Waiting To Be Signed · interviews on generative art, on-chain
← Index
Interview // JUL 2024

James Merrill

Title: Making Generative Art Less Random
Role: Generative artist
Platform: Art Blocks
Duration: 1h 7m
Hosts: Will & Trinity
Listen on Spotify Guest on X Download MP3
#068 · Making Generative Art Less Random
Self-hosted audio // press play
1h 7m
MP3 ↓

Spot an error? Highlight the words or lines you want to fix — an “edit” button appears, and the panel opens with your selection ready to edit.

Will: All right, hello and welcome everyone to another episode of Waiting to Be Signed, a special interview episode. We're joined today by James Merrill, aka ToThePixel, aka LostPixels. You've got a couple of different handles there. Trinity is not here today — she's currently in Toronto fleeing lead paint, for those of you who have been listening. So she's not making it to this one, but that's okay. We got James on the mic ready to talk to us about art. James, how's it going?

James Merrill: Hey, it's going good. Really excited to be here and to finally chat with you.

Will: We met earlier this year — at the end of 2023, I think — at a Station 3 event in the city, and hit it off. We've been talking about doing this episode for about six months, and here we are finally doing it. So super excited to have you on. Let's start off how we always do: introduce yourself, give us your background in art and coding, and tell us how you came to find crypto, blockchain, and NFTs as a way to distribute your work.

James Merrill: I've been a practitioner of digital art since around 2003, 2004. I can actually remember the exact moment I got into generative — or just digital — art in general. I found out about Macromedia Flash at the time from newgrounds.com. I loved cartoons and animation and that sort of thing. I was in New York City with my mom, who was an art teacher and also a crazy bargain finder — she'd go to New York and hunt for bargains for ten hours straight, just walking around. Honestly kind of terrible, but I feel like that was in my DNA. I was walking down the street and saw a street vendor with a book that had the Macromedia Flash logo on it. I thought, this is sick, and walked up to look at it. He said, "That's the book, but do you need the software?" I said yeah, I don't have that. So he pulled out this black bin from under the table full of counterfeit software on CDs. I bought a Macromedia Flash CD from him for twenty bucks — it had a printed label with the serial code on it. I took it home, installed it on my computer, and from then on, for years straight, I made digital art every single day.

I got really into DeviantArt around that time, found other artists there, and became extremely inspired working alongside them. I also started illegally pirating a bunch of other software — Photoshop, Fireworks, Illustrator, 3D Studio Max, all this simulation software. I immersed myself in creating digital media in all these different ways, posting it online, getting feedback. It was a beautiful time on the internet, and it spawned these digital art collectives — groups that would come together, decide on a theme, and release artwork together, with hundreds of members all making thematically similar work. That was deeply inspirational. I did that for years until it kind of fell apart. By 2012, 2013, a lot of those people had gone on to careers in design and art — we basically matured out of the system together. The website fell into disrepair, the community fell apart, and I think social media coming around really destroyed a lot of online communities too. DeviantArt became a lot of manga and stuff, which isn't really my vibe.

Will: Pregnant Sonic the Hedgehogs and stuff like that.

James Merrill: So I stopped posting there as much and went through a period of getting really into physical simulations — taking simulation software you'd use for VFX in films, or for engineering projects, and doing psychedelic, artistic things with it. I made these seven-to-ten-minute short films of weird physics reactions and that sort of thing, just to post online. My place then was Vimeo — you can still go back and watch about 45 minutes of simulation work I did. That was a lot of fun.

Then I was browsing Hacker News — I'll never forget this, another one of those watershed moments — and I found out about plotter Twitter, around 2018 or '19. I thought it was crazy cool: people using code and pen plotters to make art, visualizing sorting algorithms, taking raster images and doing hatch fills over them, all these interesting things. I needed to participate. By then my career had gone into software engineering, so I'd been coding the whole time, but never creatively — I actually didn't like generative art that much before this, for various reasons. But when I saw plotter art, I said, "I'm gonna do this." I downloaded Processing, bought an AxiDraw, and started coding and making generative art every day, posting it online.

In February 2021, some of the old digital art collectives I'd been part of started coming back together, and NFTs came around. My buddy Justin Mahler started working with Nifty Gateway, and at first I thought — digital artists have never gotten any respect, never gotten paid, and this is the first time I'd ever seen that change. This was exactly what I'd been waiting for. I immediately started FOMOing, thinking I had to do something but that I was already late. I had no idea what was going to happen. I ended up on Foundation — it was invite-only at the time — and started posting some of my simulation work. Those were my very first forays into NFTs. It was amazing: people would collect them and then want to talk to me about them. I had video calls with people like this one guy, Todd, a member of a semi-famous Christian rock band.

Will: Oh, wow.

James Merrill: I think he even has a Wikipedia page. He bought an NFT from me and was so excited, wanted to chat — and that was mind-blowing to me, as someone who'd made work for twenty years and put it out there and finally had people genuinely excited about it. I did that for a while, and then I found out about Art Blocks, because I was a fan of Tyler's — we'd talked a couple times before, through plotter art, since he did a lot of plotter work too. Then Fidenza came out, and I thought, this is amazing, but I'm late. Like, damn.

Will: What did you like about Fidenza versus other generative art that you didn't like? I'm curious, because I think Fidenza is a piece that gets thrown around a lot as a great example of a very standard thing — a flow field isn't in itself interesting, it's a really tasteful execution of it. Was it the taste level, the color palettes? What was it about Fidenza that stood out to you?

