BRENNAN MCCRACKEN: Your project, Slow Serif, uses microbial fuel cells, energy harvesting, and probability-based text generation to perform, with immense slowness, the writing of a “novella” about slowness. Where did that idea come from?
MATTHEW HALPENNY: I initially wanted to work with visualization, not necessarily with text. But text worked within my technical requirements: the fuel cells are extremely low energy, and they can take days to weeks just to generate one pulse of writing. I was exploring different hardware I could use, and e-ink is one of the lowest-energy ones. These technical factors, plus theoretical factors—everything contained this idea of slowness, which led me to use the microbial fuel cells to write text.
I also thought that text would generate curiosity even if it didn’t necessarily seem coherent. I was working with Markov chains—processes based on probability that aren’t actively learning, and that aren’t using data sets outside of what you give them. So, for this project, I gave the Markov chains three texts: Geology of Media,
1 Parikka, Jussi. Geology of Media. University of Minessota Press, 2015.Slow Violence,
2 Nixon, Rob. Slow Violence and the Environmentalism of the Poor. Harvard University Press, 2013.and A Manifesto for Slow Science.
3 Stengers, Isabelle. Another Science is Possible: A Manifesto for Slow Science. Translated by Stephen Muecke. Wiley, 2018.The microcontroller would read through these texts and create a probability bank based on which words come after other words. It gives you back all the juicy bits of the texts, but out of context and mashed up with juicy bits from other texts.
One of the ideas when I started working with fuel cells and timed release was—do you know the corpse flower?
MH: This was one of the ideas I had, where people wait patiently for this corpse flower to bloom, they watch it, they stream it, it’s very short, and then it’s over. That idea of patience, of everyone needing to take care and garden to get this result, is part of the project. The fuel cells themselves are like miniature gardens—if you want them to produce electricity, to be able to produce the novella, you need to take care of them. You need to be watering them, fertilizing them, watching them. It’s not just a technology that you can plug in and let stand by. I’m not trying to take something and isolate it, but rather taking a microcosm or a mesocosm and putting it in a new environment. For the moss gardens to work, they still need all their microorganisms, they still need the surrounding ecosystem. It’s not just one individual organism that’s producing the work, it’s a series of organisms helping each other out, and then the bacteria are producing the electricity.
BM: It’s even more knotty than I would have expected. To pick up on this idea of care and maintenance, I feel like the project short-circuits, or draws attention to, how what we might call regular publishing, or ‘fast’ publishing, is so deeply rooted in practices that are disconnected from maintenance and care. I was recently rereading Against Purity by Alexis Shotwell, which is so helpful for thinking about what Shotwell calls our compromised conditions—this idea that, whatever work we’re doing, even if it’s ethically-motivated, the ground from which we’re doing that work is so often compromised.
4 Shotwell, Alexis. Against Purity: Living Ethically in Compromised Times. University of Minnesota Press, 2018.Academic publishing in particular is something that is built upon the unequal flow of capital, that is tied to university endowments invested in extractive fossil fuel production, all these factors that make possible its ‘fast’ publishing. Here, in your project, the scale is different—both temporally and in terms of the work that’s being produced—and the logic is different too.
To return to the text itself, you mentioned the three texts that you used as material for the Markov chain to draw on. Why three non-fiction texts, when the work that’s produced is framed or indexed as a novella?
MH: I did have a bank of texts to choose from, but since I wasn’t using AI per se, I had to hold its hand a lot of the way. In practice, that meant copy-pasting an entire pdf into a text document, going through it, getting through all the numbers, getting rid of the titles that reappear—otherwise, it would just keep regenerating titles, because they came up so frequently. I probably would have added more than three texts, but that process was quite lengthy.
I see these three texts as relating to the materiality of the creation process. If I went back, I could have picked different texts—I’m not necessarily super committed to those three, but I think they were at the top of my list at that time. The memory on the microcontroller is also super low, so that dictated how many texts I could load at a time. It wasn’t a lot.
BM: It’s significant, I think, that the volume of texts you’ve drawn on is directly related to the material and technological limits of the process, that text that’s been produced so far is emerging from such a limited set.
MH: That limited set produced interesting effects…it was a little bit more chaotic. As it reads more and more, the output makes more and more sense, but when you have fewer texts to draw on, there’s a bit of juggling going on that makes for cool results.
Here is the full text:
BM: How long did it take to produce these three paragraphs?
MH: For Bohn,
5 Digital Ecologies conference, University of Bohn, 2022.because of the time frame, this is a representation… You can disconnect the work from microbial fuel cells, connect it to a battery and let it go faster. So that’s what I did for Bohn. This would take months for the fuel cells to produce.
BM: And the paragraphing—is that your choice?
MH: I’m not sure when it decides to paragraph, but I do know you could add bits of code to have it break every three, four, five sentences… But I believe the paragraphs were just its doing in this case.
BM: Same for the punctuation?
MH: Same for the punctuation.
BM: So sick. “Indeed, aesthetics becomes twisted into a soap bubble.”
MH: Yeah, that’s one of my favorite lines.
BM: There’s also a way in which the trippy logic of the sentences slows down my reading, too. When I approach the text, I feel like I keep snagging on these corners, these syntaxes or grammars that aren’t as frictionless as the kind of texts that I’m so habituated or trained to read.
MH: That experience is such a big part of the work. When I presented it at Bohn, people were thrilled to look at nothing. They were staring at it, maybe hoping that something would pop up, but at the same time you could tell people were thinking, digesting, having their moments with the work. Everyone was surprisingly quiet—it was a beautiful thing to see, and one of the responses I wanted to get from the work. That’s huge for me when I make art—I’m not trying to paint a picture of anything per se. Instead, I often work with systems-based work, trying to create experiences that are about process more so than static objects. This has holds in both worlds: you do have the static object of the text, but more than anything it’s the process on display.
BM: It reminds me of a strange inversion of the Romantic relationship to nature and slowness. Part of the quality of Romantic poetry, and specifically the experience of reading it, as I understand it, was generating an affective slowness in response to the acceleration of the world through industrialisation, the development of railway networks, the speed of information.
6 See The Calamity Form: On Poetry and Social Life by Anahid Nersessian and “Slow Time” by Jonathan Sachs.In this project, the way that the slowness is being performed—and also mediated through technologies—situates slowness not in some “pure,” discreet “nature” that is separate from the human world, separate from our fast creations, but as something that can be found in ecologies of human and non-human relations. It’s such a fascinating response to these pervasive, durable Romantic ideas.
MH: And responding to that concept of nature—as someone who comes from a background in biology, I really was hoping to explore how we’re interwoven with these natural processes. All the technologies we use rely on the environment to for their production, often in very destructive ways. Bringing our understanding of ecology together with technology, and not removing ourselves from natural processes, is important. In this project, we’re needed, we’re symbiotic.