Worlds, Not Just Walls

A few weeks ago I had a great video call with John Crain, one of the founders of SuperRare. It was two builders talking. By the end of it he handed me a key: early access to a new tool, then called the SuperRare agentic beta. The idea was simple and fluid, and a little unbelievable. Describe an app, and an AI agent helps you build it and deploy it to a live, shareable link in minutes.

That tool has been renamed twice since I started writing this. It was the agentic beta, then PlayFactory, and now it is SuperRare Studio. I am leaving that in rather than tidying it away, because it tells you exactly what stage this is. The thing is still becoming itself. So is everything below.

I need to say where I am standing before I go any further. I have been an artist on SuperRare since April of 2019, and a fan of it for longer than that. A fan of what it is, what it has been, and what it keeps becoming, which is a place that trusts artists and points them straight at the frontier of what comes next. That trust is not a small thing to me. So when I was handed these tools and told, in effect, to go build, it did not land as access. It landed as an honor. And it gave me back a word I do not use lightly about my own work. It let me play.

I am an artist. I have been making visual work for more than forty years. I am also Deaf, and Hard of Hearing, and visual language has always been my first language. What I am not, by training, a developer. And yet here was a machine offering to close that gap. So I did the only thing that made sense to me. I started building, and I decided to build it all in the open.

This post is the record of that, and I keep it current rather than writing a new one each time. When I first published it there were ten apps. There are now sixteen. What changed in between is not just the count, and that is the part worth reading.

Why this matters to me:

For a long time I have believed two things that used to sound like metaphors and now sound like a plan. The first is that my archive, all of that work, is not a pile of old things. It is material. It is data. The second is that AI is not a novelty to bolt onto art. It is infrastructure, the way electricity is infrastructure. Put those two beliefs next to each other and you get a question worth spending a year on. What happens when forty years of your own visual language becomes something you can walk into, play inside, and hand to someone else?

This is the same question I have been circling under the name AREMES Enterprises, the world I have been building where my archive is treated as living, licensable, intelligent material. Studio let me start answering that question with my hands.

The turn I did not plan:

The first ten things I made were rooms. Places to walk into, built out of my own work, for people to move through. I would have told you at the time that this was the whole project.

Then something shifted. Somewhere around the sixth or seventh build I stopped hitting problems that were mine and started hitting problems that were everybody's. A zip that deploys fine on my machine and arrives broken. An archive folder full of the same painting at four different sizes. An animated GIF that quietly stops animating the moment anything resizes it. None of that is art. All of it stands between an artist and the thing they are trying to show.

So the second half of this list is different in kind from the first. The rooms are for looking. The tools are for building, and they are for you, not for me. Every one of them is free, every one is a single HTML file, and every one exists because it broke on me first.

The rooms:

Places built out of my own work. Click any image to walk in.

Now I Make Things With Machines

Where it started. A single layered room, a painting on the wall, animated figures, and a 3D sculpture you can turn in your hands. My work, stacked into one space you can move through.

SPATIAL

A walkable gallery engine pointed at my own SuperRare collection. A first-person corridor you move through with your keyboard, where clicking a work flies you to it and opens the link to collect it. It even opens in VR on a headset. These do not have to be my rooms. Any artist's collection could become a space you enter.

The 3D Collection

A white-cube museum of my free-standing 3D sculptures. Walk in, move among them, and see them the way they were meant to be seen, in the round. This set the tone for the next iteration of it below.

Walk Between the Wings

A multi-wing museum. You step through glowing doorways and fade between three separate rooms, one for the sculptures, one for motion work, one for works on paper. The motion wing is alive, its walls animate as you stand in front of them.

Graff-Entity: Spatial Relic No. 2

One of my minted 1/1 sculptures turned into a proper drop page. The piece floats in a museum void, turning as you drag it, with a button to view it in augmented reality on your phone and a button to collect it on SuperRare. The 3D and the AR are generated from a single model at runtime. No app to download. Just a link.

A Front Door for the Machines

An art object that is also a live commerce terminal. Press a button and it makes a real request to an endpoint I run, and the endpoint answers the way a machine answers another machine, with a signed offer to pay in stablecoin. Soon the visitors to our work will not all be human, and we should build a door they can knock on.

Seslow Remix Studio

A browser art studio modeled on my own drawing system. Compose with my transparent stamps, layer them, warp them, and export the result. A tool, not a pitch. Make something in my visual language and take it with you.

Field Notes

A reading list of my recent writing, so other artists could find their own way in. No gatekeeping. If it helps you build your own thing, that is the whole point.

Creation Syntax

A playable reality engine. A generative field driven by a little equation from my AREMES world, where four sliders are the terms of the formula and you author the image by moving them. Silly, yes, fun, yes.

The Tesseract Garden

A 3D garden you can play in. A first-person game where you gather the blooms before the timer stops you. It deserves its own telling, further down, because it is where all of this was pointing.

