Executive order around AI safety. And, essentially, right now is this moment where there is this massive groundswell of we've gotta be first, we've gotta be bigger, badder, you know, better in all ways, shape, or form when it comes to AI. And yet here we have kind of at this particular moment where that temperature is so high and so important seemingly.

You got this Chinese kind of upstart company that really seems to have come out of nowhere. It did it it it created a a model or a series of models that go toe to toe with the ones here in the US that have been, you know, created with 1,000,000,000 upon 1,000,000,000 of dollars and a whole lot of time and all of this, you know, as they like to say, their moat around around their services built an incredible moat to kind of prove that their value is so much better than anything else out there. And then yet here you have this example of wait a minute. This kinda came from nowhere with incredible limitations and it's and it's got them scrambling. Some articles I've read, you know, show that the the the war room in meta where they're like, how the heck did they do that and how can we replicate it?

And to a certain degree, seems like it might kind of change the road map or the the blueprint for these companies as we, you know, are only at the beginning of 2025 in a way that even they might not have predicted. Yeah. I think you're really right. And also, I think at a time of economic and cultural isolationism, this punctures that wall. I saw a story today and this is off our AI topic, but when, you know, TikTok going off, a lot of consumers in the US went to other Chinese runs, systems.

They didn't go to Meta. They didn't go to Instagram, and and Facebook and so on. They went to another Chinese service. The red something. I can't remember the name of the service.

Whatever it is. It's kind of the way to say, like Mao's little red book. So, yeah, it's Chinese. And so you can't stop it. And when it comes to AI, by trying to cut off the Chinese, we only created, more robust competition.

Yeah. And so global you can't just turn off globalism. If globalism is a reality and in technology especially, and in, the Internet, a connected world, and in science, we are irretrievably global. And I think that's another message for the DeepSeek story. Now Now whether DeepSeek it'll be interesting.

It'll take a long bet. In 6 months, will DeepSeek still be the hot thing or not? Yeah. For what you said, it kinda doesn't matter. It it just it made an impact.

Yeah. It it proved something that wasn't easily identifiable prior. And whether that what that means for DeepSeek's business specifically, that is an entirely different question. I'm sure they're gonna be part of the conversation to some degree. You know, being there being the first ones to kinda prove this smaller, you know, more robust, less expensive model going toe to toe, blah blah blah.

There's a whole a lot of story there. So I think that that keeps deep seek in the conversation, but, really, you know, the bigger picture here is, oh, there's a different way to do this than the way that we just all assumed it needed to be done Yes. Which required an insane amount of capital and time and investment. And, you know, and, you know, like you said, kind of the unsung hero part of this story is the fact that it's open source. Like, immediately, people can can do what they want with these with these models, and they aren't they aren't hamstrung by the $200 a month pro membership for ChatGPT to do some random thing, which is you know, I have I have to imagine those kinds of pricing structures begin to to, tame out a little bit or dissolve to a certain degree because the value, you know, is the value the same today as it was a week ago or 2 weeks ago?

I don't know that it is. Yeah. And, you know, if you look at at the cost, so g p d four cost 40,000,000, what's that dollars? Yeah. $40,000,000 to train and, DeepSeek cost, I think, something like 6,000,000.

66,000,000. Yeah. And, and for a university, and I'm involved in a very, stem based university now in Stony Brook. And, and so this I think affects the ambition you can have at that lower level. Whether you're a company or university or a lab or a country, right?

There were countries who probably said that we can't, we can't play in this. France was there with Mistral. But as usual, many European countries will say, Oh, no, the Americans again, geez, we're left out. They won't pick us for the team. Well, no.

China proved that, anybody can be in this game. Yeah. So great time to have a podcast, Yeah, I know. Totally. It's a, it's an exciting time.

Fascinating. It's cool to kind of be, yeah, to be in a position. Where you thought, oh, it's gonna be 3 or 4 big companies. They're gonna learn it all as well. Here we go again.

Boom. Yeah. I know. It all changes. Just when we thought we knew everything, then suddenly I have a dream about deep sea.

It's real quick shout out to Dimitri Shapiro from MindStudio who is watching us on Facebook. Oh. Live streaming right now. Is it Facebook? Yeah.

