- 5 months ago
Centuries of art history assumed a human author. That assumption is now in question. As AI systems produce novels, visual art, and film at industrial scale, the definition of creativity — and the economic architecture that rewards it — is fracturing in real time. In this talk, Lev Manovich traces how generative AI is reshaping the way culture is made and shared. Looking at both art history and today’s digital platforms, he examines a simple but urgent question: when machines can generate endlessly, what still makes human creativity stand out?
Category
🤖
TechTranscript
00:07I
01:10Yeah.
01:12Start?
01:13Okay.
01:14Great.
01:15Good afternoon.
01:16We're going to start.
01:18So we plan for the next half an hour is I will do a short talk, try to keep it
01:22to 12, 13 minutes,
01:24and then we'll have a conversation about AI, creativity, future, and do we actually need to be more creative or
01:31maybe less.
01:35So I've been a practicing artist for over 50 years, but also a theorist of history and media,
01:43the offer of 17 books.
01:45And what I'll share today comes from a new book I'm writing, which is actually already a third book on
01:50AI, art, and visual culture.
01:56This is an example of one of my image series I created.
02:02I created dozens of them in the last few years.
02:05So this was done using the journey and then heavily edited in Lightroom.
02:09So when generative AI arrived around four years ago, my instinct was not to ask maybe it's good or bad.
02:18I'm not really interested in these moral questions because I think every technology is always good and bad.
02:24It depends on how it's used.
02:26But I wanted to ask a different question.
02:28What kind of medium is it actually?
02:31What are its properties?
02:33What are its characteristics?
02:34Right?
02:35So we have textbooks which tell us about the kind of properties or dimensions of photography or painting or cinema.
02:42But we don't have such a book for AI.
02:45So my writing is with them to develop something like this, which of course is challenging because some elements of
02:51AI change very quickly.
02:52So the trick is to get at a more essential, right, deep level properties, which will still be important, you
03:00know, hopefully five or 10 years from now.
03:03And today I will talk about previous properties.
03:07And since we're in Paris, I will try to make a few references to French art.
03:18Okay.
03:19So this is the plan.
03:21Three properties.
03:22AI, creation, post-creation.
03:25AI and artistic identity, content versus style.
03:28And AI is anti-avant-garde.
03:33Like many people, starting in the middle of 2022, when contemporary Gen AI tools started to arrive, such as Midjourney
03:41and later CGPT, I started to make images.
03:44I made lots of them.
03:45So I was using Midjourney almost every day for about two years.
03:50And I made hundreds of thousands of images.
03:52Even the ones which I thought had potential, I saved to my drive.
03:56So I saved about 20,000.
03:58Even I went through them.
04:00And the ones which I thought had more potential, I edited, sometimes taking months, right, going back and forth, returning
04:06to the same image.
04:07And eventually, I had about 4,000, which I edited.
04:10And then I selected hundreds, which went to my art website, manovich.art.
04:16Now, even this reduced number of images, let's say 20,000 to 4,000, it's still bigger than what typically
04:24any modern artist, any 20th century artist would generate, right?
04:29And historically, it was even less.
04:32So Vermeer, I think, did less than 40 paintings.
04:35And I think Bosch did also less than 50.
04:39So as many observed, in this situation, you're no longer just a creator.
04:43You become a kind of curator, right, of your own output.
04:47And I think this is, you know, new experience for many traditional artists.
04:51So this, so for example, if we look at, okay, sorry, it was mine.
04:56Yes, if we look at, for example, French Impressionism, right?
04:59The artists who were participating in the Persianist exhibitions in the 1870s and 1880s here in Paris,
05:05like a very median number of paintings we did over decades of our career, it was about 1,000.
05:10And many people did less, right?
05:12Between 400 and, you know, 800 and so on.
05:17Now, when you find an area of latent space, which you really like, right, you keep rerunning the same prompt,
05:26right?