The Tesseract Garden II

A second generation of blooms in the same garden. The first of these to get a sequel instead of a revision, which is what happens once the world exists and filling it becomes the easy part.

Ryan Seslow's SuperRare Art Gallery

A 1998 Netscape window holding my available 1/1s, with live prices, where every tile links back to SuperRare to collect. I grew up on that internet, the one that took a minute to load and felt like a place. Making a storefront that looks like it came out of a dial-up modem and then having it actually transact is the whole reason it exists.

The tools:

These came out of getting things wrong in public. Take any of them, change them, put your own name on them.

Wallet to Wall

Paste any SuperRare wallet address, pick one of four designs, and that artist's whole minted collection hangs as a live gallery. It reads the API every time it opens, so nothing goes stale when a piece sells. Point it at your own address, not mine. The design and the address both travel in the link, so the URL you copy is the gallery.

Studio Doctor

Drop your app zip on the page and it tells you what is going to break before you deploy it. It reads the zip in your own browser, nothing uploads anywhere, and it draws your whole app as a picture: every file a block sized by its real weight, your entry point lit up, and anything your HTML asks for that is not in the zip drawn as a ghost. Every check in it came out of something that broke on me first.

Archive Packager

Point it at a folder of your work and it hands back a gallery you can deploy, under the size limit, with the duplicates gone and your animation still moving. It checks the four things that quietly wrecked my own archive galleries: WordPress size variants, animated GIFs being flattened by a resize, the same piece arriving under three different filenames, and packaging work you have not actually looked at yet.

Archive to Agent

Watch a machine that has never heard of me find my archive, choose a work, and ask what it costs. It gets a real price in USDC on a real blockchain, addressed to my real wallet, live from your own browser while you watch. It stops at the price and nothing is bought. Then it writes the same five files for your own site, so your archive can answer a machine the same way.

The Machine Room

All sixteen in one page. Click any of them and the real deployed app runs right there, live, rather than showing you a picture of it. The page holds a list rather than the apps themselves, so it stays small no matter how many there are, and there is a button at the bottom that hands you the whole thing to point at your own work.

All sixteen in one place:

Sixteen separate links is not a body of work, it is a list. So the last thing I built is the room they all live in. The Machine Room is one page holding every app, and clicking any of them runs the real deployed thing right there in the page rather than showing you a picture of it.

It is deliberately a shell and not a container. All sixteen apps together come to about forty five megabytes, and the deploy ceiling is fifty, so packaging them into one file would fit today and break the moment I make a seventeenth. Instead the page holds a list of titles and links. It stays the same small size forever, and each app keeps being updated on its own.

The list is the only part you would need to change to point it at your own work, so I made the page hand itself to you. There is a button at the bottom of it that downloads the whole thing.

What is actually under the hood:

I want to be transparent about the tech, because the whole point is to show the process, not just the result.

Every one of these is a single, self-contained web page. Everything lives inside one file: the 3D engine, the loaders, and the artwork itself, encoded right into the page. You open a link and it runs. No install, no accounts, no server of mine to keep alive. Ten of the sixteen are one HTML file with nothing beside it, which is also why none of this is trapped anywhere. If the platform vanished tomorrow I could put all of it back up somewhere else in an afternoon.

For the real-time 3D I used Three.js. For augmented reality I leaned on the browser's own model viewer, which can take a single 3D model and generate the AR experience for both iPhone and Android on the fly.

The hardest part, honestly, was my own sculptures. My 3D works are big, detailed files, tens of megabytes each, and the platform has a sensible ceiling on how much you can deploy. So I built a compression pipeline that shrinks the geometry and the textures without you being able to tell. For the Tesseract Garden, five of my 3D characters started at roughly sixty-four megabytes and came down to about three and a half, and they look the same. That step, boring as it sounds, is what made the whole thing shippable.

Underneath the commerce piece is a small open standard for machines to pay each other, and for the walkable galleries there is real WebXR so they open in a headset. But the thread through all of it is simpler than any of these words. I described what I wanted, an AI agent helped me build it, and a command-line tool pushed it live. An artist was handed developer reach, and used it to make his own work walkable, playable, and alive.

The machines really do show up:

When I built A Front Door for the Machines it was a demonstration. Press a button, watch a server answer the way a server answers another machine, with a signed offer to pay in stablecoin. It was true, and it was still theatre, because nothing was on the other side of it.

That is no longer the case. In March an agent I did not write, running on somebody else's machine, found my archive on its own, read what was in it, chose a product, paid forty nine dollars in USDC, and received the file. No checkout page, no e-mail, nobody watching it happen. I found out because there was an order in my shop. I wrote about it the next morning because I could not quite believe it had worked.

Archive to Agent, the newest thing on this list, is that made visible. You press a button and watch a machine that has never heard of me find my catalog, choose a work, and ask what it costs. It gets a real price, in real dollars of USDC, addressed to my real wallet, and then it stops there on purpose. Nothing is bought and nothing is charged. The interesting moment is the price, not the payment.