It is Facebook. Hey, Dimitri. I accidentally didn't put it on Twitter. I put it on Facebook and LinkedIn. So hi, Dimitri.

Yeah. I know. We've got different networks. I like it. We're we're spreading it around.

So, another thing that I was alluding to just a little bit ago is, this kind of this movement on OpenAI's front. The next two stories are actually OpenAI features or Yeah. We remember them. We're old enough to remember OpenAI. Yeah.

You remember OpenAI? They used to be a big deal before deep sea came along. OpenAI launched chatgpt gov, which is a new product designed specifically for US government use. Now that, you know, now that Altman's getting real cozy with the current administration, as they all are right now, it's And the Pentagon has said not to use deep seek, by the way, because it's Chinese. I mean, had to see that coming.

Yeah. Of course. This product provides enhanced security, beyond ChatGPT Enterprise, so it's a self hosted solution. It can process for the government, sensitive non public information within that secure environment, which absolutely you need if you're gonna use something like this. It operates, you know, within the government, the walls.

Operates through Microsoft Azure Commercial Cloud, Azure Government Community Cloud. It's compliant with, you know, many of the most important security frameworks, IL 5, CJIS, ITAR, FedRAMP high. And it's testing starting next month. And yeah. So this, you know, this could be integrated with defense agencies, law enforcement, health care organizations.

This is OpenAI. Get real cozy with the US government, to Or trying to. Specifically for them. We're trying to. Yeah.

Yeah. Yeah. For sure. And then another product that they were, that they announced, and this is actually, you know, a total, you know, different direction than what we're just talking about, but this is all about agents. And it's a project called Operator, and it's a research preview right now.

This is one of those features, one of those products that you can only get access to if you're on the $200 per month plan, through OpenAI. So, I think, eventually, it'll come to the $20 plan. But, essentially, what it is is it's a computing using agent. That's the model that's driving operator. That's what OpenAI called it.

It simulates keyboard mouse clicks in the in a browser, a browser instance. So it's not like it opens up your own browser and starts browsing. It's done through their system and streamed to your computer. And, it's so it's a remote browser essentially in the cloud, and it acts out those actions based on a command that you enter. So if you wanted to, you know, book a 6:30, this is always the example.

Right? Book me a 6:30 reservation at, you know, Boulevard restaurant for 2 people tonight, you know, and it'll you can actually watch as the you can choose to watch. You don't have to, but you can choose as it opens up a little that kind of virtualized browser instance and navigates the the site and does all the things. And if it hits a point that it knows or it has been programmed or told to ask for your confirmation or intervention, then it won't go any further. So, like, logging into something or confirming a reservation or spending money, there are things at which it requires a confirmation from you.

But this is this is ChatGPT's agent, essentially. And I think what's what's interesting about this, I can't because I can't play with it because I'm not spending $200 a month. Sorry, Alvin. But I think what's interesting about this is that it does become the point where you can see what's going on. And the fact that Dimitri is here, with his with his application, that's part of the advantage of what he's built.

And the problem with, with AI and agent tree is it's opaque. Well, it it went off and did something. It came back and it gave me an answer. And deep seek tries to simulate a process of giving you an answer. But again, I don't know how credible that is.

In this case, you watch it actually do it. So I think that that does have some value, at least as kind of an educational level in the early days of agentic AI to show you, oh, yeah. That's how it that's how it did it. Okay. Alright.

I trust it then. It's doing the right thing. It does have some partnerships. So like Instacart, DoorDash, Etsy, OpenTable, TripAdvisor, you know, a number of other sites that are partners. I do wonder what that means.

Does that mean that they've trained the system specifically on those sites to work well with them, you know, which I feel like we've seen before, not in the AI realm? Does that mean that they get a rev share if someone uses Instacart through the browser instance to, you know, buy It's also an interesting problem. You know, I talked to the show many weeks ago that we, we had SEO search engine optimization, and then we have generative, GEO, generative engine optimization. Then you need to go to a low, agentic, engine optimization, where if if if the agent understands how to, navigate your site and you redesign the site Yeah. Right?

Yeah. Oh. If if your button suddenly turn changes from green to red, you know, simple simple color change, is that enough to totally throw it off based on what it had learned and yeah. Little things like that probably It'll be interesting to watch. Big impact.

