Sound in the Shape of Electricity

Kirk Pearson is a composer and the creative director of Dogbotic, a Bay-Area-based audio laboratory that, as Kirk describes it, “makes weird and strange sounds for weird and strange people.” I first came across Dogbotic while finishing my Masters in Cybernetics in 2021. I signed up for a class on tape loops and hacking cassette recorders, eager to get more hands-on with creative analog technology. It was among the most fun things I have ever done — and no, they aren’t paying me to say that.

Kirk makes a lot of music out of found sounds and invented instruments, and leads many of Dogbotic’s workshops.

I was eager to have a discussion with Kirk because of the way Dogbotic teaches technology. It’s unlike anything I’d encountered in education spaces, in ways that spoke to my artist-brain to make abstract concepts in electronics more tangible. It’s oriented to the idea that technology is something you can make, break apart, and rebuild to do new, personally expressive things. In an age of out-of-the-box prompt-driven creativity, it’s worth thinking back to analog tools and their models of creative misuse. 

I’m genuinely enthusiastic about Dogbotic and have written about them more than once — on the history of tapes and data, and in the development of my mushroom synthesizer. 

I wholeheartedly recommend any of their classes to curious tinkerers, go check it out before you forget. This is not an ad. It’s genuine joy at the idea that you, yes you, will take a class.

Oscillators & Elevators

Eryk: Why don't we start with what Dogbotic is? What got you started?

Kirk: Dogbotic is about five years old now. I'm a composer that does a lot of work in pretty disparate fields and styles, and quickly learned doing that under your own name wasn’t kosher. The first few years of Dogbotic were a glorified one-person shop — I spent 2018-2020 working pretty much solo as a music and sound design studio. My clients are predominantly animators and filmmakers who have a sense of humor, and they tend to be looking for music and sound that other studios might not take too seriously. Or it’s some sort of museum or design firm and they're building an installation that requires an interface to make such-and-such a sound. Basically, it’s my art practice spun into the world of a commercial agency. Just, you know, we make audio that has dubious commercial potential.

In 2020 the pandemic hit, and suddenly the custom audio industry seemed to be the first to go. And so, kind of out of boredom, I started teaching these DIY synthesizer workshops like a summer camp — I’d ship you the parts and teach people together, a little bit every week. Fortunately — and shockingly — the workshops spun out into their own thing. And now the educational aspect is a whole entity adjacent to the sound house stuff.

Eryk: Was it always online? 

Kirk: It's always been online. We probably couldn't fill the workshops with people just from the local area and keep them as niche as they are, which is part of the appeal. We're trying to teach something that seems, on the surface, a little bit frivolous, and I like that. We’re teaching analog electronics to musicians so they can build their own effects units, or teaching ways of chemically processing film that are different from a traditional darkroom photography class. Very few of the things we teach are “practical job skills,” but I think that's great. At the end of the day, there are so many songs produced with the exact same gear — but I’m most interested in hearing the music produced with something weird the songwriter built in their spare time. And I don’t think that’s just me!

Eryk: Even though it’s super niche, you're able to find this community of people connected, it seems, by ethos. You have this very particular point of view. 

Kirk: Yeah, our fundamental viewpoint is that music technology is ultimately a creative practice. The one thing that defines music technology is people from different disciplines meeting each other and coming up with something new and interesting. Almost all of the coolest projects students have made came out of a conversation between two people from different worlds. It’s really the best way to make art.

A thrill that people get out of the workshops is just realizing that, oh, there’s already a community of people that do this work! And it's validating for folks to meet others that have the same eccentric proclivities and tastes as you, especially when so much of the art one makes could be dismissed as “unappealing” or “inaccessible” by those you might hold near and dear.

“All art practice is political.” 

Eryk: There's also something a little subversive. I don't know if political is the word that you would use?

Kirk:  Yeah, I would. All art practice is political. It’s so easy to look at an artwork and call it frivolous, or not a worthwhile pursuit. By making a work of art that’s polarizing, you’re actively creating a political statement merely by calling your creation “art.” A lot of art challenges the conception of what art is in the first place, but more importantly to me, it challenges the conception of who can be an artist. 

