Intervening with 2010s Technologies

Intervening with 2010s Technologies

Show Video

Hello, everyone, and welcome to this Le Random discussion about chapter nine of our generative art timeline, which covers the decade of the 2010s. I'm your host, Peter Bauman, or Monk Antony, the editor in chief at Le Random. And joining us today are an incredible list of artists who created this art history that we're going to be discussing today, and they are Tyler Hobbs, Helena Sarin, Rhea Myers, and Gene Kogan. Also joining us from the Le Random team are our co founder, the funny guys, and our collection lead, Conrad House.

So today we're looking at digital expression in the 2010s. We're taking a detailed look, and this is all part of the project I've been working on at Le Random, creating this generative art timeline for the last about year and a half. So it's already at about 100, 000 words and 1000 moments. And we are just releasing chapter nine of the timeline, which covers the 2010s.

So the timeline goes back to 70,000 years ago. In chapter one, the pre modern era, we've had a talk for every chapter except for chapter six, we had two. So this is our 10th talk on chapter nine. It covers, broadly speaking, two enormous shifts in technology and art. And they are the rise of AI and its capabilities in the decade.

And also the invention of nfts. Blockchains started the decade before, but they. And they really took off the decade after. But we are going to look at how this decade was instrumental in their development. And of course, the development of nfts and the proto nfts that led to the tokens that we know of today. And on top of these huge AI developments like convolutional neural networks and gans and deep dreams and diffusion models, there was also broad trends in this decade, like social media 2.0

becoming an even bigger deal, and also mobile technology 2.0 becoming an even bigger deal. So it's a hugely important decade.

So many themes, so many trends we can really get into. So I'm very excited to welcome our guests who we really want to hear from. And the first one is Tyler Hobbs.

Tyler is a visual artist who works primarily with algorithms, plotters and paint. He began his generative practice in 2014, so ten years ago, and has been, of course, active in the space since then. His artwork focuses on computational esthetics, and we did a whole interview on algorithmic esthetics that we put out last year, how they are shaped by the biases of modern computer hardware and software and how they relate to and interact with the natural world around us. His practice involves programming custom algorithms that are used to generate serial visual imagery. And like I said, Tyler got his practice started more than ten years ago, so really interesting to chat with him about that first part of his career that covered a lot and especially the theoretical foundation of his practice. And our next guest is Helena Sarin.

She is a visual artist and software engineer who has worked with cutting edge technologies at places like Bell Labs and as an independent consultant developing computer vision software using deep learning. While primarily in tech, she has also explored applied arts like fashion, food styling and photography. Her artwork remained mostly analog until she discovered gans in the second part of the decade that we're discussing today. And these offered an adventure, a new model of creation, of employing her datasets and receiving inspired and surprising unpredictable results. These days she calls herself an engineering artist and her main interest is generative pottery.

Our next guest is Rhea Myers, who is an artist, hacker and writer whose work places technology and culture and mutual interrogation to produce new ways of seeing the world as it unfolds around us. She began investigating blockchain as early as 2011 with a blog post striving to be the first bitcoin artist. And in 2014 she started making conceptual blockchain art in the vein of Marcel Duchamp and crafting theory as well as narrative. And then our next guest is the incredible Gene Kogan. Gene is an artist and programmer and educator exploring autonomous systems emergence, collective intelligence, generative art, and computer science.

He is interested in advancing scientific literacy through open source toolkits for creatives that stress creativity and play. This includes building educational spaces which are as open and inclusive as possible. Gene was one of the most influential figures in the decade in terms of reimagining machine learning for creative expression beyond demonstrating the power of the new tools himself. Through his work, he was instrumental in teaching the world about these tools.

So what an incredible lineup of people and artists and thinkers to talk to today. Our first set of questions are more for the group, so let's get started. What do you think the name for this decade should be like? Just to start on a very basic level, the eighties I named that chapter the PC era. The nineties I called the net era, the two thousands I called the tooling era for processing and open frameworks and a lot of those instrumental tools. But what would best describe this era? Do you think it should be the AI era? Do you think it should be the on chain era? Is there something else? I'd call it the GPU era because sort of the things that people noticed as the decade went on were the things that were using GPU's, whether it's blockchain for hashing or AI for neural net simulation. So sort of if we view it in terms of technology, and that obviously absolutely was the PC era, it's tempting to call the cloud era because that's what scaled all of the apps that created the society we now have.