James Merrill: On DeviantArt, there was always a lot of bad generative art — no offense to anyone who posted it there. A lot of people learning to make generative art would post basically noise maps with colors that weren't particularly aesthetically pleasing. I saw a lot of that, and my impression was always that you could never get the level of control through generative mediums that you could by digitally painting things. So I preferred to digitally paint, because I wanted that exact control. Then when I saw Fidenza, I thought, this really stands out. I'd already been a fan of Tyler's — Ectogenesis is one of my favorite generative art projects of all time that he did, and he also did a Max Cooper CD cover I loved. So when Fidenza came out, I was blown away, because the execution was really good, and it was the first time I'd seen such a cohesive color palette in generative work.

I think a lot of people look at Fidenza now and compare it to generative art from 2024, and by that comparison it looks much simpler than what we see today. Look at the project you discussed with Entanglement on your last episode — Entanglement is crazy compared to what Fidenza was then. But I don't think that's a fair comparison, because Fidenza came out so long ago — I know it's only been a couple years, but so much has changed since then. Contextually, it was a hugely important project.

Will: 100%. And very excited for Entangled, which I think is actually opening up a few hours after we end this recording, so I'm excited to go collect that. By the way — in school, did you study coding? Was art a minor, or was it always just a hobby alongside your education?

James Merrill: The school question's funny. I hated school and ended up dropping out of high school — I failed Algebra 1 three times. I'll never forget it, I'd just sit in class either sleeping or disrupting it, thinking, we're never gonna use any of this in real life, why would I ever pay attention? Now I use math every single day, and my math teacher has no idea — it would probably make them very happy, given the irony. I ended up using my knowledge of Flash at the time to get a job in advertising, building all these really weird websites. I did Marilyn Manson's website at one point.

Will: Whoa.

James Merrill: And part of Justin Bieber's website. Flash was always this amazing tool for doing multimedia, interactive things — JavaScript, HTML, and CSS were subpar back then, great for making Craigslist but never great for making experiential websites. That's since changed, but at the time I thought Flash was the coolest thing ever. Then it died. But that's how I got into the field and became a software engineer.

Will: It seems like the entire time you were making art — every single day — do you ever regret not bundling that "I make something every day" gimmick together? You could be a $69 millionaire if you'd packaged that up at some point.

James Merrill: I think the thing Beeple did that was interesting is he never really deviated from a specific subset of work, so it's all cohesive. My portfolio is scattered all over the place — I just couldn't fixate on one genre of digital art for that long.

Will: So you got into plotter stuff, got on plotter Twitter, did Foundation — and then you must have gotten the attention of Feral File, right? Because in 2021 you were featured in their plotter — what did they call it?

James Merrill: Graph+. It was a group exhibition with some of the most prolific and amazing plotter artists out there. It was amazing — Casey Reas sent me an email, which touched my heart, a true legend in the space. I got to work with him and all these other artists, meeting weekly to discuss our projects. It ended up being quite successful, and quite a learning experience, because I was committed to doing thirty NFTs and thirty plotter works, which is a lot when you only have one plotter. I had a newborn at the time and had to send prints around the world, so I had to deal with logistics seriously for the first time ever. Before that, I'd done a couple of Etsy drops where I really cheaped out on packing materials, and the plots would get ruined on the way and sent back. I learned a lot of lessons there, and then I was able to tap into a group who taught me the right way to do it — spending the right amount of money, that sort of thing. That was really helpful and, I think, a big signal to me that I could commit to being a full-time professional artist rather than doing it on the side of a professional career.

Will: What was it about plotters? In the last year or so I've gotten way more into collecting plotter work, and I can't quite put my finger on why. There's something about that translation of code to the physical that's just intrinsically cool to me. But as someone actually making it — what is it that's so cool about plotter work, do you think?

James Merrill: There's a multitude of things — hard to put a single point on it, because I think the artwork possesses different qualities when you know it's generative in nature. For one, every artwork made by a plotter from a generative algorithm is unique. They can be part of the same series, but they're all different. That already sets it apart from a print — you can print generative art, that's fine too, but for me, I was really inspired seeing people use vintage plotters and modern plotters, writing all sorts of diverse algorithms, and there was always this distinctive physicality to the work that I'd never find in a printed piece.

A lot of generative art tries to mimic traditional art by adding paper textures or brushstrokes, and that can be intriguing, but for me there's nothing better than actually seeing pen ink or pencil or paint on paper, holding it up, and knowing I wrote the algorithm that made it happen. It's hard to explain, but I try to explore maximizing that little space. For my new project, BAYC, I'm trying to make drawings with ridiculous amounts of detail because I want to trick people — I love it when someone looks at a piece and thinks, "Wow, you drew that, it must have taken forever," and it's like, actually, this pen plotter did it. The most recent one had 427 meters of ink drawn and took 10 hours. If I tried to replicate that by hand, it would take hundreds of hours, but I can produce it in my studio and play that trick on people. I just love that.

Will: You'd be like, "Yeah, it did take a lot of time — I just didn't do it by hand. Twenty hours to plot, and a lot of pen swapping in between, I'm sure." You have this upcoming project — you've just started sharing some stuff on social media, not talking too much about it yet, just videos of the pen plotter going. They look insane, the detail really comes through even in what you've shown so far. You were nice enough to share some info about the project with me before this interview, and one thing you mentioned was that you're seeking to use causality instead of random values in generative art, and that this project is an expression of that — a lot of it drawn by hand and then, I'd guess, algorithmically inserted somehow. How does that work? These things are so complex — my layperson's understanding of generative art looks at this and thinks, it's like a packing algorithm. We've seen other artists do this, usually with circles because that's easy, or with an assortment of geometric shapes, which is harder because now you have to work out how to measure center point to center point and figure out the borders — all the ways you solve a packing algorithm. But here, nothing is a simple shape. To have it all fit together like these really fine puzzle pieces — it looks more designed than generative. So, to whatever degree you can explain it: how does this even work with code? It seems almost impossible.