05:26Sometimes without any changes, sometimes you make some small changes.
05:30And the system generates hundreds or thousands of images.
05:34And I think, to me, this is one of the interesting and kind of a bit of mysterious properties of
05:39generative AI that exists today.
05:41These images have more or less exactly the same aesthetic quality.
05:45And this is not a completely new situation, right?
05:47If you look at the generative art from the 60s and 70s, when artists had to explicitly write capital programs,
05:53right, you have algorithm, it uses some random numbers.
05:56So again, you can get an immediate number of images.
05:59And again, they're more or less aesthetically the same.
06:01And now we have a different system, right?
06:04It's not explicitly programmed, but I think it has a certain property.
06:08So as I said, we can find examples of such art production in the past, especially in so-called generative
06:16art,
06:16the way it was generated by a program where artists would explicitly write rules like many people in the 60s.
06:23But while this was not a rule for historical art, right?
06:26So typically, again, if you look at Prussianism as a kind of typical modern art example,
06:32every painting is more or less different.
06:34But with AI, it's a new problem, right?
06:37You become a curator of your own output, and it's very difficult to curate
06:41because the kind of algorithmic nature of the systems allows you to create endless outputs.
06:49And very often, nothing stands out.
06:52So it's a very different situation for curation.
06:56So that's one thing.
06:59Now, the second thing is about the style.
07:01So traditionally, artists are identified by very unique styles, especially in the modern period, 19th, 20th century,
07:09the period of individualism.
07:11Writers, artists, composers, photographers, filmmakers, choreographers, architects
07:16have been identified by very unique style, or if you prefer academic language, by very visual language.
07:24Now, for the first time, a new medium changes, challenges these assumptions.
07:28Because the AI tools allow us to easily generate media in hundreds of styles.
07:33So there is a separation between the style, right?
07:35The output, the surface, and the content, right?
07:38Because you write your prompt, and then you can say, like, you know, watercolor, or oil painting, or oil painting,
07:45plus pastel, plus fractals, you know, plus 1980s vintage photography.
07:49Or you can put a name of, you know, thousands of artists or designers, and the same composition gets rendered
07:56differently.
07:58Right?
08:02So these are some examples.
08:03Maybe it's too small to see.
08:05So we come from a particular website with library, which actually did an essay about it.
08:10So you kind of have the same, so each row is the same content.
08:13Sorry, no, I'm sorry.
08:15Each column is the same content, and then it's rendered right by AI in different styles.
08:20And you can see how the semantics stays the same, but the visual appearance changes dramatically.
08:30Now, what does it do for artistic identity, right?
08:34So over a few years, I kind of honed in on a dozen or two dozen styles.
08:40I like them all.
08:40I don't want to choose one.
08:42But this is not how the art world or the cultural world works.
08:46So the question is, if a creator doesn't need to develop a unique visual language, how would she separate herself
08:54from other curators?
08:55Yeah, curators.
08:57So you may be tempted to say, well, art is not just about styles, it's about ideas and concepts.
09:03This is certainly true.
09:04But if we look at the history of art, I think we'll find that successful creators, the people who entered
09:10art history, combined particular content with particular form,
09:15and, in fact, form was often more important, right?
09:18So for centuries in the European tradition, people painted crucifixion and other things from the Bible.
09:25So the content was kind of limited.
09:27The content didn't change.
09:28What changes, you know, the style, right?
09:30Between Bosch and Bruegel and Raphael and Watteau and so on, right?
09:37Again, I haven't done the count.
09:38It would be interesting to calculate if it's possible, right?
09:41To quantify other more unique kind of content, more unique compositions, and actually striven styles.
09:47But I think, at least in a kind of Western tradition, with people who entered the canon, with people who
09:52entered, you know, Louvre and Pompidou Center,
09:55we all develop very unique visual signatures.
