About Kueizen

Our Vision

Founded in 2020. Kueizen's impossibly ambitious overarching goal is to build an O'Neill Cylinder—a monolithic space mega-structure.

This is more than a moonshot. However, it aligns our horizons, values and roadmaps.

Four Years in Stealth R&D

Initially bootstrapped and self-funded. Kueizen sprung from independent research of an early (2018) physical simulation of an O'Neill Cylinder by Omar Bessa.

This early simulation showed many of the different sub-systems that would have to be built in order to make it work. Hydroponics, On-Site Assembly, Robotics Manufacturing, and so on.

During the pandemic, Kueizen's team grew (and then shrunk) considerably. Our main focus back then was to use the Mothership framework to design/optimize vertical farms and other hydroponics systems.

For this purpose, an engineering sandbox with Computational Fluid Dynamics simulations was built. Mothership's usage converged on an optimized hydroponic modular tower system.

The need for a swarm of autonomous robots

Since it is very clear that we're far off from the technological readiness necessary to build an O'Neill Cylinder as it is, the strategy then became to figure out a roadmap of the technology that would have to be deployed/developed in order to climb the steps until that technological readiness.

Currently, the most plausible way to build one of these structures is to hollow out an asteroid and build it in-site.

Humans will not be involved in this process. For this we need autonomous agents that can thrive in a dynamic environment with a very clearly defined goal: the GOLEMs.

Early GOLEM Development

The initial prototypes had some limited agency, and were not even named GOLEM at the time, these were Artificial Life prototypes with early Generative AI capabilities.

It was, and still is, a very computationally heavy process; as the Mothership Framework is extremely resource intensive.

Merging the GOLEM with Language Models

We got very early access to GPT-3, which we evaluated for integration right away.

However, it wasn't very clear at that time how that integration would look like, but later on—after plenty of experimentation—we managed to develop a prototype.

As of right now, as we continue to build this tech, many different types of products become possible with this new stack as it develops.

Enter K1V4

A new type of AI Assistant that can run locally on consumer hardware. With the new architectural insights from developing the GOLEM, we can now create this fast consumer product.

As unrelated as it might sound, we envision that the on-going development of these digital waifus will lead us to victory in the colonization of space by humankind.

We hope to dearly contribute to that goal.