James Merrill: I'll tell you all about it — I'm actually planning a full walkthrough of how this thing is built at some point, since I get that question a lot. But I'll start here: I want to get away from pure abstract generative art. I think figurative or recognizable elements aren't used enough, and I really want to play with that. I've been loving what Studio Yorktown posted today — work-in-progress shots of these crowds of people he's creating. I've had that idea myself, but it's really hard to make generative people. Ask Harold Cohen. Beyond that, I also want to get away from relying on randomization for everything. I've done work where I hand the randomization off to MIDI controllers and let people manipulate the parameters themselves — it starts to become more procedural than generative when you do that, but I like having more say in what's going to happen, and only using randomization where there's no other tool for the job.

For this project, one of the things I wanted to do was build tools to make generative art. I built something that's more of a pipeline: I do illustration, then bring those illustrations into the generative algorithm and use the algorithm to manipulate them. I'm not the first to do this — going back to Tyler and Feral File, his work with Flight does something similar with a vectorization step. I draw a lot of these elements, then use generative code to turn them back into X and Y coordinates from SVGs and manipulate them — changing their colors, scale, dimensions, that sort of thing. I commit to drawing hundreds of tiny elements — so small the pen nib can barely render some of them, but I know what they are and I care. I'll draw 10 different types of cars, 100 different types of buildings, and build a pipeline so I can keep doing that continuously. When I'm bored, when I'm at the airport, when I have my iPad, I just draw and draw and draw, then feed it into the algorithm, and the drawing becomes more and more elaborate.

The funny, tricky part is that the more you do this, the less weight each individual element has — if I do this a thousand times, you're probably only going to see one example of each sprite on the canvas. There are diminishing returns. But I started playing with the idea of a Where's Waldo type thing — making it so elaborate, but placing just one of a particular thing, not telling anybody, and seeing if people find it. Creatively, it's been a lot of fun.

To go back to how the algorithm actually works: it's basically a grid that I'm filling. There are random walkers that move around, and sprites of various shapes and sizes get placed. There's also the concept of a z-index, which is notoriously difficult when you're not working in WebGL, because the z-index determines what's in front of what — like in a landscape, trees in the foreground occlude the mountain behind them. Doing that with vector graphics is challenging. Zancan's work does this to great effect, in a way I've tried in the past, but it got so slow for me that I ended up developing a mathematical approach instead — using intersection detection to figure out the stacking. It's running an enormous number of calculations to determine whether that tree is in front of that mountain, and if so, draw the tree and not the mountain. Just doing that at massive scale.

Will: That's got to be Algebra 2 at least.

James Merrill: Dude, I don't even fully understand what I'm doing. I didn't even know there was a name for it until recently — I think it's called computational geometry. I basically use intuition and visual debugging to solve these problems, with a little ChatGPT on the side.

Will: So hearing you talk about this — you're constantly adding these sprites to a library the algorithm can pull from. Given how much time plotting takes, it's probably going to be a smaller series, and probably curated. That's where curation comes in, right? If this were a long-form, JPEG-only project with 500 outputs, you might get visual fatigue — this one has too many overlapping elements compared to that one. But by curating it and taking the time to actually plot each piece, you can preserve the diversity and individuality of each one.

James Merrill: There's no great way to do physical NFT releases — we've tried redeemables, we've done it the Feral File way, they all work decently. But I think you want to curate what you're making before you release it, because that lets you preload the release with some physicals already made. If I do 50 of these, it's going to take 500 hours of my time and my AxieDraw's time — it would be crazy to sell them all first, without knowing what they're going to look like, and then have to produce them after the fact. I don't mind doing long-form work, but for physicals, curation feels right.

I'm also experimenting with diversity within the algorithm — we talked about rare elements that appear in maybe one of every five or ten outputs, but also how I distribute color. Maybe it's completely random, maybe I use zone coloring, maybe I use different inks. Another thing that attracts me to plotter art is the ability to acquire different art supplies — I love buying different pens and inks and experimenting with them. I think I have about 60 different bottles of ink now, which I use to create different color palettes.

Will: I'm sure you know Schwittlick — he has a series called The Long Run. Have you seen that?

James Merrill: I don't know if I have.

Will: Go check it out — I don't think it's a project many people know about. It's on Foundation, I believe. The pieces are really long, almost a meter of paper. He had these old markers from the '60s — or maybe the '80s, I don't remember exactly — and he'd take six of the same color, but because they're so old, the inks had degraded in different ways. He'd draw millions of dots in a row, letting one marker completely deplete before moving to the next. So the one marked "green" ends up with pink and red in it, and traces of green hue, because each one decayed differently. He's honoring these old inks, but the generativeness of it is in how they've changed over decades. Ken Consumer plays with something similar — inks and materials that change on the canvas as they overlap through the process. But that's an aside — sorry, this is your interview.

James Merrill: I love that. Real life is the best generative system out there. I love analog photography, for example, because digital photography is so perfect all the time. I like imperfection in the things I look at, and it's really hard to achieve that purely digitally. If you make something in real life, you'll often find those inconsistencies that are really the magic of it. That project sounds like a great example of maximizing and focusing on that. I've got to go check it out.