10:00So, if now AI kind of, in a way, makes this relevant, like, how are we going to react to
10:06it as society, as artists, as curators, as collectors, and so on?
10:11Okay, so the next one.
10:14AI learns the logic of our world, thinking and behaviors, the everyday expectations embedded into millions of pages and billions
10:24of images.
10:25It learns how the normal world is organized, what entities it has, and how its entities normally behave, right?
10:32So the world consists of solid objects, whatever entities call people, like their body is bent,
10:38we walk in a particular way, where your muscles in the face work in a particular way, and this is
10:42what AI learns, right?
10:43Not with physics yet, because we know we might not arrive at the kind of world models completely,
10:49but at least it learns how it works visually, the appearances.
10:53So in this respect, generative AI can be structurally understood as the opposite of artistic avant-garde,
10:59because artistic avant-garde, and again, right, the place of avant-garde was Paris, 120 years ago, right?
11:06All these artists living in Montmartre, and gathering in cafes, and reinventing modern art, right?
11:11Picasso, Braque, Soutine, and so on and so forth, Giacometti.
11:16And what avant-garde tried to do is precisely go against the logic of normal bourgeois world, right?
11:22To defamiliarize the familiar.
11:25Here we can recall another sort of, right, important artistic movement from Paris, the French surrealists, 1920s, 1930s,
11:33who praised exactly such improbable images.
11:37So in a way, right, what modernism, especially Impressionism, valued was images which we can call statistically, right,
11:45less likely to appear.
11:47So in a way, it's opposite, right?
11:49Because AI, the way it's trained, right, and also the way training data, you know, is used, right?
11:55From a web, often, as we know, it tries to do the most predictable thing, right?
11:59This is how the next talking prediction works, right?
12:02With some randomness, but then, of course, you can change temperature, right?
12:06But more or less, it tries to predict what would be most likely response to the situation,
12:10which is exactly opposite to what avant-garde art was trying to do.
12:14And then, so this is like a reminder, right, of like how conservative art, of art we now think as
12:22conservative used to look like.
12:24So this is French Salon, right, second part of the 19th century, and there were so-called academic painters who
12:30would paint with highly finished,
12:32today we'll call them airbrushed or photoshopped images, and traditional scenes, and these people were very, very famous every day,
12:38and now we're all completely forgotten, right?
12:44And we can say that many images can learn, right, or get fine-tuned to privilege the same aesthetics.
12:53Idealized lighting, anatomically consistent figures and faces, recognizable objects, polar surfaces, legible spatial depth, and clearly defined contours.
13:06And here comes a big question, right?
13:08So people constantly ask AI, right?
13:12And also, this is what artists keep asking, right, or creatives.
13:16Well, can AI invent something radically new?
13:18But I think the more important question is, why?
13:21Why do you want to invent radically new?
13:23Do we actually need anything radically new?
13:25Yes, in business, we have this rhetoric, right, of disruption, right, which is a result of globalization,
13:30because in the 20th century, the company is more or less also a conservative, and now, I guess, you know,
13:35to win, you have to do something radical.
13:38You have to establish a new brand category, but in culture, right?
13:42I mean, we had 100 years of modern art from French imperialism up to 20th century.
13:47We had 30 years of computer graphics, and together, like, I think with every possible visual image,
13:53every possible static was already produced, so nothing is actually possible.
13:57I think where real innovation can still happen, and it will happen is in the interfaces, right?
14:03How interface to the image, or maybe we bypass the image, right?
14:06And we have some other holographic representation or representation which directly gets, you know,
14:11beamed into your retina.
14:12But asking AI to produce something visually innovative, in a way, is besides the point.
14:17So I will conclude, right, in two minutes.
14:20And again, this is just to remind you, right, with examples of famous French imperialist paintings,
14:25which were rejected by Salon, and we remember them because of a radical nature.
14:30But this was 150 years ago.
14:33So is it possible to do anything as radical today in real image or moving image?
14:38Probably not.