On a more technical level, when it comes to building or even operating synthesizers, the learning curve has always been steep, and can be unfriendly and vicious. But there's nothing that separates musicians from the engineer types, or vice versa. It’s important for us to remind people that they actually have the ability to learn new tricks, and that engineering isn't magic. I also just get sheer enjoyment out of watching people make art that they didn't realize they had the ability to make. I think that's fun.

Eryk: When did you start building this stuff?

Kirk: I've always been a tinkerer circuit-bender type. But truthfully, I didn't start building my own synthesizers in any serious way until a few months before the pandemic started. Coincidentally, I have a science education background, so it was pretty fun for me to parlay those things into the weird and wonderful Dogbotic Power Hour.

It was a personal frustration of ours that a lot of Internet resources for learning electronics are not only dense and poorly written, but actively antagonistic. In addition to the fact that a lot of instructional materials are needlessly confusing, or, even worse, just flat-out incorrect — when you complete a project successfully there’s just not a community that me or any of the people I work with have found to be very supportive. And this is coming from me (gestures at face). It goes without saying that these technical communities are even more antagonistic and condescending towards women.

So the idea with our workshops was to provide at least some sort of antidote to that. It’s easier to learn electronics in a group, and it’s more fun to base everything around art-making than technical wizardry. We like reminding people that at the end of the day they're doing this to make music. Fewer and fewer places are actually teaching this stuff as a creative art, which is emphatically the best way to teach it.

Eryk: I've taken classes in electronics, engineering, and nothing clicked into place as “real” until the tape loop synthesizer class. Both of your classes I took, the process was explained in a way that made sense to me as a creative person. I’ve never had this orientation to engineering, which kind of assumes there is a finished product to build, as opposed to someone who actually wants to make a new thing. Engineering as exploratory process just isn’t the done thing.

And I have also struggled with my technical education — I have Dyscalculia, which sounds like I’m a vampire, but it’s numeric dyslexia — so things like coding and even remembering dates is extremely challenging to me. The rigidity and inflexibility of programming is very unkind to my brain. It has always made me lean a little toward glitch — if I screw up and make something that pops when it should click, then how can I work with it? A fun part of the Dogbotic classes is that approach. 

In electrical engineering classwork I just got black smoke, and finally here I was in a situation where that was framed as doing something ... not exactly right, but as a natural risk of what you should be doing, which is experimenting. There's a way of failing that is still okay if it makes good noise.

Kirk: Failing while making art is the best way to learn! It’s also hard to “fail” at building a synth (unless, I guess, it just doesn’t make any sound). The difference between synth building and engineering elevator brakes is, your elevator braking system really has to work every time. But in the case of musical instruments, all you need is to get something that's culturally interesting. I don’t care if it doesn’t work the same way every time — the fact is, you made a sound that nobody else in the world has made and I think you should feel some pride about that.

DIY Music Technology

Eryk: I had the sense that your workshops are aligned with a feminist approach to technology. It's centering the communities we build with and through the machines. It's not just about getting them to do what they are supposed to do. It's about building technology that you can share with other people, whether it's the music or knowledge. I wonder if that is intentional, or just something that emerges from you because of the way you think about things or do things?

Kirk: Wow, I've never heard that comparison before, but I like it! The maker movement… as much as I love making stuff… is not exactly a very feminist movement. One of the ways that shows is a lot of attention on product and not on process, and pretty much no attention on critical thinking about it. We're coming out of a decade in which a lot of rich people recklessly made a lot of stuff without critically thinking about how it could or should be used. That's an important thing — we can help teach that as an educational resource and that can make the world better. Teaching people to slow down and think about how the process works and why you're building the thing you're building. You’ve gotta have a lot of privilege to have a “move fast and break things” mentality.

Eryk: Does that ethos fuel back into the DIY and analog elements?

Kirk: I like the DIY ethic because it's easier to convince yourself that the world isn't just a black box. I love working with computers, don't get me wrong. But at the end of the day, if you're programming an Arduino, there are a zillion levels of abstraction that you don't get to see. And that's fine, you don't always have to see them. But knowing that they're there is just enough to change how you approach it and change your creative process. 