Will: Looking at some of your other works in progress, it struck me how much, from a distance, these can look like circuit boards — not just roads, but tubes and wires and all sorts of things connecting the elements in these huge drawings. Was there a certain point where you thought, "oh yeah, this is also kind of a nod to circuitry"? It reminded me of a project on fx(hash) called Impressions of Order from 2022 by Nibswit, which is very explicitly an ode to circuitry. But here it just really sticks out to me that there's this connection.

James Merrill: I love that project, by the way. But I wouldn't say there's any influence of circuitry in this — it's more of a happy accident. I hear that comment a lot, and for a second I thought maybe I should put some CPUs and RAM in there. But for me it's really about cityscapes, order, complex living systems, and the appearance of disorder from a distance versus the complete control and autonomy of these systems when they're actually in use. If you think about a subway system, in some ways it's extremely chaotic, but in other ways it's on time, predictable, extremely measurable. You can wait there and it'll tell you exactly when the train's going to arrive.

Will: Well, there's still time — the hidden object could be a tiny little CPU buried in the middle of it.

James Merrill: Actually, one of the things I want to do with this project — this is kind of the first iteration — is change what you're seeing thematically to be maybe more like a computer, or a specific city. I'm playing around with that idea, and I have some things in the works to further customize and explore this as more of a framework than a solo piece of art.

Will: Right on. One other plotter-related question: you're releasing this as a JPEG — the NFT will be the JPEG, and there will also be a physical piece. When so much of the work takes place on the screen, but the final output is a physical thing, what's that process like? Are you able to look at something on a computer screen and know what it's going to look like on paper? Or are you ever wildly excited or disappointed by how it ends up looking physically? Is that easy to predict, or does it take a lot of experimentation and waiting for the plotter to show you what it's actually going to be?

James Merrill: There's a lot of experimentation. Mixing colors doesn't work very well on computers, so people are always trying to solve this — Lars is a great example of someone who's done a lot with this particular problem. When you mix, say, yellow and blue on a computer, you might not get green, you might get gray. Trying to mimic the physicality of inks overlapping is unreliable, to say the least, and there's still not a great way to do it. It's an ongoing problem, and that's one thing that definitely bites me.

Another thing is that inks have all these properties that make them more than just a flat 2D color — the angle you look at them, their saturation, how fast the pen is moving, whether it's skipping, whether the nib has an issue. All these fun little things happen in real life that you just can't replicate digitally, unless you really want to focus on that, which is not how I choose to spend my time. So there are a lot of what I'd call happy accidents, a lot of experimentation and trial and error, and that's a lot of the fun, honestly. If it were really, really accurate, like a print, I probably wouldn't be that excited. Those little discoveries are what set people apart.

I look at artists like Licia, who does a lot with watercolor and pigments you're just never going to be able to mimic digitally — that's the joy of it. And the way she renders her work digitally is interesting too, a completely different pipeline than I've ever seen before. Some of her programs write programs, so her output isn't a static JPEG or SVG file — it's a Python program with instructions for how to draw the thing you're seeing on the screen. That's so much more interesting to me than looking at a static JPEG that looks like a painting.

Will: It's got kind of a Sol LeWitt vibe. A good way to transition into some broader topics here would be to talk about being an artist versus being a craftsperson. This is a debate we've been circling around a lot in the NFT and generative art community — what is art versus craft, and which artists are making high-concept, amazing art versus which are just insanely skilled coders pounding out really impressive projects. But the open question is: is it good art? Correct me if I'm wrong, but you've said you consider yourself more of a craftsperson than an artist. How do you think art versus craft applies to code-based art like this?

James Merrill: I don't really know what I am, for the record. And I don't know if there's any line in the sand you can draw here. I've thought about this a little, and the way I'd approach it is: purely writing code and making beautiful images is probably closer to craft. Coming up with a story about that work is more like art. A lot of people in the contemporary art world need to see that backing story, and a lot of times that forces artists to come up with, frankly, bullshit that backs their work — and this is true across the contemporary art world, I'm not pointing fingers at generative artists specifically. You basically have to manufacture it. That's kind of the dirty secret, and whether people believe you or not, I think, influences your success in that market.

There's a lot of room for authentic storytelling in any art form, including generative art. As an artist, you really have to think about why you're making the work — you might have a great story, but it's also a matter of being a good storyteller. Especially in this world — social media, web3, long-form work — it's incredibly hard to tell stories. That's one of the things I'm focusing more on: moving away from purely abstract work and doing things people can resonate with a little more easily. I think about that a lot. That's all I've got.

Will: I think that's a great answer, and I don't think it's cut and dry either — it's something you almost have to grapple with for each individual artist, each individual project. Sometimes the code itself can be so crazy that the code is the art, and maybe the outputs don't matter as much. There might be something so special about how the code was written, or something the artist figured out in the process, that only a few people can really appreciate — and you just have to take their word for it, that this is really cool, trust us. And then there's other work that's high-concept, or just purely beautiful and cool, and I appreciate that too. I don't think we have an answer, but I really enjoyed your take on it.

Back when we met up at NFT NYC, you were going to an informal artist gathering — a salon-style event where everyone had their own topics, unfortunately I didn't get to go — and I know you were going to talk about this idea of maybe almost unionizing as artists, or collectively getting together to take more control of your destiny in the space. What was your intent going into it? What problems were you trying to address, what were the goals, and did anything come out of it? Or is it still this big open-ended question of how artists maintain parity against the platforms, and to some degree the collectors too?