14:39So, but at the same time, right, because we are conservative by our nature, right?
14:44You know, we live in the 21st century, but we often think in the kind of 18th century style,
14:50we still believe in the romantic idea of artists and so on.
14:55And even me, right, who is critical of all these traditions still keeps asking,
14:59because I can't help it, what can I do with AI which will be radically different,
15:04which I couldn't do it before?
15:06So we know it's a wrong question, but we have to ask it anyway.
15:09But I will conclude by, I think, a more important question, right?
15:13So now we acquired this new superpower.
15:16AI is no longer just a tool.
15:18It's a partner, right?
15:19It's an intellectual partner and a visual partner.
15:21And you have dialogue, and you develop new projects, new work for this dialogue.
15:27And also, right, we had another development in the last year, right?
15:31With web code, you know, web design.
15:33So now people can write their own tools.
15:35They're no longer bounded by existing software.
15:37So we really have a whole new level of superpower.
15:41I mean, I've been using digital media and working in this field since 1984,
15:46so it's over 40 years.
15:47And I would say it's the biggest technological revolution which I experienced,
15:51even bigger on the web.
15:53So how should we use our new superpowers?
15:56What we should do?
15:57Should we just continue doing more projects?
15:59Should we continue writing one novel for 40 years?
16:01Should we write a new novel or a new scientific article every day?
16:04I think these are the most interesting and important questions, right?
16:09How to use these new superpowers?
16:11How to meet the challenge, right?
16:15So you have this, right, kind of magical creature which says,
16:19okay, I'm here, right?
16:20I'm out of a bottle.
16:21I can fulfill all your desires, one desire every 10 seconds.
16:24So what do we do?
16:26Thank you so much.
16:41Okay.
16:42God, I've got so many questions for you.
16:45I even started, like, writing new questions while you were talking.
16:49um and one one that struck me love when you when you were saying that um ai is an inherently
16:56conservative medium so i i i'm gonna ask a fairly big question but hopefully it's if you'll follow
17:05what i'm trying to do with it i think most people agree that they don't like the path that the
17:12world
17:12is currently taking um and if all the ai can do is just remix the past how can we think
17:20our way
17:21out of the situation we're in now how do we imagine a future that's better if all we're doing is
17:27tinkering with history well i think your question right it's actually two questions because one
17:32question once one application would be to think about images art aesthetics film and now one is
17:37about the world in general sure well so i think we have to ask what is new right so or
17:43what is
17:43creativity you know margaret borden famously distinguished three types of creativity
17:48and not like one is better than the rest and the first one was about remixing yeah which things
17:53exist and you know we had great things in the past right great economic theories great societies
17:57there was this thing called democracy of course it only worked on the scale of like right few thousand
18:02people yeah um so i think there's nothing wrong with inventing new which takes best elements of
18:09what already exist and just putting them in a new combination it's new it doesn't mean right it
18:14doesn't mean we always have to have this radical creativity yeah so as i get i mean i was also
18:19very
18:19much avant-garde when i was young and that's why i invented digital art and now like i really like
18:24paris i really like you know the kind of conservative right conservative regions like i prefer some
18:30maybe the six uh to uh to maria you know which is full of young people of tattoo so i'm
18:35i'm really
18:35into i'm really into so that's why i like ai right yeah yeah yeah it's a medium for all the
18:39people
18:41yeah paradoxically perhaps yeah it makes sense um and and also uh i mean you were saying that um
18:49you know ai is a particularly exciting technology for artists why what what what makes it meaningfully
18:57more interesting than like microsoft paint yeah well so i think right what i would say for me
19:03personally so there are other reasons so first of all of course i can execute with this prototype
19:09many more of my ideas right so as an artist or any creator you have this kind of pile of
19:15ideas which
19:16gets bigger and bigger like you know recently i was able to find an open profile which shows ideas
19:21from 2001 i'm actually afraid to read it because there are probably so many ideas which i'll never
19:27execute but maybe now i can yeah right so that's the exciting part the difficult part is that every