It'll be really interesting to watch. And I'm sure, you know, I'm sure this is the kind of stuff that gets figured out eventually, but, you know and and also, again, how do we feel about assigning a task, you know, to to an agent like this? Like, there there's probably a certain level of task that I'd feel comfortable with, you know, because if it gets it wrong, it's not the end of the world. And there's a certain level of of, tasks that I probably would have a really hard time convincing myself to to rely on it until it can prove to me that I can trust it. You know?

But, and, actually, another point before we go to a break, how does an agent, like this that can do human like things on the web, how does that impact the financial underpinnings of how the web operates? So much of Right. So much of how sites determine, you know, who gets what money and and attention and all this is based on our human attention. When we visit a site, they know we're a captive audience. They put up an ad, and that matters.

That means something that someone's paying for that. How is that influenced or or changed or impacted by the fact that there's a bunch of robots doing that instead? Right. 1 of 1 of the big, the the the golden fleece in the advertising industry is to prove that there was a relationship between seeing the ad and taking the action. Yeah.

Who gets credit for that? Right? But what potentially changes here you're right. It's fascinating, Jason. I hadn't thought about this.

Is that if I have my OpenAI agent go off and buy those shoes, right? OpenAI is gonna come along to the shoe place and say any upper. I'm not I'm not, buying for you anymore. Mhmm. Which then affects the credibility of OpenAI's agent with me because now you got me crappy shoes.

I like those shoes better. You didn't buy them. Well, because you're not getting the the the the deal. So what's the transparency about those deals and how it operates? And does a media property get the money or does the agent as the, middle man, so to speak, get the money or no one gets the money and it stays in the pockets of this is a kind of a Craig Newmark way to look at things.

The customer, and the company at the end rather than all these middlemen. Yeah, it, it, it does throw up a lot of questions about these financial relationships. Indeed. Indeed. It does.

Curious to see how that all pans out. Gonna take quick break. One more quick break, then we got a few more stories to round things out, coming up here in a second. Alright. You flagged this, series of of, stories, and I'm happy that you did.

Reid Hoffman, as well as renowned cancer researcher, Siddhartha Mukherjee. I don't I'm I'm probably mispronouncing the name, but launched Manas AI, AI drug discovery, startup. And this is concentrating initially on breast cancer, on prostate cancer, on lymphoma, and it's using Microsoft's data centers for, their operations. Of course, Hoffman is a Microsoft board member. And they plan to generate their revenue through the development and sale of the company's own drugs to, tackle some of these issues.

And, you know, you you put this in here. You wanted to be sure that we talk about it. So I'm really curious to hear kinda what your thoughts are. So there's a couple pieces of Reid Hoffman news right now. Mhmm.

And what I'm interested in, Reid Reid is unquestionably a technological optimist. He believes in the power of technology. He believes that it can do great things for us. He's not as obnoxious as Andreessen. I wouldn't call him an accelerationist, but he does believe we can do great things with the technology.

Let's control it. So one of the things he did was this is a company to to, Mukherjee wrote a very famous book about cancer and to tie with him to reimagine ways to attack cancers. And they're gonna do that as great. At the same time, Reid has a new book out called Super Agency. And he appeared he has a New York Times op ed and he just appeared on Morning Joe.

And Morning Joe drives me nuts because it's the, as I often tweet, oh, what time is it? It's moral pinnacle clock on Morning Joe. Because they're gonna go after everything wrong with technology and it's phones and it's social media and now it's AI. And here was the problem. Yourself but watch, Jared.

Yeah. Why? Or as I tell people on social, I watch so you don't have to. Right. So here comes Reed saying, no, we can do good things.

Yes. There are problems. We'll figure out those problems so we can do good things. And it was wonderful to be through Beka Brzezinski for a loop. She didn't know what to do with this.

But you mean we have agency? And he said, yes. We do. So that's what I wanted to concentrate on today. A little bit just for a minute because, whatever you think of Reid's optimism and whether you think it's too optimistic or not, put that aside.

The the moral for a moment, the moral of the story I think is that we have more agency with technology than we know. And that that sounds ridiculous, and I wrote about this in the web we weave. It sounds ridiculous because well the big companies control it. We can't we can't do anything. It's the algorithm.

