In Conversation: Chris Eugene Mills

Hannah Marsland

Chris Eugene Mills’ display in the Garage, ‘a finely-tuned interference engine thwarted by a painting of an ouroboros (in thirty-six parts)’ presents a coded performance of data, that consumes and regenerates itself, as a beautiful array of digital ‘paintings’.

We sat down with the artist to discuss his experience of working in a site-specific space, and learn more about the becomings of this particular instillation.


Tell me about your experience creating an installation for the Burrard Arts Foundation Garage space?


It was a great learning experience to put on a full show, especially one that wasn’t self-organized. I’ve done a few shows that I’ve [independently] crafted together. Being on the back end of organizing exhibitions, a lot at UBC – and now the Vancouver Art Gallery – I knew what was required to make the installation run smoothly. Having said that I planned for a long time to get all my ducks in a row, and avoid hiccups.

I worked for BAF previously, from 2015 – 2017, first as an intern and then as a hired prep sporadically, which was a job that directly led into working at the Vancouver Art Gallery, thus solidifying my work experience. But I was familiar with the old space up on Broadway, I knew it like the back of my hand, but this exhibition, being the second artist to show in the garage space, I was starting from square one!

[The Garage] was really a great experience. As with any spatial economy situation in Vancouver, budgeting for emerging artists in this city is impossible. If it weren’t for the resources I had from BAF, in addition to equipment loans and help of friends it totally wouldn’t have happened.


Would you say this work is site-specific in regards to the scale and number of screens?


Entirely. The work is by definition a piece of software, the implementation is variable. My original draft proposal was for one screen; taking one of these 36 and making that the singular work, using one computer, one program. Simple.

But I think the work is enhanced by this. That buzzword ‘multitudes’ comes to mind. I was forced by constraints like natural light, the cost and unsuitability of 24/7 projection or synchronizing video walls. These constraints made me reach further and make a more ambitious work . This is definitely the largest since the big ol’ pile of lumber I exhibited for my grad, sudden collapse. If I had just one screen, one iteration of the program, I think the actual massiveness of generative possibility would’ve been lost, and wouldn’t communicate just how fleeting and random the work is. It’s funny; consciously you know the best ideas derive from limitations and detours, but they still sneak up on you.

Doing a full test of all 36 screens also revealed other site-specific conditions, such as how hot the garage space really gets, even with the windows semi-covered. I had the program testing for almost six months, running continuously, but I didn’t quite get all those screens running at once until about the week before. The reality of having them all in one room, running at once… That little room gets piping hot, I will say. It becomes a total greenhouse. I know Alex Achtem ran into this with all of her water just evaporating to dew up the glass. It was all fine of course, but that’s one example of a site-specific consideration that passed my radar.


Typically is there a very solid visualization of a piece for you, precursory to coding a program?


Definitely not. Well, partially, but it ventures so far afield that it’s effectively unrecognizable. Something I’ve been thinking recently is that coding is a material. I find myself pushing numbers like paint and rearranging probabilities like I would lay pieces of lumber together. This viewpoint kind of allows me to take artistic approaches and apply them to a very technical process. Not everything applies or carries over, but the idea of taking ten percent of your last project is one I unconsciously stick by. There are snippets of code here and there, algorithms and math that get repeated and altered from past works, and that helps define a starting point at least.

For example, I used an old 2.5-dimensional trick from retro games of stacking isometric images to make Understructures look vaguely flat and deep simultaneously. [At the time] I was most concerned with the frame of the screen and its functioning as a false window. This simple repetition, the drawing of motion without erasing the last frame upon refresh, is what makes Ouroboros accumulate endlessly. So I can start with a little tactic like that to kick off.

It’s this simple idea, “teach a computer to paint in a way that only it can,” that I ran with from the very start, not a perfect mental snapshot of the end goal but a consistent idea. I’m wary of calling it a thesis, but there we go. 

From there on, it’s just iterating, adding features and variables, letting it simmer, scheduling events, and then finding a spot where I had to stop myself. Take a breath, step back, and realize I had made something utterly boring. The code had a single purpose, one behaviour. To make this generation actually surprise me again I had to take every single hardcoded choice, TRUEs and FALSEs, every preconception about where the artwork should go, and turn those into probabilities, a chance of behaviour. Long stroke, short stroke, edge detection, complexity analysis, color vibrancy; even just brush speed. Dealing with brush speed per frame starts to make different textures and interference patterns, but you don’t want to just pick one, that’s way too monotonous, and you also don’t want it moving so randomly its not patternistic, or so fast it skips the page, etc etc. It’s becomes primarily about making a system and tuning probabilities to 80% of what you want and expect, and 20% sweet chili heat…

  • It was definitely important from the beginning to deal on a pixel-by-pixel textural level, as I always do, like Holy Generator or Architectures III, where the whole point is observing interference patterns on the threshold of the screen surface, (you can even zoom in on architectures, one of my few interactive pieces so you could open it in Chrome and get your face in it). Ouroboros wasn’t all about this, about the grid of the screen, but seeing as I was trying to achieve pure aestheticization/corruption of the source material, my personal artistic leanings had a field day!!

