‍

Beeple on Infinite Creation Machines

Mike Winkelmann (Beeple) speaks with Peter Bauman (Monk Antony) about what motivates the artist—known for his Everydays sketches—to engage with generativity and modern AI’s deep learning tools.
‍
Peter Bauman: You’ve been completing Everydays for over 18 years now. That kind of daily commitment, even for one year, is impressive. What have these two decades taught you about image-making and the value of persistence for artists in an age of fleeting attention?

Beeple: It started May 1, 2007, so just hit 18 years. It's really changed focus in some respects. At the highest level, it's just about practicing, having this daily thing to get better—just making art.‍

‍The very first ones were drawings. I wanted to get better at drawing so I did a year of that. Then I realized I started progressing a lot faster than I was prior to that. Because I'd been making art—when I started the Everydays—for like eight years, starting in 1999.‍

‍I already made many short films and all kinds of other things like creative websites and little weird, abstract things. I saw it as a way to get better—faster in terms of techniques and stuff.
‍
‍Then over time it switched to Cinema 4D, learning Cinema 4D and all the technical aspects of it. Once I had a base-level proficiency, it switched more into actual image-making in terms of trying to get better at color and composition and those things.
‍‍

Then it changed into being more about, “What can I say narratively with this work?” 

‍
‍The work became less abstract. Part of that is the tools and what's possible with them in just a couple hours. That has, you know, shifted. Now it's actually in a weird spot. It always felt like I was using the most advanced tools to work as fast as possible and now it doesn't feel like that because I could type something into an AI and make an image in, like, two seconds.‍

‍Sometimes I’ll spend, like, two hours making an image and it doesn't even turn out as well as if I would have used AI. So I'm in kind of like a weird place with the Everydays at the moment. Even maybe six months ago, the images I was making would have been very hard to make with AI because AI just had too many weird complexities.‍

‍Now the AI is starting to get good enough where it could make “Donald Trump sitting on top of a chicken, holding people, holding signs above him.” You can now be very granular in your explanation and it can get pretty close or sort of like nail it.‍

‍It's definitely a journey but I think having appropriate expectations each day is something that can keep it going because it's really one sketch. It's something I'm sitting down and spending a couple hours with. This is what a sketch looks like in 2025.
‍

‍

Peter Bauman: You have these daily snapshots of your work and thought but I'm also really interested in your sculptural work. Something that's been around but will be shown publicly later this month is Diffuse Control, a collaborative, kinetic sculpture.

Recent shows like Electric Op and Electric Dreams at Buffalo AKG and Tate have explicitly tied the history of kinetic sculpture—going back to Dadaists like Duchamp and Tinguely—to digital art. Now you’re bringing these conversations to modern AI art with deep learning. Why did you choose AI to construct your creative vision with kinetic sculpture?
‍

Beeple: The idea to use AI follows that I'm always interested in trying to make something that I've not seen before. AI really enables that.‍

‍
‍There's this idea that some people have that AI is just regurgitating things that already existed. Especially when it comes to video, that's in some sense not really the case. You can make video with it that I don't know how it’s done.‍

‍I made, for over a decade, commercial 3D animation and things like that. I still have no idea how you would even begin to make some of the video that you can produce with AI through traditional 3D workflows of modeling, compositing, After Effects or tools like that. So it does create a kind of entirely new aesthetic.
‍‍

At the end of the day, creativity is pushing the pieces around in a different way than anybody's seen before. And we've created something that can. 

‍
‍When people say it's just recreating things. It's like, well, you've not seen good AI. Of course there's all kinds of bad AI, just like there's all kinds of bad things. You could also say there's nothing new you can put on a canvas.‍

‍And it is really hard to put something new on a canvas that's a truly new, novel idea—that's not just kind of the same as this other thing somebody did twenty years ago.
‍‍

AI is no different. It's very hard and challenging to do something new with it. 