Building your own instruments pushes against this idea that synthesizers are products (synths are things you buy, with a price tag, and are made by a company). When it breaks, you get the new synthesizer. You can't fix it! Commercial circuit boards are designed to be deliberately inaccessible and confusing so they can’t be fixed, or not be modified, and that has real social effects. It results in the creation of a ton of waste. It results in an economic system that kind of encourages you to make products that are disposable rather than salvageable. It also pushes forth the idea that these electronics are somehow “more complicated than you can understand.” But that’s absolute nonsense. It’s the same as learning to fix anything.

But I also think the DIY mindset is important because it's nice to remind yourself that music technology is a stupidly vague term. Just by futzing around with this stuff and making sound, you are participating in music technology as a discipline, which certainly to me, growing up, when you said “music technology,” the image in my mind was an academic or something super corporate: places that had financial means to be doing whatever “the latest technology” is. It took me a long time to realize this phrase, “the latest technology,” doesn't make any sense. It's not really a thing. You could participate in that at home, just by making a sound that no one's ever heard before. 

“Music technology is about appropriating a thing
that wasn't previously considered a musical instrument.”

Eryk: Yeah. It is funny because that idea of the most advanced technology, a lot of the sort of early synthesizer stuff, like when you look at Stockhausen or Delia Derbyshire, the tools they're using are now dirt cheap and they were just playing with it, and it was an attitude that they had, actually — OK, Stockhausen stuff might still be expensive. He was playing with, like, a radio station. But they treated it like it wasn’t precious. 

Kirk: Right, right! It's all about the mentality you have… that’s the difference, more so than financial resources. Music technology is about appropriating a thing that wasn't previously considered a musical instrument. 

The DIY workshops we do are not based on any particular end goal–we don’t end with a single project everyone makes together. We give you a bunch of little sketches of ideas and encourage you to try putting them together in your own way and asking us questions if something isn’t working. Not having a single goal can be an incredible frustration for some people, but I think it's an advantage! People have different interests, are turned on to different kinds of sound, and I want them to build something that will solve a problem they are having in their creative practice. There are countless projects on Instructables that you can go ahead and do, and I don’t think it’s the role of an art class to focus on that. It's much cooler to say “here are 50 colors, go ahead and paint with them.” 

The projects that students make always blow me away. If it was all some predescribed circuit, I’m sure people would make it just fine, but everyone would end up with the same thing! That’s hardly music technology, that’s just “you do as I do.” 

Eryk: Which reminds me of the Black Box vs DIY. You don't know what your digital audio software is doing behind the interface unless you get into the code and look at it, so you kind of have to do music the way it wants you to do music. Whereas you can rearrange circuits and have them flow around each other in any way. 

What’s the Deal with Electricity?

Eryk: There is also a weird sense of magic in electricity that seems so interesting to unpack. The question I have written down is, “what's the deal with electricity?”

Kirk: Could you say that like Jerry Seinfeld?

Eryk: What is the DEAL with electricity?

Kirk: Perfect! I’m so glad you mentioned this. There's something alive about dealing with electric current that you just can’t get in a script editor. There’s an early exercise we do — the first oscillator exercise where you plug your amplifier into one of the pins on the NAND Gate, and you get a square wave output. Then you plug it into another place and you get the same pitch, but a triangle wave. All you're doing there is tapping into a different part of the circuit where the electron density is changing in a different pattern. That's all that's happening. But it feels like you're a doctor with an endoscope. Working with electricity is almost voyeuristic in this regard.

I like bringing up this example because it’s a simple project where that idea kind of crystallizes in a lot of people's minds — a synthesizer is a gigantic, crazy electron-pushing machine that you can tap in different places and you will get different signals from those different places. Realizing that you've built this crazy little Rube Goldberg contraption that's doing that is so much cooler than doing it on a computer. Yes, yes, an Arduino might work better and be more reliable, but you’re not really pushing and pulling electrons at that crude, simple level—you’re telling a computer to push those electrons for you. Same output, but you really learn a lot less.