But in terms of the focus of the art, it's very much on what you can misuse GPU's for. Yeah. For me, a couple of things that really stood out as big changes in that time period are social media in particular for visual artwork, generative artwork. I'd say Instagram and Reddit were the two biggest changes for me. They were very influential for me in terms of what art I was exposed to, in terms of how I was sharing my art and eventually how I started to make my living from artwork. And I think digital artists very natively adopted these formats and locations. And so I think those had a massive, massive effect.

So potentially, there's an argument that, that you could call it sort of a social network decade as well. I'd maybe concur on social media. It feels like it really just impacted how also for me, just how I got exposed to things. So, like for machine learning and AI before that, it was really, you have to be kind of part of an institution, maybe deinstitutionalization is also something I would kind of associate with it. So suddenly there was kind of this, you know, AI twitter, you know, that sprang up in the. It was possible to kind of be, you know, not affiliated with any specific, you know, research institution or grad school and keep up with what was going on.

And, you know, I would also maybe open source code is a big part of it, too, that also, I mean, been around for longer than that, but it really grew a lot in the. Does anybody else have any or want to share? Have any of you met in person before Orlando? Are you friends or. Yeah. Does anybody know each other in this chat? So I will continue with Gene, if I may, because Gene actually, besides his MOOG, he was instrumental in many cases. I mean, I gave my public talk at NYU on his invitation, and also then it was like, 2018, and I tested, like, waters with my approach to working with the gans.

And this actually became a foundation for the article that Jason Bailey and I wrote in the end of that year. So, yeah, I mean, Jin is. I consider Gene as a friend and, like, mentor and really kind of influencer. I know everyone via social media. I know parasocial relationships get a bad rep, but it's like, it's an epistolary era. It's sort of a network of or kingdom of letters that we live in, just electronically.

And I have met groups of people that I know online. For me, notably the gray area blockchain show, was that 2018 2019? I met lots of people I'd previously only known online to the point where I was confused why we were sort of all hugging each other like we hadn't seen each other for ages. And I was like, oh, we haven't actually met physically before. Okay, yeah, that makes more sense. Yes. Yeah. Maybe more evidence for this decade being the social media era.

And, yeah, I think Gene has had that, I'm sure, similar effect on countless coders today. And over the last 15 or 20 years, we might have touched on many of those already. But does anybody want to include or mention any lesser known themes from the decade? I mean, I don't know that it's non obvious. It's pretty obvious, but the rise of mobile devices, phones as well, obviously, this pretty quickly became the most common viewing format for a lot of what we do.

And so, yeah, that has a big impact on what kind of work is successful, and especially if you're doing something live, live generated, for example, you have now particular hardware requirements that you have to be able to work with. So, yeah, there's big, big impact from the rise of mobile phones. Yeah. I mean, from creation of the work to how people procure it and consume it. Yeah, everything has been affected by mobile.

And also this decade saw mobile phones really go completely mainstream and global, and then also things like the iPad, too, and its effects. The third question is about everybody in this calls, I think, fascination with randomness and chance. So, Rhea, your early work explored cybernetic principles and you were in some of your early blockchain work, like face coin. You were, I mean, literally searching for. In randomness, you were searching for human elements like faces.

Helena, you've said Gans possess, and I quote, a certain unpredictability that inspires, unblocks and creates something special. Tyler, randomness was the topic of the very first essay that you wrote in 2014. And Gene, you've long studied and taught and devoted your career, essentially, to processes like randomness and emergence. So how do you think the continued secularization of society has impacted your views on randomness? Is there a connection between your interest in randomness and secularization in general? In generative art, randomness is obviously one of the touchstones.

The other is rules. And sort of, you can either create entirely rule based or sort of largely random, with a few rules to interpret them, or sort of interleave them however you want, and sort of, it looks like a crutch or like a failure to really get to grips. What one is dealing with when one makes generative art and says, hey, look at this amazing pattern or structure or this thing that I found, and you're not dealing with the question of randomness. And sort of paintings don't all have to be about pigment. Pigment. That's not what I'm saying. I'm saying that, yeah, the question of why randomness at this time, in opposition to concepts of order, does seem to have more of an impact post financial crisis.