‍
‍With Diffuse Control, each time it's shown, it will have a number of different iterations that will continue to grow. When it launches, it will have three and we'll have more throughout this exhibition at the Shed. Then the next time, it will just keep growing more and more. They will all become distinct versions of the sculpture. So here is the first iteration. It was done with Hans Ulrich Obrist, who curated six images that form the basis of this iteration. Those images are fed into AI and form a new version of this artwork.‍

‍The idea is that this is a system for creating more and more artwork. Instead of the artwork being the output of AI, the artwork is the actual system that produces an infinite amount of artwork. It's the curation of that work of images that creates the work. It's really an analogy of curation as creation, which is a way that we use AI as well.

Also, this work was made possible using the open-source, publicly available models that were trained on billions of images.
‍‍

To me, the public models are actually much more interesting than training your own model, which I think is kind of like teaching a dog a trick. 

‍
‍Training your own model is like very narrowly defining a box for it to regurgitate exactly what you told it in a different way.
‍‍

But the publicly available models are more like, “What is my kid going to say?” I don't know what my kid's gonna say because they were trained on the entire world. So it really is more a statement on the entire scope of humanity.

‍
‍The images that are fed into this define how it's going to look—in ways I couldn't have possibly imagined. It really is this ongoing conversation, utilizing AI more to its full potential—to be this infinite creation machine versus just a very select output.
‍

‍

Peter Bauman: These public models attract copious criticism, often precisely for the automation-run-amok perception. Critique is clearly central to your work as well. How do you balance AI’s critiques beyond its creative limitations, which you touched on—from ethical concerns, including legal (copyright), data sourcing, ecological (resource intensity), economic (job loss) and socio-political—while still working within these systems? Are these concerns on your mind and to what extent do they impact what you're trying to achieve?

Beeple: I honestly think about all those things; the debates around all of them are very interesting. Let's go one by one. In terms of copyright issues, I think that part is interesting. I think people should be able to opt out. But at the end of the day, I hate to say it, if you opt out of it, it's not going to make any difference. It's sort of: That ship has sailed.
‍‍

There are publicly available models that are trained on the entirety of the Internet. Having those available for everybody to use is the most ethical thing.

‍
‍Telling a kid in the developing world it's not ethical for him to use this thing that could potentially bring his family out of poverty—I'm not sure that's the most ethical take either. This idea that the world could benefit from us all throwing our hat into the ring is not new.
‍‍

Google Images literally scrapes the entire Internet. And it doesn't show a latent idea of your image. It shows your image and then it sells ads against it. And nobody was like, “Oh, my God, what are they doing?”

‍
‍Everybody was like, “Wow, this is actually super useful.” So I think a lot of these arguments are based on protectionism in terms of being scared of this technology taking people's jobs, which I can totally see, because I do think that is the case. Those are very valid fears. Coming from a commercial art background and having many friends still in commercial art, I very much empathize with those fears. But we are where we are. So the idea that this is going to be litigated away is, to me, complete nonsense.‍

‍In terms of the power consumption of it. Look at traditional methods of things. When Pixar would render a movie, it'd take 24 hours for one frame—and 30 frames a second. These things used wild amounts of power. So that's not new. But I do believe this could get us to a point, technologically, that it could solve a lot of these problems. So I'm hopeful if we keep pushing forward, it could. I'm not saying it definitely will; I can think of many very bad scenarios moving forward.‍

‍To me, the more immediate concern is how fast this technology is progressing and what could happen from that.
‍‍

We will be gods fifty years from now or we'll be dead. An alien force is landing on the world—or we're building an alien force in the world. 

‍‍
‍Peter Bauman: A couple times now, you've mentioned how AI image making enables you to create things you haven't seen before, which you've called “the heart of what I do.” And you've also referred to it as "the role of the artist."

Is that the most fundamental connection between your various projects? Where might commentary fit? How has that connection evolved? Because you said at first your interests were more abstract, then they became more narrative. But they also became more critical. So how is commentary tied into the centrality of novel imagery?