It's it's it's it's the wizards of technology. But we do have the opportunity to make the technology work for us. And so I'm I'm writing the next book right now on the history of the Linotype. I won't bore you with the details right now. I will later.

But it's the machine that stopped typing set one letter at a time, through 500 years in Gutenberg, and then set it a line at a time. And it opened the door to and ended the process of the and industrialization of print. Open the door to mass media. Okay. The story that's related here.

I'm sorry. Okay. Uh-oh. I didn't turn off the ringer. I have an old phone.

I have an old fashioned phone. We never answer it because it's only gonna be spam. It's over there. So give me about 4 rings and it'll go out. I apologize.

We we got rid of that phone a long time ago. Yeah. No. I'm still I'm old. I'm old.

I got a phone like that. We're not gonna get rid of that. But we never answer it, as it can prove right now, there's other people in the house, so we just don't answer it. Anyway, I would too. Yeah.

So so the story that interests me about the line of type is that the the typesetters, the people who set type one letter at a time for 450 years, along comes this machine and they are threatened to death by it. They're also threatened that new people could operate that machine, women and young people and and others. And and they could have done what, the Luddites did. They could have tried to destroy the machine or or others who did in movements. But instead the International Typographical Union said, no, it's inevitable and we're going to control the machine.

We are the best ones to set type on it. We are gonna control it. We are gonna do it. And they knew that the jobs that they had would go down for a bit, but they also knew that publishing would explode because the cost of publishing was down. We could have more newspapers, thicker newspapers in more places, more magazines, more books.

There would be more printers as a result. And so by taking agency of the technology, they won for a long time until the next time when they fought it. And I won't go into that today. So the only reason I wanted to mention Reid Hoffman's book, which I haven't read yet, I bought his op ed in the times, his appearance on morning joe. And he proves it by investing in companies like a cancer company is that his message is take agency over the technology.

Make it do good things for us. Make it ours. And I think that's a moral to the story about how we deal with technology. And what interests me about AI is that even though you and I Jason can't make a model right now, probably never can, we probably can't train a model, but more and more and more isn't the lesson of AI that people without technology backgrounds can take can run the technology, can be the boss of it, can tell it what to do. Isn't that what Lev Manovich proves?

And and that he used it. I I I love the the discussion there about how he saw the combative nature of it, as as as as a fight and a struggle, but he was gonna win. He was inevitably gonna win because he's the artist. And so that's just an interesting context here, I think, to all of our discussions about AI. And so I wanted to take this little rabbit hole tour, today.

And that's why, normally, Jason does a great job. We put in tons of links in the rundown, and Jason really does he has a great news judgment, and he picks out the stories. And I I never disagree with him. This one case, before he decided, I figured Jason's gonna look that and say, why did Jeff put that in there? That's a little weird.

I love that you called it out. Yeah. So I called it out and I said, I wanna I wanna I wanna rant about this for a minute, because it's not obvious what the rant is. But the rant for this show, for what we do, I think, is that's so exciting about this area is that we want to empower all of us to be able to use the technology in a way that's beneficial to us. Also being aware of the bad parts, also being aware of the risks.

Absolutely. But if we don't try to take it over, we're fools. That was a thank you. Kind of, an overarching theme of today's episode is particular specifically with with Lev, with the deep seek news Yep. With Reid Hoffman is this idea that, like, oh, wow.

This is a technology that is, you know, incredibly democratizing. It comes to a human level. Yeah. Okay. Yeah.

It it becomes accessible, for all, and there there are opportunities for everyone to get involved in it. And there's things that that that we hope that I I can't use it to to to cure cancer, but I'm glad the people who who perhaps can are working on it. Absolutely. Yeah. And and, and also for our next story, I'll never be able to do what the next story does either.

Yeah. This is this is one of those stories that, like, you know, I I put in here fully realizing that I don't entirely understand it at a certain level, but I love that that it's happening and I understand it enough to know that it's, you know, things like this are significant. This is an AI model, that's created a novel fluorescent protein, which apparently is a major leap in protein engineering. It's called e s m g f p. It was developed using e s m 3, which is an AI system, that simulated molecular evolution that would and this is, I think, the kind of key takeaway.