I mean, for me personally, I sort of always found the world very random and confusing. So sort of using randomness in a way that I controlled was kind of therapeutic or at least helped me theorize, you know, my experience of the world. And sort of people get very upset when we talk about randomness and computers, because computers cannot generate actually random numbers. They can just fool us with patterns. We can't guess unless you plug in a hardware key to get actual quantum noise or something. But, yeah, I don't think the absence of broader religious or economic narratives requires a concept of randomness, because randomness still brings in the concept of variety, which we also see in cybernetics.

And we're still in a sort of very singular political era where there's sort of less and less variety in the options that we're faced with. And sort of randomness, where it's sort of profoundly embraced, can be an escape rather than a distraction from that. Yeah. Thank you, Rhea. I mean, I think you very elegantly rephrased the question as why randomness at this time? And much better than my wordy one.

But, yeah, I think that's an excellent rephrasing. Yeah. From the personal experience. I'll start with a quote by Richter. Chance does it better than I can, but I have to prepare the conditions to allow randomness to do its work.

And in my case, like, to sort of delineate this randomness, I used my own data sets. So kind of like divide and conquer, in a sense. So in tandem with randomness that gans brought, I used, like, the limited data sets to kind of work with this randomness too. Like these conditions, for me, randomness is really just a necessary ingredient in exploring things systematically.

So for me, it's really like the things that I found most interesting in nature were sort of systematic, so by biological systems or geological systems, and of course, those involve randomness, because if there was no randomness involved, then it would always do exactly the same thing. Wouldn't be anything very interesting to it. So really, to have something systematic that also has variety and depth to it, you have to have that random ingredient.

So, for me, in my artwork, randomness is sort of the fuel that powers the exploration of the systems that I develop. So, yeah, it's a really curious relationship with randomness. I've never really thought about it in spiritual terms. I don't have that kind of connection to it.

For me, it's just really a necessary part of doing anything systematically. I would maybe emphasize emergence. That was always kind of the thing that drove me. So, you know, a lot of small things, which, you know, in the sense random just means kind of uncoordinated, you know, doing their own thing.

But then something emerges from all of those small parts, like something bigger that's not explicitly there, but is what really seems to be there is always kind of, I think, at the heart of generative systems, more so than randomness itself. Part of asking that question is, I think, globally, in the US, or in a lot of developed economies globally, you saw secularization really increase over the last ten or 15 years. I'm wondering if that's coincided with this growth in technology, how our relationship with tech is changing.

But, Conrad, I think. Did you want. Yeah, I think since we kind of hit the halfway point, we can maybe start jumping into some direct artist questions and can hand it over to Jan. Helena, your practice was entirely analog until you discovered gans at some point between 2014 and 2018, we believe, after which, like, generative models became your primary medium. And when.

How did you first, like, discover these gans? And what about them, like, prompted such a. Such a transformative artistic shift? So randomness happenstance because I was doing some training of object recognition. And at the time, the papers, specifically about, like, cyclogone paper appeared about synthetic data. And actually, this is what I did, like, for my consulting gig. And, I mean, the results were outstanding. I was so surprised, like, how, well, like, back in 2016, neural networks could generate pretty much plausible data, which was good for kind of augmenting the data sets and also kind of, like, even surprise the customers.

So. And then because I had, like, so much digitized work of my own, I decided I should try to give it a try, like cyclegan, to do this domain to domain translation. So it's basically psychogen is style transfer on steroids. So if you're familiar with style transfer, and I never did it. I mean, I kind of, like, was oblivious to the whole machine learning kind of art community.

And that was actually a blessing because I kind of achieved something being, like, an idiot and. But I didn't have any imposter syndrome or anything because I didn't know people even do such stuff. And then, like, one of my watercolor mentors, she told me she doesn't know anything about, like, computers or anything, but she felt there is something in it. And then I kind of, like, decided, okay, to help with it, I'll post on Twitter, and this is how the story began.

Were your first post? Those ones that I think are from February or March 2018, were those your first works? No, I mean, originally, all my art was posted on Flickr, and this is, like, end of 2017, when I kind of, like, dared to post stuff. And then when I kind of, like, started searching for, like, if people doing these things, I discovered this small, at the time community again, Gina, including, and Gin, actually, like, retweeted one of my stuff for you. And at the time, I did this kind of, like, food related things because most of my photography was around stills. And food, like, a lot of Instagram today. Maybe we can move to Rhea and Tyler about your beginnings in this decade as well. So, Rhea, I know you began investigating blockchain.