19:32medium has its limitations and the limitations of genetic ai they get different because for example
19:37if i have camera right like iphone it's harder to take good pictures in the dark but i can still
19:41do it
19:42with ai because of how trained some things are easier to generate other things are harder so it's
19:49really a primitive medium and you can have to fight it yeah and that's difficult frustrating but also
19:55it's fascinating yes yeah absolutely so something that is clearly different about generative ai is that
20:02you can just make because you know there's the the kind of thing about art where it's like you just
20:06need to get a lot of clay down and then you can start shaping it so you can generate lots
20:10and lots of
20:10stuff with ai and then kind of play around with it what's the benefit that comes with that and
20:18what are the what are the challenges like when you're navigating the quantity as an artist yeah
20:23well so i think i did right so thank you for asking i think right one of my points was
20:27about this
20:28so you become a curator of your own work and actually it's very it's harder to be curator right
20:33because okay if you're good artist or designer or a photographer i mean right i mean photographers
20:38i mean usually before before digital right a photographer would take 100 000 photos in a year
20:45and then if we're lucky we would select four or five to go into our kind of website
20:50and if we're really lucky we would say one is one of my best photographs yeah and it's probably
20:55exactly the same but it's actually harder i think it's harder to curate yeah yeah so that's i think
21:01the challenge like the scale is actually not necessarily a good thing so you know it may be
21:05better to be a vermeer who actually made like 38 paintings and you know everyone is special
21:10so can we make ai which just allows you to make one painting in five years but it'll be interesting
21:15yeah i mean as an editor i appreciate the appreciation for curation definitely um and i guess
21:23you you were sort of setting this in the context of previous waves have there been
21:29any technological developments that you would compare to this so
21:35so there has been some publications including one by my student where she compared like a public
21:41reception about ai she looked like hundreds of new york times articles in 2223 to become reception of
21:47photoshop i mean reception of photography in the 19th century and she found that there is maybe five or
21:52six themes which always come up in the public discourse like is it going to are people going to lose
21:57job
21:57is it going to lead to this killing uh can you make art of it because machines doing art
22:02yeah so we are most of these ideas i would say but we are definitely right some some new issues
22:08yeah because medium is different but most of them i would say so we kind of project the same fears
22:13every time yes enormous technology is almost irrelevant so if something new comes up if we have some
22:19brain interface tomorrow it'll be safe people are going to lose their jobs yeah artists are going to lose
22:27waves of gatekeeping because you know the reality is that everyone is inherently creative but of course
22:32some some people don't want you know they just don't want the competition i guess i don't know
22:38um but you but you see that as well i suppose um and you've talked a bit about like the
22:44actually do
22:45you know what never mind that let's talk about jobs because that is a fear that comes up people talk
22:50about this all the time what what's your take on on the sort of future of artistic jobs so there
22:56are
22:56very authoritative publications right you know by a team of scientists or places like stanford
23:02etc and um which i think look at what happens with job in the kind of history of automation
23:07right because this is not like now we have automation of thinking automation of culture but before we
23:13had automation of you know manual labor and automation of you know file keeping and uh counting right in
23:191940s and they claim that actually of course what are changes but ultimately more new jobs are created
23:27but personally i lived right through this period for example i was in 96 97 living in l.a and
23:33hollywood was struggling right adoption of digital media and basically have its old editors who didn't
23:38want to learn it but realize i have to learn it and use it otherwise i will not have a
23:41job
23:42but when new job new jobs appeared like motion graphics and one thing about technology entertainment
23:48which is our track because every company needs to produce endless amount of content now right
23:53you have to make a new post i know every 10 minutes etc so uh yes ai systems can do
23:59more and more
24:00of it but we can don't trust them which means actually there's going to be more jobs for creatives