Eryk: You also get the sense of electricity as shape, sound as the shape of electricity, which is a beautiful thing to grok once you grok it.

Kirk: I totally agree. I'm also astonished at how poorly understood that is. The general public knows what a waveform looks like. They've been on SoundCloud before. But so few people, including me, many years into my musical journey, had no idea that that was literally a drawing of what your speaker is doing. That's such a simple thing. The shape is sound.

Tape Loops are Algorithms

Eryk: I wanted to talk about cassettes and tape, physical media. Currently, we are in a bit of a generative AI, digital fever dream. We're thinking a lot about generativity as if it's new. We've lost track of the fact that this goes back eons, goes back to the 1940s, where people were taking pictures of oscilloscopes. This came up around tapes and tape loops, and you mentioned that tape loops could be generative too. 

Kirk: Generativity is like the creation of a recipe. It's the creation of a process that you're doing. When you create generative art in one form or another, the output and the input are almost arbitrary, so the object of marvel becomes the process itself. Truthfully, I think that's the marvel of DALL-E 2 or whatever — it’s the software, not any of the images. I'm not wowed by the artistic merit of any of the images it’s produced (they have none), but the true marvel is this infrastructure that's built to translate between natural language and image. It’s the marvel with tape loops, too, or making a circuit that slowly evolves a melodic pattern over time. The input is arbitrary, but the format of the output is what makes it novel.

I like making generative art because it feels like you generate meaning out of arbitrary inputs. When you make a tape loop, or generative circuit, or whatever, you've somehow created this weird little robot that seems like it can understand culture in such a way. Of course it doesn't really understand culture. It's an unfeeling thing, but you're able to glean something with meaning out of it. A lot of generative art is like staring at clouds, and I mean that in the best possible way. You stare at clouds, and eventually you see a pattern emerge, and you're like, oh, that means something to me.

Eryk: I think we talked about Brian Eno’s Music for Airports. The way it was built, which I did not know — I knew the piece, but I didn't know how it was constructed.

Kirk: A classic. As far as I know, it’s just Eno playing with tape loops of various lengths. So you have a gigantic loop of tape, and it reads through the tape and then reads through it again. But of course, because all your tape loops are different lengths, you never end up with the exact same pattern of overlapping stuff at the same time. That's all I know about Music for Airports. Side one, disk one is the banger. The rest are OK. 

“The tape loops don't have a sense of when it is going to be a powerful expression,
but we do, as listeners.”

Eryk: The funny thing about Music for Airports is that the result of that generative music is the same as the result of these AI tools. There's this musical score that people find very beautiful — maybe they find some sides of the album more beautiful than others — but it's a few notes put onto different pieces of tape of varying lengths. They overlap with one another almost arbitrarily. Someone puts together a recipe, as you say, and then finds beauty in it.

Kirk: Right. The tape loops don't have a sense of when it is going to be a powerful expression, but we do as listeners, and Brian Eno does as an editor, too. Audiences don’t give themselves enough credit. They’re the ones doing the heavy lifting right at the end.

Eryk: That's one of the interesting things about making music, about making generative music, is where does Brian Eno come in and where does the tape loop come in? A few years ago someone wrote a kind of so-angry-it-was-funny response to my suggestion that Music for Airports was an algorithm. But it is, because you have a range of possible arrangements that can result from various interactions, and they’re quite large! So this comparison to AI generated art, well, it’s just scale. But that leaves the question, too, about recipes. It’s easier to lose sight of this when you have 5 billion different lengths of tape instead of just the seven. Where do composers come in then? 

Kirk: I agree with you it is an algorithm, and algorithms can certainly be art. But I do have a hard time parsing between if the finished work is the art object or if the recipe of creating the work is. The more I work in the musical world — and I'm sure you identify with this — the less of a definition I have for what music is. You just continue to find counterexamples of things that previously didn’t fit into your little mental box. That's an exciting thing, to be confused about what your own discipline is! Better still, it’s exciting to not be concerned with whether something is or isn't music, but to be concerned with whether or not you think it's interesting.