Beeple: That's a good question. I was always just innately drawn to seeing something and doing something that is completely new and pushing the boundaries of what art can say—how art can look. I don't know that that was necessarily a conscious decision but I think, to me, the art that has stood the test of time were new things. At first it was like, “No, that's not art.” Then it was like, “Oh, yeah, okay, I guess that's art.” Those are the touchstone pieces that we're still talking about a hundred years later, whatever. I've always been drawn to that.‍

‍At the beginning of my career—or maybe most of my career up until the NFT thing—it was on a much more aesthetic level of like, “Okay, where can I see something aesthetically I've never seen before?” That's what has really drawn me to digital art because as the technology progresses, it allows new techniques and new workflows that can produce images that you've never seen on a canvas.

It's very hard to do something truly new and novel that somebody hasn't done before. Personally, when I look at, I hate to say it, many contemporary paintings, I'm kind of like, “Yeah, but isn't that kind of like this thing—with just a slight tweak—that somebody did a hundred years before?” It's really much harder, I think, to do that in certain mediums.‍

‍Since the NFT thing, I've looked at the work that I'm doing in a broader context—through the framework of the canon of art history. Before, I was not thinking about how my work might connect to the rest of art history.‍

‍I went to school for computer science—not art school—so I knew very, very little about art history prior to this. I still don’t know a ton but certainly more than I did four years ago.‍

‍Now I'm thinking a bit more high-level conceptually: How can I say something or do something new with this technology? That's where AI really facilitates new interactions. It allows me to go to a curator and say, “Give me six images and it will form a new basis for this artwork.”‍

‍I'm giving up control. I didn't know what images people were going to pick. I'm ceding a huge amount of control over what the artwork even looks like.

Peter Bauman: You mentioned that since “the NFT thing,” you started thinking about connecting to the art historical canon. It also seems to have spurred a conversation about control and chance. This was highlighted explicitly by Patrick Moore, past director of The Andy Warhol Museum.

He wrote for SXSW London “From Warhol to Beeple,” in which he points out how “the connections between Beeple and Warhol are myriad.” How do you feel about that comparison? What does Warhol’s legacy mean to you?

Beeple: That's obviously a very kind comparison—a big honor. There are definitely, in many ways, unwitting parallels there. I just didn't know anything about Warhol prior to about five years ago. Now I obviously know more about it. His work is super interesting and I do see many strings that I'm pulling on that are very similar in terms of pop culture. But it’s also trying to use the most advanced tools. That was something he really did, including screen printing and Polaroids.‍

‍I feel very fortunate and inspired by the tools available today and the just infinite amount of possibilities from them, especially the speed at which they're like, “Oh, now you can do this and I can do this and I can do this.” I have no way of knowing but I'm assuming that this is something that Warhol would have been very inspired to be able to create with.
‍

‍

Peter Bauman: He was certainly interested in using new techniques for image mass production. You mentioned your learning experience with art history started after you reached a certain level of commercial success four or five years ago. Can you talk more about that?

Beeple: Yeah, it was because of that. After that experience, I started talking to people that were completely apart from the people I was talking to before.
‍‍

This bridge started forming more with the online digital world and the traditional world. So I gravitated in some ways more to the traditional world because I was also very new to crypto. I'd only learned of NFTs like four months before the big sale. 

‍
‍I was pretty acutely aware that there were many people in the crypto space that did not give any f– –k about art and that this was just the thing to gamble on. They would move on to the next shiny new thing to gamble on as soon as that presented itself.
‍

In contrast, the traditional art people—even though we didn’t agree on many things, let's say—were people who I could tell had spent a lifetime thinking about art as I had.