24:04in culture industry yeah yeah yeah it's in even training roles supervision role editing role
24:10in fact we're going to be more need for people like you yeah maybe even for me yeah yeah no
24:15i i i hear you
24:16on that because you know sometimes people talk about because obviously ai is very good at coding
24:19now and people will say that um you know oh software developer jobs will disappear but it's the
24:26opposite they will increase because people are making more code without knowing what they're creating
24:31so they will need more people to understand what they're doing i don't know if that applies to art
24:35as well well it's also you know like as i said uh maybe it may change in the future but
24:39often right
24:40you you know you i spend a few hours i kind of hit this point and latent space which of
24:45course a
24:45metaphor because not one point even i generate 300 images and when i start editing one and when i
24:51spend days editing one or if i even if i write let's say i write an abstract of my talk
24:55of ai
24:56i get first version in 30 seconds and then i spend two hours editing so my experience i'm actually
25:01spending more time because if i write it myself like i don't make certain logical mistakes yeah but
25:07the i write it it looks good even though you wake up next day you say what is this so
25:11it does save
25:12time but nobody is saying it's safe it's safe a company like it may company 500 more efficient
25:18yeah yeah usually 20 30 percent again it can change in the future yeah but what's interesting because
25:23it's thinking but it's not thinking like us right yeah alien thinking and you have to clean up after
25:28the alien yeah yeah yeah yeah really kind of makes things a bit it's very efficient but it makes
25:33things a bit messy yeah can we talk about the word slop because the word slop oh yeah people
25:39people will refer to uh anything ai generated as slop do you think that that's a fair word to use
25:46as a historian of art and media like i'm not so convinced because if you look at like typical
25:5218th or 70th century paintings right or if you go to like some you know some city like what was
25:58the
25:58city where uh mozart right was working oh yes yeah yeah we don't say yeah we don't say slope it's
26:05a
26:05kitchen but it looks really like bad taste to us so i think it always existed right yeah yeah and
26:10i
26:11think we have to understand that the images which are floating on the internet were a great images by
26:15professionals the professionals often can't even show where it was work because they were like under
26:19contract so you have images you know done by people who never get art training it's the same thing
26:25which happened with instagram as selfies yeah 2013 everybody thought instagram was all selfies
26:29if we do the you know actually analysis and turn out it was only 20 30 percent so i think
26:34the slope does
26:35exist but it's only one surface out of many yeah yeah yeah i think that that's fair can we also
26:42talk
26:42about the nature of truth because obviously as a journalist that's something that i'm interested in
26:46and obviously you know now we're seeing ai being embedded you know not just not just through um
26:52as a tool that artists are using but at the layer of just even the photo you're taking there's there's
26:58you know ai capabilities in your google phone so when you take a photo of something it isn't quite
27:05the thing that you are actually photographing what how how how do you feel about this as an issue
27:11like what can we trust i actually think i actually think it's incredibly interesting issue
27:15but personally i'm less interested in the question of trust i'm actually
27:20interested just a bit of a different angle for example this morning i was watching kind of
27:23interview which we did with some people from apple so in the new release of ios 27
27:29we're going to have new features where you take a photo and then they turn it into a gaussian
27:33you know a gaussian splat right which means you can rotate it change you know camera
27:38and then the people at apple spend miles debating ethically like how much you can change it so for
27:45example if you move a camera and we decided we will only allow users to move camera 20 percent
27:50like this yeah yeah because the idea is to improve your memory but not to allow you to generate new
27:55memory yeah so what does it mean to improve your memory right and where do you grow over borders
27:59so i think this is kind of gray line where something becomes something else yeah yeah so i think for
28:05like
28:05through for ethics politics people lying that's definitely maybe more binary but for a lot of our aspects
28:12of human experience because we are analog analog creatures yeah it's like this kind of gray zone
28:17and i think this is very very interesting right yeah for sure so we're going to remember this event