With Music for Airports, I think Brian Eno found a really good recipe that resulted in an ok-tasting dish. But it’s not a piece that could have been written in the manner I was taught to compose, so it was certainly eye opening. The rather conservative approach I had growing up for what music was, was that you needed to have your words pre-formed. You needed to know what your melody was before you wrote it down and embellished it. But something as straightforward as taking a couple of different tape loops and playing them all at different speeds already complicates that. Whether or not you get something that's worth sharing, that's up to you. As a process, it's very liberating just to realize that there's more than one way to compose.

Gaussian Noise

Kirk: Also, I want to talk to you about your Gaussian noise AI project. Can you tell me a bit about that? 

Eryk: Oh, sure! I've been playing around with AI, and I am really interested in generative art and where it's coming from. I look at Fluxus stuff from the 1960s, where generative art started, or at least, got its vocabulary, though lots of it goes back further. Point is, this didn't start with GANs or in 2014.

So Gaussian noise, every image you make with Dall-E 2 or Stable Diffusion, etc., starts with static. Diffusion is basically trying to clean that image up. The machine is designed to be a noise reduction machine for images of pure noise, which is a really cool idea. But the only information it has for reducing that noise is whatever your prompt is. So you type in the prompt, and it understands, from looking at billions of images and their captions, how to reverse noise to that prompt or to certain pieces of that prompt. And then it gives you something totally new, with rare exceptions.

So the Gaussian noise thing that I've been playing with came from a question I had. If you start with a piece of Gaussian noise, and then it’s removing Gaussian noise in a direction toward some other image, what if the image that it is removing-noise-toward is more noise? Where does it go? It's like ping ponging now. It's reducing noise, and then it adds noise, and then it reduces that noise — or some combination of the new and old noise — and then it adds noise, and it keeps pushing and pulling, and it makes these weird, recursive fractal things. Or it did.

A lot of the models were updated and don't let you do that anymore. They turned it off, or restricted its behavior. The new Stable Diffusion just gives me black and white photos of some guy. 

But Stable Diffusion 2.1 seemed like it just didn't know what the hell to do, so it gave you these wild textures and shapes. It looked like this early generative video work, like the Vasulkas or Nam June Paik or whatever. It wasn’t really connecting with any of the style information it had learned from the Web. It was unmoored and cool, and they're getting rid of it and making it just generate a picture of a random dude.

Kirk: It sounds like a Greek myth. You posed in the mirror of chaos! That's such an interesting idea, asking an engine to create something random when it can't create anything random.

Eryk: Yeah, and that's what I'm playing with, is glitch as generativity, recursive feedback where the error creates something new from the noise. It's like putting two mirrors next to each other, but putting two mirrors next to each other and just some dude shows up.

Kirk: Yeah, I like that analogy. I feel like at some point you're going to come up with some very profound statement about how humans think about something based on how the AI breeds noise in a particular image. I guess my question to you is, do you think that the most successful pieces of AI art in the near future are not going to be works of art made by an AI, but they're going to be pieces that deliberately misuse one? Kind of like you’re doing.

Eryk: Yeah, I hope so. I hope it is! This brings us back to DIY and creative misuse. Creative misuse is so important to building new vocabularies in art forms around new technology. When people invented film cameras, imagine what it took for someone to say, “OK, I'm going to take that piece of film and I'm going to cut it with scissors and I'm just going to make the thing that happened afterward happen first.” That's kind of mind blowing that someone even did that. And with digital stuff, what are we going to discover? What is the language that we are going to discover with it? We won’t learn it by reading the instruction manual and doing the thing that the designers tell us to do with it. There is something exciting about throwing the instruction manual away and breaking things.

Kirk: When you watch really early films, I'm talking like 1890s early films, like pre Edwin S. Porter and all of that fun editing stuff, they really look like stage plays. They look like someone put a camera in the middle of a proscenium theater and walked away, which is not a value judgment. They just didn’t yet have the vocabulary of film, to move the camera or to edit. You had to wait a good 30 years or so before that stuff started to happen in cinema. 

