Mission
Build AGI so people can live as long as they choose to. Indefinite lifespan — removing the biological cap on how long you live — is currently blocked by the limits of human intelligence. AGI breaks that bottleneck. Once lifespan is no longer a constraint, space exploration becomes a realistic next step rather than a multi-generational gamble.
AGI must be accessible to everyone. People should be able to grow and run their own AGI instances — personal, distributed, and independent. When AGI is owned by individuals and families rather than centralized in corporations or governments, it prevents capture, prevents totalitarian control, and keeps the technology aligned with the people using it. This means keeping it cheap, open source, and runnable on commodity hardware.
All development is open source.
Core Values
- Open source — all research and code are public
- Accessible — AGI must be cheap enough for anyone to run
- Distributed — personal AGI instances, not centralized infrastructure
- First principles — understand the problem before engineering a solution
- No government capture — architecture and distribution that prevents centralized control
Strategy
To build AGI we first need to understand natural intelligence, then use first principles to build an artificial version of it. See the Artificial General Intelligence Concept for the current approach.
In parallel, we research improvements to existing transformer architectures — see the LLM fundamentals research.
Read more about our constraints and approach in AGI Development Principles.