James Merrill: That was at TokenArt, and it's an unconference — the participants guide the conversations, which aren't lectures or presentations so much as facilitated discussions on specific topics that the attendees choose. I came up with a deliberately provocative title, because there's something funny about how, whenever you have a job, threatening to unionize scares your boss. Not that we have bosses, but we do have this relationship with platforms where there's a give and take.

Unfortunately, there have been a lot of stories over the last couple of years of bad behavior from various platforms — though I have to say, plenty of platforms do a really good job too. But there are a lot of common issues and friction that artists experience, and because we're essentially sole proprietors, we each have to deal with it alone. That doesn't necessarily make sense to me. A good example is negotiating a contract with a platform — we all have to do that individually, and when we don't coordinate, historically bad things happen.

For instance, I know of a recent case where a platform is holding an artist's physical goods and won't give them back until the artist pays money that was never part of the contract. All this confusing, terrible stuff makes me wonder why these people operate this way — and it's because there's not a lot of oversight. What does happen is artists talk. We hear about this kind of thing, and we won't work with those platforms in the future. I certainly won't work with a platform knowing how they've treated another artist. So some of it is just that — having a community and a conversation about our experiences.

But there are levels to it. I don't want to say "union" necessarily — again, it was a provocative title. It's more like a cooperative, a collective, a guild. In the conversations we had, it was more about what this even is and what the benefits would be. Starting a labor union isn't the right answer, but providing a safety net for artists who choose to participate is closer to it. We want to be part of this movement of generative art and take it as far as possible, and to do that, we need to figure out how to work together in various ways. That's a big part of it — and the other part, which you alluded to, is how we shape healthy relationships with platforms.

Will: It's a really challenging issue. When I first started thinking about this—not unionizing so much as advocacy for artists—it was around the end of 2022. The market was still okay but starting to cool, and we saw a lot of new platforms coming out. Since we're a show that talks to artists, we'd ask them about these Genesis drops platforms would always advertise, and say: the platform builds legitimacy off the success of your work, so what concessions did you get? A minimum guarantee? An upfront payment? And in almost all cases it was, "Oh, no, the people just seemed really nice, they really love artists, so I let them have this project, and they're gonna market it a little as the genesis drop." Which is cool, but it's the responsibility of the people who run these platforms to own that and compensate you for lending your voice and legitimacy to it. That's unfortunately what makes platforms succeed—a series of good releases that collectors like, so collectors build trust in the platform and the people who run it, and it becomes a momentum thing, because then more artists want to release there too. Art Blocks is a great example: they showed that projects generally mint out and go up in value—or, I guess, until recently.

So I think, this is where a manager comes in, an agent comes in. We've seen people like Simon Says pop up to help manage and negotiate on behalf of artists, but not every artist is going to get representation like that. What would it look like collectively? You'd probably need one person whose full-time job was to advocate for the group and say, here's the standard for anyone in this collective.

James Merrill: I think that's just it. Going back to the landscape—if anything, there's going to be more and more consolidation of platforms. So we really need to get it right in terms of how our relationships with platforms evolve, because there will be fewer options out there. There won't be another store to go to if you don't like the prices at this one, especially in a down market. In terms of the conversation we had, it doesn't need to be someone within the current community—it's a matter of finding sustainability through various means. Artists are actually pretty good at that: we produce a product we can use as leverage to purchase time from legal avenues or an agent, establish a boilerplate we share and update, that sort of thing. I'd like to see something like that happen. The management of it is worth studying, just to understand how to facilitate that structure without it becoming your full-time job. That's the biggest challenge, honestly—finding people willing to give away their free time to further the whole movement.

Will: Is there anything in the works right now, or is that a secret?

James Merrill: I can't say anything's happening anytime soon, but there are levels to it, and we've built up a lot of different assets as a community over the last couple of years. That's just going to continue to grow.

Will: Playing devil's advocate—it's a hard time to be asking for more, to be establishing a floor on artist compensation and rights, because it feels like every platform is either breaking even or losing money right now. Going to them and saying "it has to be this way, you have to sign this document, you have to advance this stuff"—they're going to say, "we might be closing in six months." Having talked to a lot of platforms, it just feels precarious. Do you get that sense? Is that something you think about when deciding who to work with—will they even be around a year from now?

James Merrill: My art practice—I'm not a crypto artist. This whole thing is amazing, and I'm happy to have experienced it, but in some way I look past it, because especially with plotter art, there's no deep necessary connection to NFTs. And I'd say to platforms: you're definitely going to close in six months if no artist will work with you. We are the people who make the product you sell. That leads to one of two things: either they come to the table and negotiate too, or they go to contemporary big-name artists with teams of ghost coders who make generative art, which is fine too, but I worry that's the direction we're headed. Then a lot of the magic we had will have matured away, in a sense. The best times for me were really Hic Et Nunc, living in the warehouse as young artists doing amazing things, when there wasn't a lot of prestige or gentrification of the space. I feel like that time is gone to an extent. But if things get bad enough—if collectors leave, artists leave, platforms shut down—that time will come again. That's exciting to me. I'm not rooting for anything to collapse, but I love the freedom of being a digital artist, and I don't know that the world we live in now is optimal for that level of freedom.

Will: Unfortunately, the vibe has shifted toward AI stuff—unfortunately because this is a generative art podcast, not an AI art podcast. I know a lot of people love that stuff, but if you look at the most successful releases this year, they've been AI. I'm thinking particularly of what Verse has put out—every time they do an AI auction, the stuff goes nuts, and whenever they do generative art, it's like, "that didn't do as well as I thought," or "wow, that price is really low." I wonder if the trajectory of AI will follow what we've seen with generative art: a lot of people doing it, everyone's excited, and then in eighteen months it'll be really sobering—everyone looking at their bags going, "why did I collect all this, why did I pay so much for it?" I don't know what your personal opinion on AI is, other than that ChatGPT helps you with math. Any hot takes on what's going on with the AI markets, or that style of art in general?

