Jason Bailey: Thanks for agreeing to an interview after what’s been a pretty wild ride this past year.
Tyler Hobbs: I’m also glad to have the opportunity to talk again. I figured it was just a matter of time.
JB: So how did you get started in generative art?
TH: I started to get really serious about being an artist in my early 20s, after studying Computer Science at UT (The University of Texas). I started working as a programmer after school at a tech startup in Austin, working on a database engine, which was a really fun job. But I was always working on artwork in my spare time.
Eventually, I knew that the artwork needed to be more personal and needed to involve my skill set. Programming was such a big part of my life, it really molded how I thought about the world in terms of patterns and processes. So I figured I needed to involve that programming in my artwork in some way. I made a few failed attempts at integrating math into my drawings and paintings — mathematical concepts and equations, and some other bad things like paintings of coding environments.
I didn’t know about generative art at the time, or I didn’t know what it was formally. Maybe the closest references I knew were things like screensavers, or a couple of digital artists here and there. But it didn’t seem like a real possibility to me. I asked myself: “I wonder if I can write a program that creates a painting?”
I’d been doing these mathematical drawings by hand — very tedious, very laborious. And I was wondering: “Can I just cut out the middleman and have the program do this stuff directly?” So I started putting together some works with Matplotlib, hacking it a little bit to create my first generative art. And then later I found out about Processing and Casey Reas and the whole tradition and started to educate myself more.
JB: It sounds like you went through the struggle of finding it yourself, inventing it in your own way. When you discovered Casey and the Processing community, what was that moment like for you, as someone trying to put your finger on this thing that you want to do?
TH: It was really interesting. First of all, I was glad that there was better software because Matplotlib is not a great tool set. But to see someone like Casey who was able to talk about the connections that went all the way back to the 1960s, about the traditions of the art form, and the ways it’s related to other traditional artists like Sol Lewitt. To me, it was super valuable to understand the work within a larger art historical context.
I was fortunate enough to see him talk at UT for a mural that he was installing. He talked about his process, about how he thinks about curation, and what he thinks is special about generative art. Definitely, his thoughts were very influential for me getting started. On the other hand, I felt lucky that I had at least done my first works with my own naive approach, because I think that helped set me up with a little bit of an independent mental space of the type of work that I was trying to do. Rather than coming at it from a computational direction I was coming at it from a painting direction. I think that was really beneficial for me.
JB: In my masters program, which was a digital art program, I had an older professor who said that the problem with generative art or digital art is that there’s no entropy, no accident. You just program it in and then it produces what you ask it to do. But as someone with a painting background, for me the two couldn’t be more alike. I’m curious about how similar or dissimilar your creative process as a painter is relative to your coding. Are they more similar than most people realize?
TH: By deeply involving randomness in the program, you’re inviting accidents to happen constantly. But beyond that, no coder is perfect, right? Often you write code that’s buggy, or you accidentally change the wrong number or something like that. And you get really bizarre results from the program. I’m certain that almost every generative artist has experienced that multiple times. And it’s highly relatable to kicking a paint can that splatters on the canvas and does something that really surprises you. It’s almost more prone to accidents than painting and drawing.
It took me a while to make the mental shift from thinking about one particular output to thinking in terms of an entire output space. My first generative works were more like constructing a singular image, which is what you do when you’re creating a painting. But generative art requires this shift to thinking in terms of the entire process and the entire system. That’s the biggest shift between painting and generative art for me.
JB: Some people separate AI artists from generative artists. I tend to think of them all under the same family tree, maybe as cousins. But a lot of the artists I talk to actually think of the algorithm, or the code, as the art. So while they will exhibit or, in some cases, sell the output; to them, their baby, their creation, is actually the code which produces near infinite outputs. So where do you put the emphasis as an artist? Where is the art? Is it the code or the output?
TH: For me, while it depends on the exact algorithm, the art is typically the entire output space rather than the code. Why do I say that? I consider myself a visual artist and coding is a means to get there. It’s like the difference between saying that a painting is the canvas and the pigment. Yes, technically it is, but it’s also the visual result.
To me, the code is not so interesting by itself. I don’t care if the code is clean or particularly novel or anything like that. Really, it’s about the potential visual output space of the generative program. That’s what excites me.
I’ll also say that, for many generative works, curation is a very important part of the process. Sometimes you’re putting works out directly with no curation, but it’s common as well to curate the work from the algorithm. So I might generate 1,000 images and curate it down to five, which is what I did for a recent set of work. And that’s also an artistic and challenging part of the process. You have to spend a lot of time thinking about how you best capture the full range of expression from the algorithm — the best parts of it — to show that variety without making it overly repetitive. So, in some cases, the artwork is the full output set from the program and, in other cases, I think that the final curated output is really the work because that curation is an artistic step in the process.
JB: Which leads me to this idea of “long-form” generative art, which I don’t think I’d heard before reading your article. Did you coin that phrase to describe what’s going on with Art Blocks and fxhash?
TH: So I felt like what was happening with Art Blocks, particularly at the time and now with fxhash and maybe other platforms, was that there was really a new branch of generative art that was starting to happen. Artists were working in a way that felt fairly distinct to me from the generative art that I had myself created and observed before.
And so I came up with this term “long-form” to try to capture what I thought I was seeing happen. And in that essay, I focused on two points. One is that rather than generating a handful of images we’re generating hundreds or over a thousand images from the same algorithm and expecting all of them to be interesting. So that’s a really large output size for a generative algorithm.
