ALCI is our digital art collection that uses Bittensor and generative art algorithms to output 3D models. The pieces represent the evolution of shape and form. When we think of traditional evolution, a few things come to mind. For example, a modern mammal was once ocean-dwelling, like some fish, and then evolved to become more mammalian, then lived on land, and eventually branched into the mammals we have today. The whale started as a land animal and, through sea-bound adaptations, became the ocean-dwelling beast we know today.

But this isn't purely a biological concept. We can understand the evolution of products and technology in a similar way. The first "computer" was the abacus, then the Greek Antikythera mechanism, and then ENIAC which gave us something more similar to modern computers. We then rapidly improved upon these with generations of minicomputers and microcomputers, then the first PCs, and eventually smartphones, which have undergone countless iterations since.

This is how we typically imagine evolution. As each new iteration forms a branch, some form of selective breeding, Darwinism, or technological progress spurs evolution. And some branches inevitably die off.

With Alci, we apply this idea of evolution and branching pathways to 3D objects. What is the evolution of shape, or better yet, what is the devolution?

If we take the human being as an example, from the outside, we appear as a cohesive form. This can be broken down into a series of connected rectangles, triangles, and other polygons. As this form devolves, we become a more uniform shape with fewer points and fewer sides. Then, we reduce to a singular line and eventually a single point.

Yet at any one of those stages of devolution, would we still be able to infer that this was once a human? As more and more form is taken away, we're left with a starting set of points from which countless other forms could emerge. This parallels evolutionary branches; every organism comes from a shared ancestor, and that lineage persists through modern expressions.

Every square, hexagon, pentagon, and so on shares an ancestor in the triangle, and the triangle shares an ancestor with them in the line, and the line shares an ancestor with them in a singular point. The point, or maybe even space in general, is the ancestor to all shapes and forms. With every point added, assuming some entropy and randomness, we get a new shape every time.

Tibetan Monks spend hours drawing beautiful mandalas from sand and then sweep them away as they finish. They do this to symbolize impermanence and detachment. It's all the same sand they reuse for every new mandala, something akin to these ancestors/branches. You could recreate an old mandala, but through entropy and randomness, the probability becomes 0, and new final Mandalas(outputs) are formed every time, all with the same sand.

This diagram is a little shoddy, but you can hopefully understand how the concept scales once we reach thousands of points and in 3 dimensions. As points are added, we result in new shapes of different sizes and dimensions.

Our generative art process is as follows:
We first generate 3D objects using Bittensor's Subnet 17 (SN17). These objects represent real-world entities while incorporating variations in form. We have 50 subsections and an arbitrary number (from 1 to 6) of iterations within each subsection.

For example, a subsection could be "landforms" and include: mountain, volcano, glacier, canyon, etc. Other subsections might have less unique iterations but still represent aspects of physical reality: food, animals, architecture, technology, daily objects, and more.

Coming from SN17, these objects are now 3D models in point cloud format, which is easily manipulated. Each object iteration is passed through four different algorithms. We've written these algorithms to remove and add points in various ways. Point Clouds that are generated with SN17 are generated with a method called Gaussian Splatting. Through this, points carry a few properties like location, color, opacity, scale, rotation, etc. For understanding our evolution concept, all you really need is location, and the concept was ideated from the perspective of a point cloud.

Our art collection displays a snapshot of SN17's outputs, and we hope to see the outputs evolve over time. Holders will be able to see the original SN17 output & prompt that resulted in their Alci object.

We believe each algorithm embodies a distinct mode of thinking or philosophical concept. These are abstract ideas, but that's the point: we're devolving existing 3D forms and then manipulating them to express thought.

Even if the final object is abstracted beyond recognition, the applied algorithm might allow someone to imagine how it came to be, regardless of its original form and without knowing which algorithm was used.