  • This work is notoriously hard to record, trying to post to instagram just made the video go to mush. The video streaming to the displays is just so complex, you get to this level that only can be generated, that’s actually better to be generated instead of playing a ten-minute video the size of a terabyte on loop.

[Lastly] the Windows XP glitch renaissance!


Can you expand on the relationship of the ouroboros and your visual theory of  “derivative overdrive”?


Derivative Overdrive started as a joke, about iterating off failed artworks until you can’t explain how you got there, like trying to deep dive into explaining bitcoin half-drunk at a party only to find yourself talking about alpaca wool. It was an anti-perfectionist artistic exercise for myself, about purposefully taking the road less traveled when a plotter print failed, scanning it, 3D modelling it, taking scribbles, and idealizing them into a waveform, and on and on. Like free association.

It gradually meant something very different to me.

I was grappling with the immensity of available data, (what one fantastic but defunct new-media blog called “prosthetic knowledge”), something too big to keep inside, a superstructure or hyperobject. Either in the form of technical literature or simply living by the social media feed. But being overwhelmed by information and becoming a datahoarder isn’t that novel I don’t think. For example, finding a 13-disc collection of clipart from 1999 at a thrift shop in Nanaimo and knowing it had to be preserved and backed up, like a shiny rock at the beach.


I read the story of the ampersand a while back. 

Traditionally, when reciting the alphabet in English-speaking schools, any letter that could also be used as a word in itself (“A”, “I”, and, at one point, “O”) was repeated with the Latin expression per se (“by itself”). This habit was useful in spelling where a word or syllable was repeated after spelling; e.g. “d, o, g—dog” would be clear but simply saying “a—a” would be confusing without the clarifying “per se” added. It was also common practice to add the “&” sign at the end of the alphabet as if it were the 27th letter, pronounced as the Latin et or later in English as and. As a result, the recitation of the alphabet would end in “X, Y, Z, and per se and“. This last phrase was routinely slurred to “ampersand” and the term had entered common English usage by 1837. However, in contrast to the 26 letters, the ampersand does not represent a speech sound—although other characters that were dropped from the English alphabet did, such as the Old English thorn, wynn, and eth. [extra trustworthy source]

This really outlines what I mean about the degradation of intention and attribution, and the evolution of meaning, as mis-memory smooths out the wrinkles, aestheticizes and abstracts. It’s a necessary function of human culture, for newly complex ideas to percolate, but it does feel like a continuous loss, thats never been faster than in the digital age, when data can be quantified, i.e. youtube hours of video uploaded per minute.

The loss of abstraction, the bridging from our tactile understanding to the macro and aestheticized, that’s the material I’m working with! The meat I’m biting into. In the image attached, a simple analog-to-digital converter, you can see how this very interview was recorded behind the mic, how our voices end up as a bitstream.. When I talk about derivative overdrive, it’s the difference between the two waveforms, the little area inside each step that’s lost or gained, over and over and over again…

The idea of an ouroboros came really late in the game. It was largely used in the rhythm of the title for its allusions to the program painting – not just with the images I provided it with but also its own screenshots- and it JUST won out over a casual autocoprophagia reference. Half the tuning of the work was just finding a sustainable 2nd or 3rd generation, that kept a lot of the aesthetic identity I wanted even after being mired in semi-randomness and purposefully out of my hands. I don’t control the flow of each individual stroke, but the identity of the work is defined. This recursiveness, autopoiesis and eternal unproductive work and stability reminded me of the ouroboros and its inability to even let go of its tail if it wanted to, stuck in that compulsive loop.

The rest of the title is derived from the idea of complete detachment from both source and iconography, leaving only a flat image, tricking the viewer’s brain into a state of pareidolia, struggling to keep up with the onslaught of formless content. P.S. the “inference engine” in the title is the viewer, I hope they get that. I like to think my artwork is finely-tuned, but I’ve got nothing compared to the human visual cortex. 


Were there any sources of information in particular that inspired the development of this piece, besides the engineer Paul Buran?