James Merrill: I can't really speak to market dynamics—I don't follow AI auctions much. AI art in general, I like a lot of it. I'm a big fan of this guy, Data Velvet. But I have to think back to my past: being that person twenty years ago uploading art on DeviantArt, posting analog 35mm photography on Flickr, and then having it all sucked up and used to power these machine learning algorithms doesn't feel good. There's no credit given to any of these artists, they receive no royalties. Not a fan of that.

If I had to splice it: the technology's cool. There are a lot of different types of AI—look at Casey Reas's recent work with GANs versus Midjourney. It's not all the same. So in general the technology is interesting, useful, and cool. I do worry it's going to kill all the creative jobs in the world, or maybe we all just become artists, I don't know. But the training data is where I get iffy, because if it's trained without permission on things people put time and energy into and released freely on the internet, their intention was never to power an algorithm that subtracts attention from their blog or portfolio to power something else they had no desire to be part of. It's done without any permission given.

Will: I have similar concerns about that particular genre—the prompting stuff in particular. I've heard the analogy to sampling in early hip-hop: artists were ripping samples without compensating anyone, and sure, they're creating something transformative and new, but now you have to license your samples, and if it's a big commercial artist, they pay for that. The recording industry went through an entire process around this. As an artist you could take a punk stance on it, but I don't feel like a lot of the AI art being made is punk at all. It doesn't nod toward the fact that it's stealing and ripping off—it's trying to be more serious and refined than that. I'm sure there are artists doing that, but as someone just casually observing, I'm not seeing anyone lean into the fact that, yeah, you're stealing and you know it. It feels like that should be acknowledged, maybe even be part of the work, because all of this imagery that gets prompted up is coming from somewhere.

James Merrill: I think people are afraid to comment on this, because it's okay to say you don't like how AI art is made. For me, the question of whether it's a transformation or a derivative doesn't even really matter. Back in my days as a Photoshop artist, the worst thing you could do was take someone else's JPEG, add filters, transform it, and pass it off as something you made—that would get you blacklisted, no one would want to work with you. That's a maximal example, and I'm not saying that's exactly what's happening here, but if you're training your own models—if you have a portfolio of your own work you're feeding these algorithms—that's really cool, that's authentic to me. If you're just using off-the-shelf stuff that's already trained, I don't know. The hard part is, a lot of times I really like the imagery. It's the kind of stuff I find inspirational, the kind of thing I wish I could make by hand but can't because I'm not that talented at painting. So it's the best digital art I've ever seen made with the least amount of effort, because it's trained on years and years and hundreds of thousands of sweat hours and equity put in by people who actually painted it.

Will: That's got to be partially why it's doing so well—you can release a batch of five hundred images and people will buy them because they look cool. But is there infinite demand for cool images, or will we fatigue on that? You can't just release five hundred images every three months forever—or maybe the price will come down over time, I'm not sure. A lot of people don't seem to care about how these models are trained or who's making money off it—not even on the artist side, but the intermediaries, the companies who scrape the data and build shareholder value off great artists using their product. Every time an artist uses it and makes a splash, that's great advertising for the product itself.

James Merrill: I think it kind of goes back to art versus craft. Is there a strong narrative backing to this work, or is it aesthetically pleasing and kind of vapor? We talk about "generative goods" a little, and when is IKEA going to sell generative art? IKEA is totally going to sell A.I. art before they sell generative art. There's such a big market for people who just want cool images in their house, and A.I. art is a great fit for that. That's why it's proliferating everywhere. That's why NVIDIA is the number one stock right now -- because this is going to change everything we know about being creative with digital assets.

Will: The worst case scenario for some of these companies -- there are lawsuits going on right now -- is that it goes the same way music has gone. Already some of these companies are paying to license Reddit's data. Google's paying a licensing fee to Reddit to get that. Quora's another site where you don't just get access to the data, you have to pay for it. Eventually they'll all just pay for the data, and there might be really specialized tools -- one that pays for great art data and is just the piece for prompting up abstract art or landscapes, that kind of thing. We might see a fragmentation in the market versus these big A.I.s that seem to have been trained on everything. And then they'll pass the cost on to the people who use it, because they have to pay for licensing fees, which means it won't be as free or as cheap to use as it is now.

James Merrill: The big question is: do they already have enough to have escaped the velocity of needing to do that? Have they sucked up enough free stuff from the internet to power what they need? From my understanding, it's a lot of specialized information now -- if you're a mathematician, OpenAI, or I think it was Scale.io, may hire you to give expert summaries of things. That's where I see this going. But have we passed the point of it getting its fill of digital art? Which is kind of sad -- then it's locked into what it has and it's never really going to evolve. The way we use it may evolve, but the training data is kind of there. I think we may be getting to the point where it doesn't matter anymore.

Will: I don't know what some of these products would look like if they were trained on purely free public data. I can't remember if it was Midjourney or Stable Diffusion that got in trouble because they'd taken Shutterstock images, and people prompting up images found the watermark showing up -- because it was so specific to Shutterstock that the model couldn't get around it. That became proof they'd taken the images without paying to license them. So how much worse would that model be once you remove that set of photographs? I have no idea. But clearly the people making these things think they need that data badly enough that they're willing to not even ask permission or pay for it.