The second aspect that I thought was really interesting and affected the work was around removing that curation step. So with Art Blocks, the output from the algorithm goes directly into the hands of the collector in the form of an NFT. There’s no opportunity for curation or cherry-picking by the artist or the platform or the collector. This also kind of creates a need for a very high consistency of quality in the output — you don’t want bad outputs going to your collectors.
So I feel like this combination of needing to produce a lot of variety for this large output set and this high bar for the average quality of that output is a really interesting challenge for artists. I feel like a lot of the works that came out in 2021 broke new ground in terms of generative art. There are certainly some precedents such as live interactive generative programs where you can click to generate new outputs. But, to me, it really feels like there’s something distinct about the way all those outputs are kind of captured and immortalized in the NFT platforms of Art Blocks and fxhash.
JB: I think we have similar backgrounds in terms of starting in analog and moving to digital. We also both look up to some of the abstract painters of the 20th century. But I wonder whether that holds us back in our concept of art, because maybe the most obvious thing that software can do is be dynamic and interactive. Are we not being a bit skeuomorphic and nostalgic in the work that we’re creating that, aesthetically at least, looks kind of like hard-edged abstraction? Even though we’ve got this medium that allows for so much more, we’re still preoccupied with static images. How do you think about that?
TH: I think about this a lot as well. At a high level, I absolutely want to use the computer to break new ground — to do things that could never be done before. I think it would be a waste to stay in the realm of what could always be done by hand. And there are many ways you can do that. It doesn’t have to be animated or interactive, though programming does make that much easier.
For me, I’ve always just connected and resonated most strongly with static 2D imagery. If I go to a museum or gallery, that’s just the kind of work that has always spoken to me. And so I feel personally pretty comfortable focusing on that for my own work. The paintings that have always spoken to me have a strong stylistic influence on my work as well.
But I know that by working through the computer, through programming, there’s no doubt in my mind that this is taking the work into a new area. And I think it’s totally okay to draw inspiration from the choices of colors, textures, forms, and structures of these painters that we really enjoy and admire, and to re-examine those ideas through the lens of algorithmic creation. I think this is a really interesting remix.
I also think that often art is most effective whenever it helps to create a bridge between the old and the new. Sometimes if work is entirely new, it doesn’t create enough of a starting point for the viewer to grasp what’s happening and to engage with it. And I think sometimes the work can engage the viewer more successfully if it uses more comfortable, relatable aspects. So I definitely don’t have a problem with using my work as a bridge from the old to the new.
JB: I’m curious as to how you think about postmodernism, which was really about acknowledging that there are all these things in the world that go beyond art for art’s sake. Indeed, there are people who would argue that art’s job in a time like this is really to grapple with all of the social, political, and existential issues that surround us right now. Would you take offense to the claim that most generative art, maybe your generative art, is art for art’s sake? How do we justify or explain what we’re making when the traditional art world has in many ways moved on?
TH: You’re absolutely right and it’s definitely a question that I think every artist should be thinking about. I think it’s totally fair for artists who so wish to use their art to address climate change or other systemic woes. I don’t think that every artist needs to concern themselves with those things, nor do I think every artist is necessarily capable of effectively concerning themselves with those things. But I do think it’s important to have a meaning and a philosophical tribe behind your work.
Here’s what I think is the value of my work and of generative art in general. You know, every day we’re spending more and more time in a digital environment. All of the tools and interfaces that we use are being constructed with code by teams of programmers. And so we’re really starting to live in this programmatic environment. And most of that software is built by corporations who have a goal of essentially extracting money from you, right? I don’t begrudge them for doing that but that’s just the facts.
It’s so important to me that we have artists working with those same materials, and that it’s not just people with commercial goals that are driving all of our software. Artists of the 20th century found this need to concern themselves with the building blocks of the day, whether it was Carl Andre using bricks or other artists using concrete or steel or glass. They found it imperative to work with the same materials that we find all around us in our daily life. And they asked a number of questions: “How can we take these same materials and bend them for artistic purposes? How can we make our environment more enjoyable and humane? How can we make it more ecologically friendly?”
And so I think code needs the exact same treatment. I think we need to develop a way to play with code and a way to create beautiful things with code, a way to create humane, expressive things with code. And the wonderful thing about code is that it’s also incredibly scalable. So I think that the benefit of doing this type of artistic exploration could have some really wonderful payouts for many digital and software environments. There’s also just so much fresh, unexplored territory. I think it’s fine to be a little bit selfish as an artist to really enjoy your medium. And I most definitely do.
In Part 2, out later this week, Tyler addresses the thorny issue of fair use in generative art. Jason also poses questions from the NFT community, probing the origins of Fidenza and which artists Tyler admires most…
Tyler Hobbs is a visual artist from Austin, Texas. His work focuses on computational aesthetics, 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. Hobbs’s most notable project, Fidenza (2021) — a series of 999 algorithmically generated works — has become one of the most sought-after fine art NFT collections of all time. His drawings, paintings, and digital works have been privately collected around the world. Hobbs has presented two solo exhibitions: “Progress” (2018) at Galería Dos Topos in León, Mexico, and “Incomplete Control” (2021) at Bright Moments, New York. His work has been auctioned at Christie’s in 2021, and at Sotheby’s in 2022. He holds a BSc in Computer Science from the University of Texas at Austin.
Jason Bailey is the creator of the art and tech blog Artnome.com and founder of GreenNFTs and ClubNFT where he serves as CEO.