I felt that way about VR for the longest time. Virtual Reality is kind of falling by the wayside. It was going to be this “next big revolution in storytelling” ten years ago, and I have not yet seen that. We’re still using VR like it’s a proscenium theater. What we do in VR is, we take you and we put you next to the dinosaur or in Bjork's mouth or whatever, and that's as risky as it gets. We don't yet have the Sergey Eisenstein of VR, who is going to say, wait, what if this doesn't have to represent someone in a first person view? We certainly aren’t there with AI generated art. We haven't even scratched the surface yet.

Creative Misuse

Eryk: This brings it a bit full circle. The DIY accessibility of analog made creative misuse par for the course. A synthesizer is creative misuse by default! It's a creative misuse of electricity and circuits. We call this creative misuse a synthesizer. But that's not what electricity was, quote unquote, for. It's not what breadboards are for. Do you know what I mean? Whereas with the new stuff that we're seeing, this new breed of AI, especially, there's interfaces that are designed that give you very limited control over what you can actually tweak. And so the Gaussian noise thing for me is exciting because you could do it through the prompt window, but there's got to be so much more that you could do if you could get behind the interface.

When we talk about digital, we forget the hardware and products — the materiality. You have all that, but that’s not the interface anymore. You can't do much with the materiality of a circuit board you can barely see, nevermind touch! My phone can access all kinds of cool things and toys, but I can't change them. I can play with the toys but I can’t play with the factory that makes the toy. I can’t take the apps apart to make new ones. 

Kirk: Yeah, I don’t see most people getting under the hood of any of these AI tools and I feel like that makes them even more intimidating. We know they're going to replace people's jobs, and we don't even know how they work! I wonder if they become declawed a bit when you see how they work. Once you spend enough time on ChatGPT, by which I mean 30 minutes, and you realize, oh, yeah, it can really only structure an argument in a couple of ways.

I think an important caveat to what you bring up is that building a synth — with the means that we teach — is an accessible, genre-breaking activity today. It was less so in the 1960s when these parts were a lot more expensive. A lot of what makes this activity fun is the fact that they are kind of obsolete and they're not intimidating.

A computer nowadays is greeted as a magical machine. You hand someone an Arduino and you say, this is a computer, and they'll go, Whoa. They won't realize that, yeah, it's a bunch of little switches. But hand someone a resistor and you say, yeah, this is just a crappy conductor. They're like, okay, I got you. The Dogbotic identity is 50 years past its prime in some sense.

A Misusing AI workshop is an interesting idea, but it feels like, in some ways, it's scratching a different itch than playing around with clunky through-hole components. But they are similar in the regard of pushing things to the limit and seeing what creative expression comes out of it. 

Eryk: The economics of it also touches on the politics. Right now, I think these AI tools are artificially inexpensive, the same way that Lyft rides were artificially inexpensive, because they're trying to hype it up. They're burning through a lot of money to make these things accessible. It costs millions of dollars to train DALLE-2. This costs a lot of money to build, so you and I can't build them. We can build other things — I teach a class on GANs, which we can go in and do on our own — but I’m talking about these billion-parameter models. We can use them, and pretty cheaply. But I can’t go out and train a model on my own 5 billion images, for lots of reasons. They're expensive, energy intensive, all this stuff, so the guard rails are high, and customization is rough.

But what we can do is almost, like, pirate it? We can go in and we do creative misuse or play with it in ways that are unexpected before they patch it up. There's an element of taking it back a little bit — this is internet data, the human species made it, etc — and I find it fun and a little subversive to try and break things. But ultimately, you're pretty restricted in what you can do.

Kirk: I really agree with that. I forget who said this about advertising… Vonnegut maybe… but it was (paraphrases) “feel free to steal whatever advertising you want and do whatever you want with it… You didn't ask to be fed advertising, but you are, so screw them. You've earned it.” I totally feel that way about AI. This is a very big, fascinating machine that's been built. And the sad part is we know that the people in power are not going to look at AI the way you and I do as artists. No conglomerate is saying, like, “oh, great, now we can have coders pair program with this AI,” they’re saying, “now we can lay off 2/3rds of our workforce to satisfy the investors.” We can’t pretend like the enthusiasm is anything except that.