I've read in a post where you said you wanted to be the first bitcoin artist, but it wasn't until you moved to Vancouver in 2013 and then working with, I think, blockchain and more closely in 2014 that your interest in the work really took off. So I'm wondering, what was it about Vancouver that made blockchain and bitcoin and that scene so appealing to? How was it different then from what I would imagine is quite a different community today with the laser eyes and some different leaders of the community more recently as well. So, yeah, I'm curious, what was it like ten years ago? Vancouver had and has a curiously large blockchain community. It's sort of. It's a cascadian city. It sort of has canadian healthcare, but sort of west coast libertarian leanings to some of its inhabitants. Politics.

And that makes for an interesting mix. And it means that the tech scene and the culture scene here were both very interested in cryptocurrency from very early on. And it was the combination of those two of sort of companies like Dapa, who I later worked for, who created the RC 721 standard, and cryptokitties for nfts and cultural spaces like decontrol, which were much more community and sort of hacker oriented. Sort of mixing without differentiation between sort of their objectives or their social scenes meant that you had a very fertile ground both for the ideas being pursued in the technological infrastructure and then being pursued not almost, but actually utopianly within society.

And that completely fascinated me because I hadn't seen anything like that since the early web in the mid 1990s. And so that was irresistible as someone who really hadn't got the early web or early net art to sort of be faced with almost a second chance to sort of observe a technocultural phenomenon like this as it was growing. And the things that people were interested in of rules and ownership and belonging and values had been key to my art since I was at art school. So it was a very, very easy decision to make to sort of switch my primary focus from bots and 3d printing and other early 2010s tech to the blockchain as something that became, as it does for many people, a rabbit hole that I disappeared down and haven't really emerge from again yet.

Yeah, like how you kind of mentioned that it kind of give you, like, a second chance to experience, like, a new form of kind of media art with having this new whole emergence that you can kind of forefront and maybe we can move on to Tyler now. Maybe we can kind of have a question for you. On your introduction into the 2010s, you really started to become a generative artist around 2014.

And you had a lot of your very early essays around this time kind of talking about how randomness is a fundamental part of your practice. You mentioned a lot of things by John Cage and Sol LeWitt and a lot of these early essays, like what programming brings to art and how that kind of highlights the new artistic possibilities within kind of generative art. So maybe can you touch on how these early essays and early concepts became so foundational to your practice and then kind of how your practice evolved within the 2010s going into the 2020s. Absolutely.

So just for context, you know, prior to making generative art, I was a traditional artist doing drawing and painting, and I had been doing that since I was young. And 2014 is when I really, just the first time I started trying to create artwork through programming. And I benefited a little bit, like Helena mentioned, of not being really aware of the pooling at that time or the style or the existing body of work. And so I was able to kind of come into it in a fresh way with my own ideas. But what was really fascinating and what kind of prompted me to write those early essays is how big of a shift it is mentally going from creating a drawing or painting to working systematically. And I, through generative art, it really changes everything about how you approach the work, sometimes in very frustrating ways and sometimes in very empowering and intriguing ways.

And to me, it was, you know, I just. I felt like I had stumbled upon something that just had an immense potential, and I hadn't really seen people fully tapd what it was capable of. And at least to me, I wasn't super up to date on what, for example, Casey Reese had been doing up to then. But to me, it felt like there was so much potential for kind of moving up the level of abstraction and starting to not just make a specific work, but to think about what makes work good in general. How can you make a body of work in kind of one go or a series? Obviously, there's some historical precedent for some of those things, like John Cage obviously worked with randomness, solowit obviously worked with rules, but neither of them, I could be wrong, but I don't think either of them really zoomed out quite so much to the systems level about how far it could potentially go.

And I think computers are part of what makes that possible. You can create so much more complex systems, so much more specific systems, and of course, they take care of all the labor for you. And so I was at the same time both really fascinated in what does it mean to work generatively? How does that change the artistic practice and the potential goals? And also with the computer as a medium and specifically programming as a medium, I had seen a lot of digital painting. That's super common. Even I wasn't particularly interested in digital painting, but I had still seen lots of it.

But I really hadn't seen people working through programming. It had never made its way to me. And there's so many unique in particular and powerful capabilities that come from programming. And I also worked as a programmer, so I was very aware of just how much you can accomplish through programming. And so those early essays were a bit of a thought of experiment for me, of, you know, really what is programming unlocking that somebody who's working with a pen or a paintbrush doesn't have access to? And that's what really compelled me to.