No defined image sources, I mostly skimmed Google images, wikipedia, getty, etc.

It was more the idea of an empty image, a technical image turned into a memetic farce of itself. Same tactic I exaggerated with Understructures, one theme or dataset, in that case floorplans, representing a tangible purpose and potent with information, completely obfuscated due to the overload of abundance.

Baran’s imagery was just the best example I could imagine regarding a technical image potent with graphical purpose, which had devolved since then, into the idea of an idea of an idea. Baran is talking about the very idea of a serialized packet of data with checks and balances and the optimization of a hard network. Without that concrete understanding we wouldn’t have anything we do today in the post-digital age. When you see that image now, its faded. Its graphical purpose is superficial and pervasive, the abstracted idea of a “network”. But I’m not being nostalgic, I wasn’t even alive then! The truth is, there is no longer a need to understand the internet on that level, just like there is no need to build a Saturn V rocket engine (wait a sec), and in most regards it’s damn good we don’t! But still there is a significant loss.

If we’re discussing general inspiration, I have a running list:

      1. The Algorists – a small movement of early generative art mostly run on massive mainframe computers in the 60s/70s after hours in university labs.
      2. Gego / Jesus Rafael Soto – Venezuelan Op Sculpture
      3. Sol Lewitt’s instructional works
      4. Cy Twombly and Cerith Wyn Evans –  use of gesture alongside critical imagery, using the two interchangeably
      5. Timothy Morton’s Hyperobjects – Ecology, the inseparability of you as a node in a network (entrapment, inescapability, subsumption), and the idea of “Grey goo” on the periphery of understanding in a network, where the complex thought of reaction and antireaction become a wash of abstraction. Where I stop talking about individual facts/ideas/events and move to a macro level to chat about “Climate Change” or “Net Art” and then zoom back in, losing details each time.
      6. Artie Vierkant’s Image Object paper and artworks – The idea of this abstraction being represented in the flat image. That a photo of a thing is inexorably tied to its source, the THING, but is a crude version of it, but still it reflects on the original.    
      7. Trevor Paglen’s Operational Images paper – addressing the aesthetics of an image with purpose, both politically and formally.
      8. Pixel Sorting and Datamoshing – i.e. the importance of glitch but not relying on the glitch phenomenon itself to be the art.
      9. Nicolas Sassoon or Casey Reas (Operation TEAPOT) – artists who concern themselves with the pixel raster surface.
      10. Evan Roth’ Red Lines series – he approaches the network physically, going to each locational where an undersea cable is laid, but he doesn’t record the cable, he just shoots the environment in this pungent infrared akin to the fiber optic frequencies. 
      11. Sondra Perry – artist who investigates race, class, abstraction, and representation, using the digital to address much larger implications, yet never leaving the materiality of that digital medium..
      12. Harm van den Dorpel’s Autobreeder series  – he used genetic programming to evolve each formal element in his compositions, relinquishing authorial control. 
      13. Ian Cheng – creates live simulations that explore the capacity of living agents to deal with change.

Are there any elements in this work or others that you want to develop further, or that you weren’t satisfied with here?


In the last 2-3 years I’ve become drawn to text art, or inclusion of text. At present, I just focus on titles, though I am working on something that will be up in September that relies much more heavily on poetics.

The tone of the titles I create isn’t always the same as the tone of the work. However, the title does determine the final work itself. I always have a drafted list of titles kept [on hand], that vary widely in between being ornate and direct. Titles are the most artistic process that I go through making a work, or at least they fit that guilty romantic vision for me. It’s a process where I feel least constrained and]can absolutely iterate creatively without any technical effort at all, versus spending months (or in this case a year-plus) coding a program. 

I used to jot in a sketchbook but recently I’ve been leaving comments in the code as I go for reference. Sometimes a codebase will be left alone for months at a time, but with the notes I can come back and think “oh, good someone knows what’s happening here!” It’s a point of reference. Sometimes they’re tongue ‘n’ cheek but most of the time they’re practical. The last thing I was [working with in a sketchbook] was doing plotter work, machine drawings as a way of iterating, and I’d like to get back into that. More plotter artwork and more random iterations. I think this would help bridge the time in between big projects, since it’s the percolation [of thought] where things really happen, not the one night you stay up until 7am working on one iteration of a piece of code that’s been stewing. 

I’d also like to follow my nose on the idea of machine agency, releasing more control to the generative process. This may take me over into making AI and adversarial networks, but who knows?

‘A finely-tuned interference engine thwarted by a painting of an ouroboros (in thirty-six parts)’  will be on show at the BAF Garage until the 27th of July.