“It's tricky playing with a new medium because you don't know if what you're doing
is properly critical of it, or if it comes off as a ringing endorsement.”

It's tricky playing with a new medium because you don't know if what you're doing is properly critical of it, or if it comes off as a ringing endorsement. A studio project that Dogbotic did a few years ago was the first ever Alexa Audio original, where it was supposed to be a “Choose Your Own Adventure” style video game, but for your Alexa. So there was no screen. You were just having conversations in a natural language, with characters, and then the story would change, and then you would go down a different narrative path.

I loved the idea that it was a bunch of kids hanging out in a room together, not looking at a screen, and interacting with a story. Writing music for it was really interesting because you had to write things with all sorts of weird patterns that could recombine. 

But we were really hung up on accepting the job for the first few days, not even because it's Amazon. It was because, by working on this project, you're kind of giving a tacit endorsement to the idea that there'll be a hot mic in your kitchen. By getting impressionable children to interact with this thing without a screen, you're introducing that this is going to be the future — “you don’t know who might be listening, kids!” 

I'm sure the people in the early era of television weren't thinking, oh, this is going to be a medium where sycophantic demagogues can lie about their job history and then be elected to Congress. They were probably just doing their jobs. But that's a tricky thing that keeps me up at night as a new media arts person. We have to be a step ahead of the rest of everyone else.

Eryk: I get that 100%, because every time I create these images of Gaussian noise, I'm giving stable diffusion some money because that's just how it works. 

Kirk: Are your ethical concerns about AI more or less your ethical concerns about capitalism? 

Eryk: Yeah, I feel like that typically boils down to that.

Kirk: And even if you are subverting the AI, you are, in a way, making ads for it because the tech is still so novel. It’s like talking about early TV. Were those people just using a new medium to see the new limit of human creativity? Or by doing that, were they also normalizing the idea that there will be ads piped to you through a box in your living room?

Eryk: There are ways of using tools that are sort of oriented towards control or power, and there's ways of using tools that aren't. Nam June Paik did this “Good Morning, Mr. Orwell” satellite broadcast in 1984. Three live satellite feeds from New York, Korea and Paris and he just, like, got drunk with John Cage and George Plimpton (and Laurie Anderson and the Thompson Twins) and mixed — superimposed, even — the feeds while millions of people watched it. 

It's interesting when these technologies are used in ways that ask, “but what else could it be?” What could it be if it weren't this corporate tool? In the same sense that public broadcasting asked that question, or could be asking that question? Paik’s thing was on WNET, which is a public broadcasting company in Buffalo. So there is that. It’s hard to call that an advertisement for television, and more like a proposal for how we could use television beyond control and power. That was stuff that Paik was specifically asking. 

Kirk: It’s pretty sad that what captures the public imagination is using AI to make art, and not using it to… oh… fix traffic problems, facilitate complicated surgery, create more accurate weather maps, and so forth.

I believe Frank Oz said, about Sesame Street when it was being developed, that the goal was to take lessons learned in the TV advertising world and apply those techniques to literacy and so forth. Let’s get AI to help with stuff like that. But the people who own AI don’t have your best interests at heart. They’re not concerned with literacy, they’re interested in maximizing profit. Asking envelope-pushing questions about the ethics of media is not a good way to generate profit.

Eryk: We have a cool public radio station in New York that is an experimental radio station and I'm really desperate to work with them in some way.

Kirk: There's something community based about “we have an experimental television station” that doesn’t have the same ring when you say “I have an experimental YouTube channel.” Making art with a community of people — which you inherently have to at a radio station–is my preferred method of creative work. You meet people! You learn new things! You have your perspectives challenged! Ironically, it’s much harder to do that on TikTok. We’re three years into COVID-19, but I’m so optimistic by the in-person art making that I’ve been seeing over the last few months. It seems like after a protracted era of sheltered-in-place individualistic art, we’re finally going to start butting heads again. And I’m so excited for that.