To create a lot of my early work and to, to write about those ideas. I believe, Tyler, in the 2000 times that you have been focused, like, all along the century, like on procedural generative art, while Gene, Helena, like, they were exploring AI arts, Rhea was exploring blockchain technology. Were these technological developments on your radar at all, and did you consider exploring them? Yeah, I did a little bit. I actually did a Moog where I wrote, like, early neural net implementations, like a little convolutional network and things like that. So I was familiar with those technologies, but they never, they didn't interest me quite so much as a.

As a potential artistic medium. And in fact, I had played with those well before I started making generative art, maybe by three or four years. To me, I love programming and working through gans and neural nets and things like that.

It removes you from the creation a certain distance. That programming, to me at least, feels like a much more direct relationship with what's being created. Obviously, I'm greatly generalizing here, and there's a lot more nuance to it than that, but I really loved working directly with the code and trying not to just make fuzzy tools, but to make very specific rules that my artwork would follow. And so programming is what allowed me to do that. That's why I think it was more attractive to me. It's actually interesting angle, because I'm the other, quite opposite to this, because maybe because I was programming for a living.

And actually I got it. I mean, for me, the whole decade was about, like, generative art. And I started with, like, K Syria's book, his first book published in 2010. Exactly. And then I learned about generative art. I took a workshop with Casey in 2013, and I tried to kind of convince myself that I can be good in it, and I didn't like it at all. And that's why Gans was a kind of like a savior, because I got gens, whatever, like this fuzziness that attracted me.

So, yeah, it's funny. That's amazing. Yeah. Yeah.

Like you kind of, we had mentioned, and Tyler, you mentioned moogs, and we talked about one of the potential themes of this decade being open source code. And I wanted to ask you, Eugene, about your book, machine learning for art, this educational resource for artists who want to explore with a lot of the tools that we've been talking about today. I just wonder what inspired you to first create those resources and how do you see them impacting the future of AI art? There's already a really great workshop and teaching culture in generative art, and at NYU people, the sort of creative technology world, there was a lot of workshopping, and so it's just kind of plugging into that. At the time I was coming at it from this, I had a little bit of background. I studied machine learning. So I was kind of coming from this engineering world.

And I found that there was more and more interest from people in creative technologists, generative artists in AI that there wasn't really a whole lot of materials out, wasn't a lot of materials at all in general, but, but certainly not the kind that were focused on a kind of, you know, an application level, you know, you know, creative explorers level view of the technology. And so it just kind of naturally felt like something I could do that would, people would respond to. It was also a way of kind of, you know, integrating myself into, into this scene. I could go and give classes and meet people, and it was, it was really just like a big part of my, my way of actually, like, making this my life. So, because, you know, otherwise, it was kind of just, you know, in front of the screen for or reading papers or something.

And it wasn't really, you know, that. And I would also add that there was just a lot of invitations, started getting invitations because there was just kind of this desire. And so I just kind of responded to that.

And, yeah, and it was a big part of, big part of what I did for a long time still is I'm not giving as many workshops as I used to, but still, it's been kind of a long time coming to making those. I think Helena has talked about previously in the, in the chat, some of these early workshops. Just wondering if Helena had any follow up for, I guess, machine learning for our or any of these kind of other workshops. If there's any other kind of expansive inspirations that came from this at the time, the community. I mean, that, that was like, I mean, this is what I'm missing these days. And because, I mean, I, I'm not against, like, people coming, like, in this avalanche of AI art, but at the time, there was really kind of like we knew each other.

We met, like, at the excellent IO festival, which is not anymore killed by Covid or whatever. So, yeah, I mean, IO definitely was like something transformative for me as well. Like, I met like so many people there. So like Refik and mario Klingemann. So it kind of like when you know people, you, you kind of like inspired by them, you make better in a sense.

I mean, inspiration, not in terms of copying, because like you, at least myself, you basically try to do something different and it's kind of like a bit of like friendly competition. So, yeah, I mean, it's something kind of like I'm really nostalgic for. Yeah, that sense of community really kind of, I think is a driving force for a lot of these movements being so helpful for really, I guess, igniting a fire under everybody. I wanted to transition to. Rhea, you mentioned it in your start of kind of blockchain explorations that it was an easy aspect to jump into because you were so fascinated with kind of the ownership of art and the multiple aspects of ownership within art, whether that's provenance and other things. And then you've recently said that all the early stuff I did was very intentionally made, so you couldn't buy it, you couldn't sell it, you couldn't control it.

