Most people treat “outlier” like a performance upgrade.
Work harder. Stack habits. Follow the proven playbook better than the next guy.
That produces high performers inside the existing system.
It rarely produces people who rewrite the system.
To be an outlier, you have to literally be different.
Not “a little unique.”
Not quirky on weekends.
Not meditating while still optimizing for the exact same metrics as everyone else.
Literally different in the ways that actually move the distribution.
I developed seven algorithms.
Not seven pictures. Seven sets of rules, each one a different procedure for building a form out of nothing, and each one answering the same question a different way: how does a thing come apart and put itself back together?
There are no image files anywhere. I did not draw 3,333 pieces. I wrote the things that draw them, and then I stopped deciding.
Here is why that is harder than it sounds.
If the artwork has to live entirely on chain you get a few kilobytes. Total. For geometry, lighting, colour, all of it. That rules out meshes, textures, and every renderer anyone actually uses, which is why almost all generative art is 2D. It isn't a taste thing.
So: signed distance fields, raymarched per pixel. Shapes blended into a single surface with a smooth minimum. Real depth, real self occlusion, true surface normals. The code is the artwork and the artwork fits.
The renderer was not the hard part.
Legibility was. A round shape can only ever be a bead, so round shapes arranged in a circle read as beads in a circle no matter what the description claims. Elongating every element into an ellipsoid is what made the forms finally read. Proportion mattered more than correctness: one of the seven was right in every measurement and still read as the wrong thing entirely until I changed nothing but the proportions.
Speed mattered too. When the art is drawn on demand instead of served as a file, a sixteen second render is not slow, it is unusable. One bounding sphere test took it to about a second.
And every version of it gets swept across fifty thousand hashes before it goes anywhere. Every single one has to produce something.
Give it a number, it draws a form. Give it the same number in a hundred years on hardware that does not exist yet, and it draws the same one. Same curve, same colour, same place.
Seven algorithms. 3,333 outcomes.
The work is not a picture of the process.
It is the process.
Website is coming soon.
There is going to be a way to get WL, first 100 people to uncover it then solve it.
You won't need to connect your wallet.
Official website will be on the official X, mint will be announced before. So don't click random links.
Stay safe.
COLLABS ARE OFFICIALLY OPEN 🦫💚
The wall is down
Looking for projects and communities ready to build with Robin Bandits
Let’s take over Robinhood together
Retweet this and tag your collab managers & communities below 👇
WL Submissions close today at 18:00 UTC ⏰
We’ve received 40,000+ applications so far.
We know some of them are bots, so the team will carefully filter submissions and select eligible wallets for the final WL.
Next post: WL Checker goes live.
🚨 WL SPOTS ARE NOW OPEN! 🚨
The community waitlist is officially LIVE! 🔥
👉 Join the Waitlist :
https://t.co/CNONT7l82p
You can now submit your entry for a chance to secure a WL spot for the upcoming mint.
The Hollow mint is coming to @RobinhoodApp
Good luck everyone! 🖤