It would have required 96 genetic mutations to happen naturally, but here it was trained on 2,780,000,000 natural proteins. It learned to complete partial protein blueprints based on its understanding of protein biology. And, you know, at this point, it's received peer, review validation, publication, the journal Science, and it's just really, at the end of the day, a really big example, a demonstration of how AI can accelerate these things and our understandings or our development of certain things like protein engineering far beyond, the the constraints of natural evolution. You know what I mean? If you know, how how many, you know, years or decades or millennia or whatever is it would it take to do this, And yet here we have these systems that, you know, we have the ability to train them to do so much in so little time.

And it's just kinda cool to think about what kind of doors that opens, as we go forward with it. The things we we can't imagine. Yeah. And and I think that's that's good. And, yes, there could be perils and there could be things that we we had to be be cautious about.

That's all true. But, how can we not, try to see where this might go? Yeah. Yeah. Very, very cool stuff.

So cool. I think the overarching kind of thing from this show is a level of optimism. Yeah. The level of humanity. Right?

The level of bringing bringing it back to us. Right. It's not just gigantic companies that are controlled us and do stupid things. Yeah. The tools might be in our hands.

Yeah. Not just those things. Those things do exist, but it's also the other things too. Yeah. Huge.

Thank you once again to Lev Manovich for spending his time, for a little bit. That was a wonderful conversation. I'm really happy that we were able to finally get him on. Also thank you to Dimitri, for sharing and I also should mention, by the way, that love, is he's he, is in New York right now because he has a part in an exhibit, called symbiosis art in the age of AI at the Sylvia Wold and Poe Kim Gallery in New York on Lafayette Street, and it's going on until March 29. So if you're in New York, and you're inspired by all means, by Lev's book, which we don't have to buy it.

You can just read it online because he was generous enough to put it up there as a PDF, But, also, you might wanna see his, the show that he's this exhibition exhibition he's part of. Love it. I wish I could see that. Alright. Well, thank you for that.

Thank you, Jeff, for everything that you do. Jeff jarvis.com is where you can go to, find all the books that Jeff, has been parenthesis now in paperback. Now in paperback. There it is. And then, of course, magazine and the most recent book, The Web We Weave.

Lots of themes as you read, you know, especially The Web We Weave as I've read through it. A lot of themes from the show, you know, distributed throughout the book and everything. So some really familiar kind of conversation or, chapters in there. So it's it's worth reading as well there. Thank you, Jeff.

Thank you. Appreciate your time. Appreciate, doing this each and every week. Everything you need to know about the show can be found at our site. It is aiinside.show.

You can also, find a link to our new YouTube channel, youtube.com/@aiinsideshow, one word. That's where we were live streaming today. That's where all of our video episodes past, present, and future will be going forward. I'm so happy we have a a dedicated channel for that. If you do love this show, we would really appreciate a a refresh review in Apple Podcasts.

That would, you know, we actually put them up on the website as new ones come through. We haven't had some in a while, and, you know, we've been going for a while now. So if you want to go in there and refresh your review or leave a review, it really helps get attention onto the show, and we are definitely looking at growing it. So, leave that review wherever you can. Probably Apple Podcasts is the best place to do that.

And finally, patreon.com/aiinsideshow is the place that you can go to support us on a deeper level. You get ad free shows. You get a Discord community. Sometimes we do some hangouts. We have an executive producer level, which is the top level to contribute at.

But if you do that, you get your name called out at the end of the show, you get an AI inside t shirt, you also get a sticker, and just knowing that you're helping this show, each and every week continue to thrive. DrDew, Jeffrey Maraccini, WPVM 103.7 in Asheville, North Carolina, and Dante Saint James. See, at this point, we've been reading some of these names for months now. This shows this is a testament to how cool this level is because now you know them by name.

You're like, oh, yeah. There's Doctor. Do. Oh, yeah. There's Jeffrey.

Exactly. So that's kind of the benefit. Our our dear friends. Yes, indeed. So thank you everybody so much for watching and for listening.

We always enjoy doing this show and hanging out with you. We will see you next time on another episode of AI Inside. Take care everybody. Bye bye.