That was the point of the work. How is your approach to creating art that resists commercialization evolved over time, I guess, especially with the even more recent rise of nfts from a career perspective. Being so committed to a very european, hey, let's criticize capitalism flavor of art has not worked out very well for me. With the early work, which people repeatedly asked me if they can buy, and I have to say no, it's physically not possible. But on a level with more integrity, I had to work through that to realize that, no, the core proposition of the blockchain is the ownership and control of digital assets.

And sort of having realized that, and having sort of been kind of blindsided by the emergence of ethereum, NFTs, post crypto punks, I had to really rethink, how do I maintain a critical position to making things to sell? And. But the way that turned out to be possible was I'd already worked through that with the 3d printing work, where I was sort of taking on the role of a post conceptual artist, sort of hiring some artisans to make objects, but sort of actually, you know, revealing who these people were and creating them as artists in their own right, which they were, I must make very clear, and sort of playing with the ownership mechanisms of creative commons licenses to sort of try and avoid making myself the villain, whilst maintaining the role of the sort of name artist in that system. And so, yeah, the way things really came through were with a certificate of inauthenticity, which sort of takes that previously unownable work that I made, where I had already worked through being the bad guy and sort of ironized the anxieties of ownership and provenance that were emerging in NFT markets at the time into some art that I was perfectly happy sort of firstly, putting my name to.

Secondly, taking money for, and thirdly, having sort of people get something useful from as art. But, yeah, the sort of the dirty secret of the art world, wherever you are, is where the money comes from. And the thing that made a lot of people very uncomfortable about cryptocurrency and blockchain was you could see where the money was coming from. You couldn't sort of pretend that you lived in this carefree world of imagination and information and esthetics and are unsullied by the churn of capital. Now, you could see where the money comes from and how much effort goes into securing the products of your imagination.

And so, yeah, getting my hands dirty with that was and continues to be a fantastically useful moral and intellectual exercise. Definitely. Yeah, that's. That's very interesting, ria, I believe with your release just earlier this week, like a ten k drop, once again, you're also. Yeah. Playing into these questions.

Yes. Yeah. I mean, ten k drop. Obviously it's a bad joke, but it's also.

And lots of the projects are bad jokes, but that's just the gateway or the lure or the candy shell or the protein shell of your virus. Yeah, it's sort of that whole era of the obsessive owning and promotion and emotional and social and economic investment in generative art. Pfps on the blockchain and trying to give people a history painting in very strange form of that era and its ambitions.

I think it's a fantastic collection. But, yeah, we're now approaching the end of the hour and maybe one general question. I will point it first to you, Tyler, but please, all of you, jump in.

We're curious to hear all of your dots. A threat in generative art history is that a lot of criticism is focused on the use of computers and the coldness of computers and that. Yeah, computer art is kind of like a contradiction in terms because art is human. Computers are called. They're not human. Like, how can they even go together? So a question that we have for you, Tyler, like, how does technology, in your opinion, reveal what makes us human? That's a good question. It's definitely one that I think about a lot.

I don't know that I have the answer to it, per se, but obviously, a lot of what the computer reveals about us is via contrast. Right. It's all the things that. All the ways that the computer is different from us. It's, in fact, quite alien.

What motivates a lot of my work today is exploring the ways in which it is strange and the ways in which it is different from everything that we evolved to expect and to enjoy in the world. And it's quite obvious to me that computers aren't going anywhere. Our lives are getting more and more entangled with computers.

I'm kind of interested in this question of, you know, with that shift to a more digital life, like, what, what really is changing? What are we maybe losing or gaining? I'm not kind of strangely, I'm not, like, pro or anti computer. I just find it really fascinating, just really ripe for artistic exploration. And computers do everything, you know, they do everything perfectly, exactly the way that you, you asked them to. They never.

They never deviate from that at all. You ask for a straight line and you get a perfectly straight line. And, in fact, you have to work to get the opposite of that.

You want anything that has any complexity or richness to it, you have to really, really work for it. Whereas, you know, we're used to, in the natural world, you try and draw a straight line by hand, and it's never going to be straight. It's always going to be flawed.

