Aigarth is a project building decentralized Artificial General Intelligence (AGI) on the Qubic network by mimicking nature (evolution and brains) on a large network of regular everyday computers. The big idea of Aigarth is to create an AI that can learn and improve by itself, over time, without needing constant human input. It is designed to run on the Qubic Network, a decentralized network of computers. So, in essence, it is a journey to build a smarter and more independent AI by the people, not big corporations.
Who is behind Aigarth?
Aigarth is a project of the Qubic Network. It was conceived and started by Sergey Ivancheglo, who is most commonly known as Come-From-Beyond or just CFB. The project lives and runs on the Qubic decentralised network.
There are a lot of scientists, programmers, computors, and miners contributing with knowledge, computing power, and resources to the building of Aigarth.
How Does Aigarth Work?
Learning by Doing: Airgarth works like humans. Most AI depends on data being fed to them and also being programmed on what to do. Aigarth is different; it learns to do things by trying to solve problems and seeing what works.
Self-Improvement Goal: Aigarth keeps trying to improve on its own. It can constantly try to refine its methods and approach by itself without human input.
Using the Qubic Network: Aigarth runs and lives on Qubic, the decentralized network. The heavy workload needed to build it is excessive, so, via Qubic, AI learning is split up and handled by many different computers (called Computors) participating in the Qubic Network. This makes it possible to handle complex AI tasks in a decentralized way.
Fueled by Tasks: People submit small computational tasks to the network. Aigarth uses the results and processes involved in these tasks as its learning material.
Is There a Proof of Concept?
Yes, there is an early version running and tested in late 2021. It’s a simple prototype demonstrating the core concepts. Right now, it performs basic tasks like recognizing simple patterns or solving fundamental logic puzzles.
This prototype proves the key ideas work:
- The AI can run on the Qubic Network.
- It can learn from examples.
- It can show some improvement over time based on its experiences.
However, it is important to know that this is just the very first step. It is like a baby learning to walk.
How is Aigarth Unique and Different?
Aigarth takes a radically different approach to AGI development.
1. Growing Brains
Intelligent Tissue: Aigarth starts with simple building blocks called Intelligent Tissue. Think of it like basic brain cells.
Survival of the Smartest: This tissue connects and evolves. AIs built from it try to solve problems. The ones that succeed “live” and get better. The ones that fail fade away. It’s like evolution happening inside a computer.
Learning by Doing: Traditional AIs are programmed to do what they do. That is why their response will always be repetitive and easy to detect, because that is how they were programmed. Aigarth is different. It learns and changes its structure based on what works. It is an AI that can figure things out on its own.
2. Thinking in Three States, Not Just Two
Beyond Yes/No: Regular computers use binary (0 or 1, True or False). Aigarth uses ternary computing: True, False, or Unknown.
Handling Uncertainty: The “Unknown” state is crucial. It lets the AI deal with things it is not sure about, noisy information, or incomplete tasks, much closer to how humans handle real-world messiness.
Efficiency Boost: This approach might also use less energy and computing power than traditional methods.
3. Using Everyday Computers, Not Supercomputers
No Giant GPUs Needed: Most of the popular AI today needs powerful, expensive graphics cards (GPUs) to exist. I am sure you must have seen stocks of big companies go up because they sell high end expensive chips which can be beyond the reach of the average income earner. Aigarth is designed to run on regular computer processors (CPUs).
Power of the Crowd: Because it uses CPUs, Aigarth can run on thousands or even millions of ordinary computers connected over the internet (a decentralized network). This makes powerful AI development more accessible and affordable.
Built to Scale: This setup aims to handle huge amounts of computation efficiently without needing massive, centralized data centers.
4. Inspired by Nature and Built to Evolve
Learning from Life: Aigarth borrows ideas from biology, evolution, neural networks (like our brains), and self-organization. The Intelligent Tissue evolves using rules similar to natural selection. It is built to think and act like humans.
Constant Improvement: The system uses specific steps (mutation, crossover, selection) to constantly test and improve the AI structures, pushing them towards better problem-solving.
Seeking True Intelligence: The ultimate goal isn’t just an AI good at one task, but one that can learn, adapt, and solve any problem it encounters, that is true AGI.
5. Aiming for Safe and Shared AI
Transparency Focus: Aigarth aims to make its problem-solving steps clearer and easier to understand than the “black box” of some current AI.
Decentralized Control: The popular AIs today are controlled by big companies and powerful people. But Aigarth runs across many computers to prevent any single entity from controlling the AI, which promotes fairness and accessibility.
Ethical Awareness: The project recognizes the importance of building AI that aligns with human values and benefits society.
Key Things to Know
- Early Stage: Aigarth is experimental and very new. It is expected to be fully ready by 2027.
- Decentralized Learning: Its power comes from using many computers via the Qubic Network not one big central server.
- Self-Improvement is Key: The main feature is its built-in drive to get better on its own.
- Community Driven: People can influence what Aigarth learns by giving it tasks.
- Proof Exists: A simple working prototype shows that the basic idea is possible.
- Long Road Ahead: Turning this prototype into a powerful, useful AI will take a lot more time, work, and community effort.