James Merrill: This is the duality of being an artist too. When you use social media, you're effectively giving up all rights to your work, and it's going to train someone's model. But you need to do that, otherwise no one's going to know who you are. That's a common problem you see being exploited everywhere. And -- slamming Adobe here for a second -- they're some of the worst offenders. Yes, they make good products, but they've made a lot of awful moves along the way, maximizing value at the expense of creatives. I absolutely cannot stand that.

Will: I want to jump back to platforms for a minute -- platforms, chains, artists surviving this horrible market. It was around the end of last year when we first met and talked about some of the Art Blocks changes. It's been a while now; they've announced a lot more, including Studio and their new take on Curated -- going more focused, more revolutionary, really trying to break the mold, probably only doing five or six releases a year. How do you feel now, as someone who's done a Curated release with them? You're obviously in the Art Blocks community. Do you like the direction Art Blocks is taking? Is Studio something that appeals to you? I know you listened to our interview with Eric, so you've heard his vision there. How do you think about Studio versus other open platforms?

James Merrill: Art Blocks -- love them. I did Ori, a Curated release, in November 2022, and it was a life-changing moment for me. I worked on that project for a year. Immediately after I finished Feral File, I committed to doing a long-form project and getting it curated on Art Blocks. I tweeted that, then put my head down and worked constantly, every night, on Ori to get it to that point, because I really believed in the work. I don't think I've ever been so focused in my life.

Art Blocks is great to work with, and they're in a really tough position, so I like that they're evolving. We need a revolution -- big changes. It would be really easy for someone like Art Blocks, at the top of the heap, to get complacent and stick with the recipe that's worked in the past. I commend them for trying new things; that's exactly what we need.

Studio intrigues me. I'm also a big fan of Highlight -- I know you've spoken with Nat, and I love how agile they are, how they lean into open editions, low costs, working with Farcaster and Frames. That's what I'm going to measure Studio against: can it introduce newer features at a quick cadence, stay close to the current sentiment, try new and interesting things? Because the whole point is that it's supposed to be a place for artists to experiment. I loved the "die with the most likes" drop that went up recently -- that was amazing, and it spoke to me because I hate Microsoft Teams so much that I thought, this is a great use of this technology. I want to see more of that -- artists doing really small editions, maybe even generative one-of-ones, all these things we haven't tried before. That's exactly what we need to restart this market. I hope collectors come along for the ride. I think Studio and Highlight are compelling in similar and different ways, so I'm measuring them together.

With Curated, the bar is obviously much higher now. I'm excited to see what the first couple of releases look like, because they need to be epic -- they need to blow our minds, be as good as good can be, and not just repeat what we've seen before. If Fidenza 2.0 came out on Curated, I don't think it would be enough. It needs to be as cutting-edge as we've seen, with a strong narrative backing, maybe some tie-in to performance art or an exhibition installation -- that's the mark they need to hit. I understand why that means only a couple of releases a year.

I think Verse is doing a good job at a lot of this too, so that's where the real competition is -- and competition is good, because it creates pressure to innovate and stay ahead. Art Blocks is in a super challenging position: big team, first to market, a lot of baggage because they had to innovate first. The way their smart contracts are written, the problems they had to solve -- like creating autonomous Dutch auctions. For a while that wasn't a thing, and artists had to go in and change the pricing themselves, which is crazy to me. They were first to solve that, and now everyone does it, so there's this commodification -- they can't just lean on "you can publish generative NFTs here" anymore. They have to do a lot more. I think they have some good ideas for how to do that. I don't know where it's going to go, honestly, but I want to participate. I don't know if I'll be using Studio to start. We'll see where it goes.

Will: It's scary to me -- I tried to get at this with Eric a bit in his episode. Of any platform, Art Blocks probably has the most staff, by a multiple. So to take your marquee thing -- probably your biggest source of revenue and fees -- cut it down severely, ask for so much more, and also make things more experimental that people might not want to collect... The Operator drop was great, this generative dance performance thing, but you can't just do six generative dance performances a year. Not everyone wants to keep collecting generative dance. And if it's something really multimedia, off the wall, and hard to enjoy because we don't have displays that let us appreciate animated works or sound pieces the way we do flat, still art -- what happens if these Dutch auctions go to the bottom, only mint out half, and they don't get the fees they were hoping for?

The compensation is supposed to be Studio, but artists have a lot of options for open platforms. If Studio doesn't have exactly the tools they want, you can go to Highlight, you can go to fx(hash). Maybe a project isn't good enough to make it into Curated because it doesn't have a year-long plan of performances and installations, so it goes to Verse or another curated site instead. So I wonder what their projections are for how much they'll make off Curated, and how much of that gap Studio is supposed to compensate for. It's not just the platforms that are consolidating -- artists too are becoming more precious with their work, and some people have been punished for experimenting and putting too much out there. So if you have this open side for people to put stuff on, hoping to collect fees that way -- I know it's crass to talk about it, but they need fees, to pay their staff and exist as a platform. I don't know how they make up for the volume on Studio, especially while trying to be experimental and take risks on the Curated side. I don't know if that makes sense.

James Merrill: I think what they probably did is analyze the collectors. They likely have the best analytics and information on the market of anybody, because they understand and have relationships with a number of big-name collectors, and they've seen through their own private accounts what's happening. I'd guess they divided the market into groups — high-end collectors on one side, lower-end but more frequent collectors on the other — and tried to make a place for both to exist.