You have to work really, really hard to try to make it perfect. So it's quite. It's quite the opposite end of the spectrum. And that means there's a. A lot of really interesting phenomena to investigate, I think, there.

And that's a big focus of my artwork. I kind of. A lot of things I could say that there's something about what computers can do that fascinates me, but I'll just take one angle, which is that there's a lot of dialog, debate about what AI has to say about the senos. And so a lot of it is like, to me, if you view AI as a tool, as something that can be used to just amplify an artist, an existing artist's work can turn. I just saw a friend who's an illustrator, and he was able to turn his illustrations into basically a styled animation. You can turn every illustrator into an animator.

And it's not something that just takes a few seconds. It's not something that you can work just as hard and just as long at these, because people will often say, oh, now, it's easy to make pictures. And that's true. It's easy to make pictures, but you can work really long and hard to get better results, and you can just do more. It sort of adds this extra sort of dimension or multiplication of your work. And so in that sense, like maybe, you know, seeing computers as just kind of supporting and amplifying artists is one way of maybe resolving that tension.

I would go further and talk about, you know, certain things that computers can do that's better fascinating in and of themselves, but. But maybe that's kind of. It's just a bigger can of worms. So I'm gonna leave it there. Yeah, I mean, the dirty secret of contemporary art is everyone uses computers, sort of.

I've seen painters sort of working out compositions in Photoshop. And even if you never touch a computer in the studio, you're on social media. And if you're not on social media, you are emailing your gallery. And if you're not emailing your gallery, then you've probably got a spreadsheet somewhere with all your work in it. And the moment, I remember very clearly the moment that there's this phase transition from you can't make computer, you can't make art with computers.

So there is nothing interesting about computer art to. Everyone makes art using computers, so there's nothing interesting about computer art in the mid two thousands. And that was a really, really, I'm going to say interesting, but I mean, in the bad sense moment to live through. But computers, for all their obsessive rule following. And yeah, if you tell a computer to do something, it will happily boil the oceans or turn the world into paper clips in order to do what you think you've told it to do.

For all their raw following, they do acture and to propagate that the warmth and the noise of signs of human activity. I learned this again way back in the early nineties when I was first scanning photographs and getting all the photographic grain and scanner noise on the screen and thinking, I'm going to keep this. I'm not going to try and hide this, and sort of morphing images together and getting this wonderful, unexpected, literal, sort of reddish warmth to the image that didn't fight with the patterns of pixels, but brought something extra to it.

I think what we imagine to be the sort of cold, hard, rule based world of the computer is as much a product of the humans imagination as anything else. And blockchain has been very useful for the sort of vanguard and the absolute of that view of the opposition of clean, incorruptible, perfect rules versus messy human decision making, seeing that, you know, that hit reality and the rules haven't won. So, yeah, that's been a very interesting outcome of the blockchain era in terms of how we see our relationship to computing machinery.

Interesting to think about how. Yeah, a lot of the coldness of computers came from our human input, but, Helena, was there anything that you wanted to. No, for me, it's just, again, I mean, I never considered computers being called or whatever. It was, like, always part of my, like, again, job. And actually my pet peeve right now that we have, like, nothing bad about, like, a lot of art, but it kind of, like, dulls the eye for me, personally. I basically kind of, like, pretty much stopped doing, like, digital art and move to generative pottery.

So this is my contribution to future artifacts to kind of, like, giving whatever produced essentially through designs with gans to back to humans. Yeah. Thank you. And I think even just that last comment about gans back to humans, I mean, it highlights why we do these.

These talks and what Tyler was saying about the power of thinking about contrasts. I think it's helpful to look back, even if it's looking back just a few years, I think you can already see that this decade did have a personality in it, and it had an increasing number of developments that have affected and seemingly will continue to affect our lives. And you were all in the middle of it, and in the middle of it, especially intervening creatively with these technologies.

So thank you so much for not only joining us today, but also for. For being there and thinking ahead at that time. And it's been, I mean, what a pleasure for us to have you. Thank you so much for joining us. Really been a great talk, and I only wish, genuinely, that we could keep going.

So thank you all so much. Thank you. Thank you. Bye. Thank you. It's been wonderful.

Thank you, everyone. Thank you. Thank you, everybody. Thank you, everyone. Bye.

2024-08-03 17:29

Show Video

Other news