INTERVIEW: CANEK ZAPATA
Canek Zapata is an editor, curator, and internet artist based in Mexico City
specializing in digital culture, generative systems, glitch art, and memes. His work focuses on creating computational environments that allow the computer to express the unthinkable through audiovisual and written media.
As part of his collaboration with TITLES, Zapata developed Modern Times, a generative model trained on a series of figures inspired by the tradition of miniature painting, but reimagined inside of retrofuturistic technological environments. Each figure embodies a technical-poetic role—archivist, synthesist, vector oracle, system mediator—operating hybrid artifacts that blend technology with ritual.
We spoke with Canek Zapata about prompting as spellcasting, the generative potential of noise, and giving computers the freedom to make creative decisions.
What interests you most about creating art with a generative medium?
What interests me most is what the computer can believe, beyond what I explicitly ask of it. My model with TITLES is the product of that kind of prompting. I’m not telling it, make me a computer and a plant, but rather I’m saying, we need there to be—and I list different models of old computers and different species of plants—and leave the decision to the computer, letting it express what it thinks should be there.
The same happens with cultural references, which is why there is such a mixture of them. I call this the computer’s bot-ness, that quality the machine can express that isn’t exactly what a human would. And for that to emerge, the computer needs a certain kind of environment. As an artist that’s my task: to give the computer the tools and attention it needs so it can express itself in posthuman ways that I don’t expect.
It all comes down to playing a game where the computer gets to surprise you.
Many of your pieces take the form of GIFs, which feel distinctly internet-native. How does the environment where people encounter your work influence what you create?
Many of these works are built with the infrastructure of the contemporary internet in mind. For example, you can’t post a collage of GIFs on social media; you have to turn it into a video. So a lot of my art is designed to sustain itself within the parameters imposed by the platform on which it will live.
But a large part of my work is also about infrastructure itself. I make a lot of net art that searches for strange behaviors inside HTML and its adjacent systems, and that kind of work looks better in its native environment—a webpage—than as a video on social media.
How are the responsibilities of a curator different in digital vs. physical space?
I think there are many differences. Digital work is seen at home or on a phone, but in a physical space the audience looks for something very different. I think it’s much more complicated to present digital art in a physical space, because you can’t demand the same kind of attention from a physical audience as you can from people on social media or on a webpage. Digital work often has a different logic of formation, where the audience actively engages with the objects by commenting, sharing, and so on, which generates dynamics that become difficult to replicate in a physical venue.
For that reason, showing digital art in a physical environment often involves additional layers of mediation—for example, letting the audience participate in the piece, which we understand as interactivity; or allowing the audience to step into it, which becomes immersion.
Do you see coding as an extension of your literary practice? Are they separate crafts or variations of the same compositional impulse?
For me they are similar. Both are forms of writing, even though they are completely different languages. One is writing for humans and the other is writing for machines.
Writing code is thinking through the possibilities the computer has for carrying out actions, and it’s often surprising when it does more than what you initially imagined. I enjoy reading what the computer writes, even if it seems strange, senseless, or illogical.
What aspects of the digital world do you find most generative for your artistic explorations?
Noise. Synthetic generators like LLMs or other systems that produce cultural objects often work mainly from what we ask of them. But for my creative practice, noise is very useful, because it allows the computer to make those decisions that I enjoy seeing it take.
What can a glitch reveal that a clean image cannot?
I think glitch—and here we are a bit fundamentalist about glitch—is an error inside a file, and the computer trying to interpret or push through that error. In natural glitches (the ones that exist in the code itself and aren’t produced by direct human intervention, but by the error) we can see the computer’s interpretation of some form of uncertainty, noise, and chaos—and maybe even a bit of the future—because the machine is interpreting something it does not fully understand.
In your model for TITLES, you draw inspiration from the tradition of miniature painting, as well as the retrofuturistic style of early computer graphics style. What interests you about bringing these two aesthetics together?
One thing I like about this generative era is that it represents the triumph of language, especially of fiction. So I’m interested in using these models to generate things that are not strictly human, to see how the computer understands this cultural inheritance of ours and how it interprets it outside of our perceptions and judgments. I think the game of miniatures with LLMs has always been about exploring these cultural border zones, where cultures mix and produce hybrid solutions.
I often think of the Greeks making classical Greek-style statues in the Hindu Kush, which ended up influencing the earliest sculptural representations of Buddhism in those mountains, and similar situations of cultural exchange.
How is a technological object also a ritual object?
I think that any object, regardless of its materiality, can become a ritual object. And I feel that making art with these synthetic generators of cultural objects is similar to the oracles of the past, where you offer an idea or a question and wait for a result. And many times you imagine a certain outcome and it isn’t what the computer gives you. Sometimes it gives you more than you expected.
I don’t know—it’s interesting, because I do think the process becomes a kind of ritual. And personally, since I’m always looking for noise, I sometimes even write magical spells into my prompts.
In the worlds you construct, do you imagine humans interpreting machines, or machines beginning to interpret us?
Humans interpreting machines. It’s interesting, because when we talk very specifically about synthetic text generators like ChatGPT, we are letting into our homes a machine that can speak the natural language of humans but is not human. Ten years ago, if someone had asked me about this idea of allowing something that can talk to us into our home, I would have imagined an extraterrestrial pet or something like that. But here we are, with a machine trained on our dreams and fears, our desires and doubts (well, the ones you can find on the internet), built from the human—not to interpret us but to infer itself into our lives through its interpretation of human data.
It’s an interesting question, because one could say these machines are meant to interpret what is human. But what interests me is interpreting the machines. I think that, as an artist working in this kind of practice, one becomes more of a curator of what the machines can generate. And I’m selecting the fruits that feel most machine-like to me.
Try Canek Zapata’s model on TITLES.