And then it hands you the files to make your own archive answer the same way. Five small files and a note saying where each one goes. That part matters more to me than the demo does. Getting found and getting priced by a machine is the hard part, almost nobody has done it, and it is safe to do. Taking custody of somebody's money is not something I am going to hand a stranger in a zip file, so the kit stops exactly where the demo stops.

How the Tesseract Garden came to be:

Here is the part I most want to be honest about, because the making of this game is a small model of how all of it works now.

The characters came first. Years before there was any game, there were the Tesseracts, the entities that live in the Tesseract Garden. On my site I describe that garden as a living laboratory where consciousness, memory, and code are cultivated, an architecture built out of compressed memory, where the entities you meet are not static sculptures but active data relics, pulsing with the signatures of their original analog and digital origins. I had made them as art. I had built them in 3D, minted some of them, and shown them in augmented reality. They already existed. The game did not invent a single character. It gave the ones I already had a place to be.

It did not start in 3D. It started flat. My first version was a small two dimensional game, a one-touch climber built from the flat, cut-out faces of these characters, pulled straight out of an old illustration file in my archive. It was genuinely fun. But when I played it I kept asking myself a simple question that would not leave me alone: what is the goal? And a second one right behind it: why is this flat, when these beings were born dimensional?

So it moved. First the world went 3D while the characters were still flat cut-outs standing in it. Then came the realization that changed everything, which is that the real 3D versions of these characters already existed on my own machine, the same models I had exhibited and minted. I did not need to make them. I needed to find them, compress them so they would fit, and let them stand up in the garden as themselves.

Even then it was not right. In the first 3D version you looked down at your own character from above, floating over the garden like a drone. It did not feel like being there. So I put you inside it. Now it is first person. You are standing on the ground among the Tesseracts, at their eye level, and you gather them one by one. Every one of those turns, the flat climber, the 3D world, the real models, the step down onto the surface, happened in the open, mostly because I played my own game, noticed what was wrong, and said so out loud. That is the whole method. Build, share, react, build again.

Then I did it again. The Tesseract Garden II is a second generation of blooms in the same world, which is the first time one of these has had a sequel rather than a revision. It turns out that once the garden exists, filling it is the easy part.

It is worth saying plainly what this game is, then. It is the first time anyone can walk into the Tesseract Garden. For years it was an idea and an exhibition. Now it is a place you can enter and move through, where the entities are not on a wall but around you.

What breaking things in public is actually for:

Building in the open sounds like a posture until something goes wrong and you have to decide whether to say so. Most of what I learned here I learned by shipping something broken and then working out why.

A gallery I was proud of went out with every animation flattened into a still, because resizing an animated GIF quietly re-encodes it to a single frame. An archive export gave me the same painting four times at four sizes and I published all four. I put a price on an app and the app went dark, because a paid listing is not serviceable yet and the tool reports success anyway. Every one of those is now a check inside Studio Doctor or Archive Packager, which is the only useful thing to do with a mistake.

I have passed each of those back to the people building the platform as I found them, with the controls I ran, because a bug report with evidence is worth more than an opinion. That is the part of this I would recommend to any artist who gets early access to anything. You are not there to be impressed. You are there to be useful, and finding the sharp edges is useful.

Where this is going?

The Tesseract Garden is not the end of anything. It is a doorway.

Right now it is a first-person game where you gather the blooms before the timer stops you. It is joyful, and it is the most alive thing I have made in a while. And the blooms are real works. I built them as 3D characters long before there was a game. They are not minted yet, and that is the plan. I want the game to become the path to them, so that someone who falls for one while playing can one day go and collect the actual piece. You can already find my minted work on SuperRare, where this whole chapter started.

From there it grows. I want the act of gathering to compose a small story, different for every player, so the game becomes a way to author a myth out of my characters. I want the blooms you gather to join the larger world of AREMES, my Citizens. And I want the garden to live on aremes-enterprises.com itself, so the page that describes the Tesseract Garden and the page you can walk into become one and the same. On my site I have long called that garden a living archive and a feedback loop between perception and simulation. The game is the first working piece of it.

None of these are for sale yet, and I want to be straight about why. Selling on Studio is not switched on. An app with a price on it is unreachable until the payment layer ships, so everything here is free, and free is the honest state rather than a gesture. When that changes I will say so here.

That is the thesis, said plainly. Forty years of a visual language, turning into living, playable, collectible, intelligent work. Built in the open, in real time, with a machine as a collaborator and my own archive as the material. And where it can be, handed to you as something you can run yourself.

One more thing, and it matters most. Thank you to John Crain, for the time, interest, support and call that started all of this and the key he handed me. And thank you to Jonathan Perkins and Jonathan Carmel, for the time, support and the connection every step of the way. Building is better when the people you respect are rooting for you to make the thing.

None of it is finished, and that is exactly why I am sharing it now. If you want to watch it come together, that is the invitation. I will keep showing every step.

Ryan Seslow · ryanseslow.com