28:21and when like somebody says okay we took you a picture and we realize no actually we that's not
28:26how we stand we were standing different position yeah yeah right so things like that yeah yeah yeah and
28:31is there anything that worries you about the use of ai and r so i will tell you i'm a
28:36very strange
28:37person i'm kind of like uh maybe a bit of a pervert uh culturally right because you know i come
28:43i was
28:43born in the soviet union right in the ssr where everything was gray with some red slogans and things
28:49were get very boring and then uh i came to america and i look at i look at the world
28:54as an alien alien
28:55sociologist and i'm just interested to see what happens uh and this may be very very strange
29:05uh i'm actually not interested myself to make the world better yeah because i think world will at
29:11least for a long time will always be good and bad like it always was yeah i mean i do
29:15think with life
29:15of billions of people improved tremendously in the last 40 years right education health and so on
29:19so i'm just interested to observe it and point out paradoxes and complications uh and you may hate me
29:26um but that's who i am and maybe it's okay to have like a freak like this in the room
29:31yeah i think
29:32it's important to have more freaks in more rooms to be honest um interacting with each other which
29:37you know i do have a kind of politics my politics is when i teach students or write my articles
29:41to
29:41simply first of all help people think clearly give people more terms and concepts to understand their
29:47practice and also just to show them what alternatives yeah so for example looking at the eye in the context
29:53of
29:53impressionism or french realism it gives us more ways to think about the eye as opposed to if you
29:58only compare it for example to photoshop right yeah so the politics of intellectual and cultural
30:03diversity yeah is that okay yeah yeah better than nothing right absolutely yeah i've just one final
30:09question um because i know we're running over a little bit um it sounds to me like you're you're quite
30:16excited um about ai because it because it is quite a transformative exciting new technology what
30:23dimensions of it do you find particularly thrilling so i would say for me personally uh is like you know
30:29as i already said right i can execute or prototype lots of ideas but to me the single most important
30:34thing about ai or gen ai at this point and two years from now it could be different is the
30:39fact that
30:39it's not just any medium like a photo camera or i'm etching or reinforced concrete which change
30:46architecture right it is actually kind of persona so when i'm creating it becomes a conversation it
30:51becomes a dialogue and not only i have a general knowledge it has a kind of media knowledge right
30:56so you can look at my painting and say okay here's the things you should change you should change and
31:00this is how so yeah it has not just cognition i think yeah it has what we can call media
31:06cognition
31:06it knows about conventions of 1920s animations and notes about k-pop etc etc so it's like a new tool
31:13right a kind of a conversation partner and immediate art historian in the same and this is completely
31:20unprecedented right yeah and as somebody who always like to talk to people that's why i can't answer
31:25very briefly unfortunately the fact that now my whole life consists of conversations with ai and then
31:31there are some outputs i find to be very exciting because i never understood all these artists who go to
31:35every studio and let's stay like alone for a whole day yeah how can you create something i would just
31:39get depressed yeah no no absolutely i think you do it's it's helpful to bounce ideas off yeah so my
31:45idea of conversation uh we were like equal in a way right because i'm better in some ways yeah is
31:50better
31:50in our ways i mean i can use another model i think it's a very it's like a new dimension
31:54of art yeah
31:55because most art was before developed alone i mean except hollywood where we work in teams
32:00uh but traditionally we think of artists working alone and now like you right it's not it's not a
32:07hollywood team it's not for five thousand people who are trying to produce a market segment but it's
32:11actually you and talking to somebody else but somebody else also combines voices of billions of
32:16people historically so it's very very amazing yeah that it's fascinating um i have so many more
32:22things i want to talk to you about but maybe i'll talk to you about later um at this kind
32:27of uh
32:27era that we're in definitely time and attention are the most valuable things that you can give
32:31anything so thank you for giving us your time and attention and thank you love for
32:35giving us your brain thanks everyone