What I'm curious about is the latter: turning people who are interested in low-cost editions, which I think is a trend that's here to stay, into people who'll spend one or two ETH on a single piece because they love it that much. That's the real challenge. Getting more low-cost collectors into the space is a great idea, and I think that's what Erick is going for with his generative goods concept — creating low-friction experiences, like cards you can hand out. That's all great, and we need to keep doing it. But the big question is: how do you turn someone buying their first NFT into someone who'll spend serious money on one? We saw that happen organically in 2021 — it wasn't hard, because people bought low-cost NFTs that suddenly became worth enough to go buy something for 10 ETH. Now it's a lot more challenging, and I don't know what the answer is.

Another angle they're taking is appealing to the traditional art market — bringing over collectors who buy physical works and converting them to digital in a way that makes sense to them. It's a better time than ever to do that. We've never had a technology that creates ownership around digital goods before — we finally have that — and it's a matter of convincing people it's worth doing. I think Art Blocks has done a good job working with places like Christie's and Sotheby's and other major institutions to show it's possible. Verse is also doing this well, giving meaning and context to something that's otherwise just a nebulous NFT — showing its value by displaying it in specific, curated, high-touch ways online.

Will: We're big fans of Verse too — we've had Jamie on a couple times. In a lot of ways their strategy is different from where Art Blocks seems to be heading. Last time we talked to Jamie, he made it sound like even though there are a lot fewer collectors now, there are still enough with deep enough pockets that if you keep offering them really good art, they'll collect it. He wasn't as concerned about crossing over new collectors or building the base — more about being at the top of the heap for the collectors who are already here. Verse has made a pretty good case for that this past year — they've put out incredible projects, many of which would have easily made it into the old Art Blocks Curated. Have you worked with Verse before?

James Merrill: I have not. I love them, I speak to them regularly, but I haven't officially released with them yet.

Will: Any possibility Busy ends up there?

Busy — James Merrill

James Merrill: I don't know.

Will: You've got to get your boilerplate contract out — get them to guarantee you twenty grand upfront and X amount of marketing spend.

James Merrill: I'm a radioactive dude, still trying to figure this out with everybody else. That's why I lean on other artists — I need that support, because we're all trying to figure this out. It's such a new thing for so many of us that having a community to back you is key.

Will: Whatever you do, some kind of live show would be awesome — getting the pieces up on display somewhere so people can really come in close and look. Or you'd need a website tool that lets people zoom in to see the detail, because even in the lower-resolution images you've shared, you can tell there's so much going on that would reward really deep viewing. What are some of those wavy, vortex-looking things supposed to be?

James Merrill: They're fun. I'd say they're kind of like a body of water, but with a bit of a psychedelic portal element too — almost like acid paint. It's up for interpretation.

Will: Gotcha. We've been going for a while now, James — let's do some rapid fire and call it an episode. What do you listen to while you code?

Busy — James Merrill

James Merrill: The Lane 8 Summer 2024 mixtape, as of yesterday.

Will: Is that hip hop? What is it?

James Merrill: No, more like deep, melodic, progressive house music. But I listen to all kinds of stuff — I'm a huge fan of Burial. Do you know who Burial is?

Will: Yeah, Untrue — I know the classic stuff, but I haven't kept up with his more recent releases. He doesn't put out much, right?

James Merrill: He's the world's coolest artist, in my opinion. He invented his own genre of music, he's completely anonymous, doesn't play live, and just makes stellar music. He put something out two days ago with Kode9, so check that out.

Will: Wasn't his original stuff what people were calling dubstep before dubstep changed — like that was dubstep first, and then it became the bass-heavy, wubby stuff it is now? Am I wrong?

Busy — James Merrill

James Merrill: That's accurate, but I think that was a sign of how immature the dubstep scene was at the time. To me, dubstep was more about these really low-end frequencies with a magnetic effect. Burial really created future garage, which isn't dubstep, but there wasn't really a term for it yet, so he got slapped with that label since dubstep was so popular around the same time he came out.

Will: All right, I'll check out Lane 8 — I like house and progressive stuff. Who would you like to hear us interview on the show?

James Merrill: Licia He, the person I brought up earlier. I think she's amazing — she does some of the best plotter work out there and has a really interesting way of thinking about generative art.

Will: She had work in the last Graph+ release with you, right?

James Merrill: Correct.

Will: We'll reach out to her. Last one — other than Busy, is there anything else you're working on? Any code-only or digital-only release you can preview for us?

Busy — James Merrill

James Merrill: I do have another thing going — a collaboration with an artist I truly love who doesn't do generative art. It plays with the idea of extracting randomness, but instead injecting this person's mindset, their decisions, the way they look at art — turning that into data and feeding it into the algorithm. It's an exciting way to explore what happens if generative art isn't so random — using randomness to the smallest extent necessary to achieve long-form or unique outputs, rather than leaning on it everywhere.

Will: Right on — maybe that's a Studio release. James, anything else before we go? Did you have a good time recording?

James Merrill: I did.

Will: Have you done podcasts before?

James Merrill: A couple. I've been pretty quiet this year because I've been heads-down working. But nothing else to add — just thank you guys, you're awesome. I love your podcast. It's great to have people talking about this stuff in the space.

Will: Glad to hear it. Everyone should follow James on all his socials to keep up with Busy and find out where it might release. I'm excited to see the video and some higher-res images soon. Thanks so much — we'll call it an episode there. Bye, everyone.

Busy — James Merrill

James Merrill: All right, see you guys.

Will: Grail of the Week. Could be time. We're waiting.

James Merrill: We're waiting to be signed.

Change log

  • Initial transcript — auto-transcribed (AssemblyAI) and readability-edited.