These algorithms might be difficult to visualize at first. Here's one example:
If we relate an algorithm to liberalism, how might we apply a Rawlsian max-min principle to a set of points? We might remove points from dense areas of the object and redistribute them to sparser areas. Or perhaps we introduce some anarchy and remove points altogether, such that "everyone" is worse off. We can have Nietzschean algorithms, Platonian, and so on; we can have algorithms that are completely unrelated to any defined philosophy.

Our first algorithm purely removes points and establishes connections between them, this is a form of anarchy in that we effectively "burn" the points that we remove, as they are never redistributed. Let the fire clear the field. Then we plant something real.

Our second algorithm builds on the first one, except we apply more world-building; we first redistribute points from high-density areas to low-density areas. In addition to this, we also mirror points and create semi-symmetrical structures.

The third algorithm represents a point in time when I was angry. This algorithm produces crude, raw, and rough outputs. The form is consistent with the original object, but the abstraction adds sharp edges and makes the objects uncomfortable.

The fourth algorithm tries to represent some abstraction of Nietzsche. There is randomness that fights any imposed structure, but with directional force, this can be seen as an almost Will to Power. There's a flair of Eternal Recurrence as echoes of the original form appear through scaled/transformed duplicates.

Hopefully, this brings color to how we write and apply these algorithms. They are expressions of ideological reasoning in spatial form.

The latent space of LLMs is an invisible, multidimensional landscape where meaning, context, and structure are embedded in mathematical form. In a metaphysical sense, latent space is the pre-linguistic ether of thought: an abstract aether where all forms of our language coexist, shaped not by grammar but by proximity and transformation. What we query and generate from an LLM is just a projection from this latent realm.

Our iterations give us another layer in our evolution concept:
How do our algorithms differ in their transformations across object iterations?
What different outputs emerge from the same base shape?
How similar are the outputs within a subsection?
These are the types of questions we're exploring through the collection. We believe a similar layer of latent space exists amongst these points we transform. Even as the actions of our algorithms are predetermined, a level of abstraction and randomness sits atop, allowing new structures to come forth.

Alci comes from Alcibiades. He was a student of Socrates, but can be seen as a rejection of him, not by argument, but by the sheer force of life lived differently. Where Socrates stood for reason, restraint, and the examined life, Alcibiades embodied passion, ambition, and the beauty of experience itself. If Socrates sought wisdom through discipline, Alcibiades reached for glory through intensity. He was not the quiet philosopher, but the storm on the horizon. Not the blueprint of virtue, but a living challenge to the idea that reason alone should guide us.

Alcibiades, like all great people, represents more than just his achievements and actions.
What remains is the example of a man's life.
It's about a man living a larger than life story of power and defiance, offering an alternative to the bleak mediocrity that defines the "middle management" of our world and their denuded vision of life. They are not just uninspiring but spiritually diminished: domesticated, petty, myopic.
Alcibiades struck life into those around him, he was not merely a man but a mirror of what life could be if one dared.
In a time when vision is punished and conformity rewarded, a life lived vividly, regardless of "achievements", becomes an act of rebellion and perhaps, a kind of salvation.

Alcibiades and other Greek figures live throughout our collection. We often leaned into Platonian elements of form and color, while also allowing Dionysian energy to seep in. To Plato, form serves as true beauty, it is the primary essence of a thing. It's rational, real, and independent of sensory attributes, while color is mutable and can be deceptive. Color is "a flame that flows from bodies", and we approached our art initially devoid of color through this reasoning. Yet, while sticking to this concept, we slowly introduced colors and came upon this dark palette that deeply resonated with us. It applies this motif that is underlying in the work: What remains in the physical world as we accelerate through the digital world? What kind of things must arise to inspire others? What structures must come forth?

And in that process, much like Alcibiades diverged from Socratic ideals, we moved beyond Plato, too, not abandoning his insights, but expanding them. We began exploring more emotion and passion through the algorithms themselves, allowing irrationality, intensity, and beauty to take form alongside structure. In doing so, we let the work live more vividly, guided not only by reason but by feeling.

In this matter, the truth is that we haven't seen anything yet.