Wall Street Dads’ Intelligent Design — A Technological Breakthrough In NFT Creation

It turns out creating 10,000 unique human beings holding hundreds of different objects, by combining 1200+ hand-crafted elements, across 25+ layers in a random yet controlled and coherent way is actually a very complex ask.

A Proprietary Visual Engine Built From The Ground Up

To handle the task, we built the Imagen engine from the ground up to simultaneously handle the complex combinations of rules, connections, exclusions, probabilities, and logic needed to make every WSD look seamless and amazing.

This makes it look easy

Watch the Hair!

For example a single dad has hundreds of different hair style options, spread across the chest hair, beards/mustaches/goatees, eyebrows, and the hair on his head ( if he’s lucky to still have some).

But like in real life, we play with our head’s hair color, but usually not the other stuff. So the algorithm makes sure that everything’s coordinated, except for those special hair colors.

The carpets match the drapes

Getting a Grip On Complexity

A unique feature of each WSD is the symbolic elements that they’re balancing, which adds to the variety, depth of each personality, and also the complexity of creation.

The fact we have different objects that require different grips means we needed to create separate grip types.

Because each hand holds a different object which require a specific grip type, each hand is generated independently from the other and from the body.

All Hands!

And if you look closely, every hand layer must coordinate to exist both behind and above the object that’s being held.

Way More than a Lasagna

In an ideal world, making a Wall Street Dad would be as easy as flattening sheets of paper over one another, like layers of an onion, or a nice lasagna. But because of how art elements work volumetrically within space ( like how an object is gripped both behind and above by the hand ), the same goes for just about everything else.

Hair and helmets, and glasses, and cigars, and parachutes, and bodysuits, they all interact in complex ways, needing to exist simultaneously both behind and in front of each other, in a mind-bending spatial puzzle.

The same goes for special background and foreground settings, which are coordinated to bookend the Z-axis of our NFTs in a delicious daddy sandwich.

Deal with it

In the example above, you’ll notice:

• The shades slip in front of and behind his hair, and his hair is both in front of and behind is face.
• The bucket of tendies and guitar are both held between a rear hand and fingers that grasp them from the front.
• The parachute suit has to be worn over his body, but the parachute itself must be behind the body.
• The plains environment is coordinated to exist both in background and foreground
• And this all happens after all the correct grips for the hands, skin colors and hair colors are coordinated among a dozen other layers that make up the body and face

Here’s another example:

• The shades, cigar, and knight helmet are intertwined with each other
• The lawnmower and cash-filled briefcase are both held between a rear hand and fingers that grasp them from the front.
• The swan floaty, even though it’s an outfit ( and outfits are layered below headgear ) builds the swan head as a separate head element that is above the knight helmet.
• The rocket interior and exterior is coordinated to exist both in background and foreground
• And again this all happens after hand grips, skin colors and hair colors have been generated and coordinated

And here’s an updated example with our upcoming Moms:

We’ve pushed the complexity even farther, with hair that interlocks with the body and facial elements. Hands, gloves, and nails coordinate and grasp around the elements, accessories like the hat and bag surround the figure as well.

Visual logic: Keeping track of it all

Which means Imagen must simultaneously:

A) Build matching layers above and behind each object for both hands independently
B) Match the grip type to correctly grasp the objects being held
C) Make sure all body parts share the same one of 13 possible skin colors
D) Make sure all hair layers share the same one of 8 possible hair colors but allow for special uncoordinated hairdos
E) Ensure that hundreds of linked elements appear correctly together if called upon to appear

And these are just the visual checks! We haven’t even gotten in to the programming of probabilities and rarities.

Here’s the nonsense that happens when we turn it off:

Chaos Ensues!!!

Check out our 0G (“Zero-G”) collection on OpenSea and see if you notice hundreds of other ways our elements are intelligently layered and sandwiched between one another. And how they all look like intelligible representations of people, unlike the above.

Probabilities and Rarities

At the micro level, we’ve covered the considerations that are taken on a image to image basis.

But we’re also generating a 10,000 set, and are controlling elements at a macro level so that the entire set as a whole has the distribution of attributes that we want.

While taking all the complexities of the visual logic above into account for each individual piece of art, the engine then factors in probabilities that we program into it for controlling rarities.

The team pre-plans the odds of existing for all elements to ensure a unique blend of common to rare to legendary traits while also ensuring that certain combinations of elements make the cut. And we run the generation until the perfect 10K set is created.

Imagen’s dreams: simulating a compilation of 10K set.

Beyond Unique

We do more than just a simple check to verify that there are no duplicates. Built into Imagen is a way to eliminate entire attribute types from counting towards uniqueness — traits like backgrounds, skin tones, hair colors, tattoos and certain facial features like noses.

In other words, you will never find any 2 WSDs that are identical except for a different environments, different skin, tats, and hair colors, or have a different nose. Those would all be recognized as duplicates in our book, even though they are technically different. Because to us, they are not different enough from each other.

No similarities. Not even close.

So that creates even further separation and uniqueness between all 10,000 Wall Street Dads, and the only way we can even self-impose such rule sets is because of the vast library of 1200+ unique assets that make up all the WSDs.

This is one of the reasons why WSDs, like real Dads, look so unique from each other.

Generating the Future

A significant amount of thought and up front work went into developing this engine and all the capabilities needed to accommodate such a wide, rigorous, and demanding combination of rules, logic, and probability.

It’s one of our key assets that we’re leveraging for building out all the members of our Wall Street Family collections in the future. And like all Daddy’s, it makes doing hard things look easy.

Check out Wall St Fam for more information on the project, follow us on Twitter and join our Discord community for the latest. You can also view the first 3,000 Wall Street Dads at OpenSea and let us know if you can spot the ways that our WSDs are intelligently created.