‍
‍I found that there were a lot of interesting conversations there. Then I saw a bit more of a lineage to what I was doing and what this moment represented in terms of new art movements—again, not at first considered art. I could kind of see that beginning to play out. Then it became much more interesting.‍

‍Also for a long time, I purposely did not want to know anything about art history because I was very scared that if I knew art history, I would just start subconsciously following all these rules.
‍‍

Now I have a bit more, I guess, confidence in my work that I won't do that. Now I want to know what's happened in the past so that I don't inadvertently just do the same sort of thing and then think, “Oh my God, this crazy thing nobody's done before!”

Peter Bauman: It does seem like you consciously picked up certain rules as you went along. You mentioned how part of the experience with Everydays was that you learned about composition and color.

Some curators I’ve spoken to insist that artists need to “speak the language” of contemporary art—that success depends on playing the institutional game. Your rise seems to challenge that. Are you an exception or do you see yourself modeling a new kind of independent path?

Beeple: I know what you're talking about and I completely agree. It's very true that before you had to, like you said, speak this certain language. Even broader than that, it's like you had to be ordained by certain gatekeepers in the traditional art world. From there, those people would disseminate the information that this person is worthy of your attention. Then the masses would just be like, “Okay,” because there was no other choice; they didn't know of anything else. These are the people that controlled the ways of seeing art through museums and books and whatever.

In the NFT moment, there was a lot of attention paid to the technology of NFTs. What was a little bit missed was the fact that I had amassed millions of followers prior to the NFT thing without any institutional validation in any way from the traditional art world.‍

‍A big piece of why this happened to me in particular is because I had more followers than any other digital artist. I was kind of the most popular digital artist. So it actually makes much more sense that that's why this happened specifically with me. I used social media to convince millions of people of the value of my work—millions of people who don't have a bunch of power. They're not sitting on the board of whatever; they're not editor of this art newspaper or whatever.‍

‍They're just an average person who's like, “Hey, this is interesting. I'm going to share with my 30 people.” Growing organically that way is something I think you will see more in the future. I was just kind of the first instance of this because it coincided with this new technology that blew up and allowed people outside the current system to value art. It revealed this other system, this more democratized system for finding new artists—artists who are doing things that people find interesting. They were able to validate that by spending a lot of money.
‍

‍‍

‍That doesn't mean that the museums are going to go away or XYZ gatekeepers and curators are going to go away—or their value is even diminished. It's just another path; both can coexist.
‍‍

I do think you will see more people come up the way I did in terms of convincing a massive amount of people of the value of their work versus convincing fifty people at the top of the art world of the value of their work.

Peter Bauman: What you’re saying reminds me of other art historical impulses to democratize art that go back quite far but particularly to the 1960s with groups like GRAV, who were literally bringing exhibitions onto the streets of Paris—intentionally taking them out of museums so that they could reach people directly.

Beeple: I look at social media as more akin to public art than a lot of people recognize. People can be like, “Oh, social media, that's where you see blah, blah, blah and everybody gets angry and whatever.” And it’s definitely all those things.‍

‍But it's also a place where you can release art to the entire world publicly for free—for anybody to engage with in a dialogue, too.
‍‍

Reframing social media as public art, I think, is something that you will see people do over time as they realize that a bronze sculpture sitting in a park is something that everybody can come up to, look at and experience. What is a social media post? It's, like, the exact same sort of thing.

Peter Bauman: This growing prominence of public art—that digital tools, like social media, have enabled—seems aligned with the institution your studio is modeling. I spoke to Dmitri Cherniak and he was really impressed by how seriously you're engaging with the potential of museums in a more digital age. To what extent do you see your studio as a prototype for a new kind of institution? And how does that relate to the rise in prominence of public art?

Beeple: What we're doing with the studio here, I look at it as a sort of incubator—or sort of like a prototype for ideas of what museums could add to the experience. I'm really very much about adding things as opposed to being like, “Burn this f– –ing system down or whatever.”

We have an experiential space in the studio here with wraparound projections. But the thing that's different about it is we have couches and tables. There's a bunch of seating.
‍‍

It's really meant to be a space where you come in and see many different types of things and spend hours. 

‍
‍Whereas, mostly when you see an experiential space, it looks a certain way, you kind of stand there for 15 minutes, and then you're like, “Okay, cool, I got it.” And then you leave.‍

‍It looks that way for months or sometimes years at a time. To me, that's a very underutilized way of looking at what that medium can be in terms of a digital space that can change instantly. And they can show you many different moods and environments—messages. We've really tried to experiment with different ways of utilizing a space like that—ways of also not just exhibiting things but also curating things too.‍

‍We're in a sort of a small city here. We're not in New York or LA so we've really been focused on events and bringing people together at specific times. But for each of those events, we put out a call for art and if it's roughly on theme and not complete butt trash, then we will show it.‍

‍So we basically curate a group show with hundreds of artists in two weeks. That's something that is very fundamentally different than the way museums operate. But with museums wanting to be more inclusive and engage a diversity of voices, I think it's a way of programming that you will see them adopt.‍

‍It's something that digital art allows that traditional art doesn't. If you're doing a group show with hundreds of artists all over the world, you're talking about years. And it will cost an immense amount with traditional art. You're shipping this and insurance and who's the handler and how are we gonna light this and where are we gonna put this wall thing?
‍‍

It could just be like, "Send me the JPEG. Send me your handle.” And if you have the system set up to ingest and show these things quickly, you can have a different type of curation. 

‍
‍Again, not to say anything against how museums operate with more criticality—the education and the contextualization that they do around works. But I think there's a different way to look at it so they can also provide a way to reach the multitude of voices from the community that they serve.

Peter Bauman: These museums have to worry about their brand, which does serve a purpose.

Beeple: There's nothing wrong with that but if you're showing somebody's artwork for a very small amount of time, which is facilitated by digital art, you're going to be in the museum a total of ten minutes. You can't do that with a painting. Physically, that doesn't make sense. Are you going to show somebody's painting for ten minutes and then swap it out?‍

‍It doesn't have to be this overwhelming endorsement of this person and, “Oh my god, this is the biggest thing.” It's like, this is a thing; it’s kind of interesting. It's ten minutes interesting. But to a lot of people, that’s a massive thing.

Peter Bauman: You're advocating for a more participative institutional model. But you are also interested in participation in your practice. Works like Human One, Tree of Knowledge and Diffuse Control all emphasize audience participation. Why is that important to you, especially given that increased participation aligns with your ideas on institutional reform?

Beeple: That's a good question. Having artwork be interactive is more about utilizing the medium of digital art to its full potential. When you have static digital art, it’s not good or bad, but I anticipate more art being dynamic in the future. Art will be something that feels very alive and is reacting to you and things that happened, as well as your input.

‍‍Finding artful ways to interact with work is an interesting challenge and something that goes beyond what I like to call “hand-waving art,” where you walk up and see that it's some sort of filter or something, you wave your hands, and it's like, “Oh, look, there's me.” And that's it. That to me is kind of a shallow interaction that I don't think says a whole lot about you or the human experience.

Looking at different ways that you can facilitate types of interactions really showcases the full potential of digital art. With Diffuse Control, the audience can slide and blend between those six different images in real time. People can collectively engage with the piece at the same time.

If two people do use it at the same time, trying to control it, it starts to go into this glitch mode. Then, if a bunch of people are using it, it goes into this collab mode. They can work together to show a different piece of the state. They can also generate prints from the artwork as they're moving it. They can kind of like take a snapshot and make a print. There are a lot of different ways that you can make artwork interactive.

Peter Bauman: Participation in both cases seems motivated to showcase the full potential of digital art—the things that excite you most about its future and potential.

Beeple: I think we're just at the absolute beginning of what's possible there.

‍
‍
----
‍


Mike Winkelmann (Beeple) is an artist and graphic designer from Charleston, SC, USA who does a variety of digital artwork including short films, Creative Commons VJ loops, everydays and VR / AR work.  
‍
‍‍Peter Bauman (Monk Antony) is Le Random's editor in chief.