Get Involved
Hello friend, We are at a crossroads in our history where intelligence is being embedded into our core systems.
The software we have relied on for the last 30 years was built on a specific promise: it is repeatable, predictable, and deterministic.
The arrival of artificial intelligence in these environments introduces a fundamentally different challenge. Its behavior is probabilistic. Similar inputs can produce different outputs, making runtime policy enforcement, oversight, and auditability difficult.
SAFi grew from a framework for understanding human cognition and has become an open-source governance engine for AI agents.
It is designed to place deterministic governance boundaries around probabilistic systems. It enforces policies in real time, governs tool calls, and records every decision for audit.
The four design laws behind SAFi
1. The Law of Probabilistic Entropy
A probabilistic engine cannot be relied upon to enforce absolute boundaries on its own.
2. The Law of Structural Determinism
Critical policy gates and execution permissions must be enforced through deterministic logic, not semantic inference.
3. The Law of Faculty Isolation
A single engine must not generate a response, evaluate its own reasoning, and authorize its own execution without separation of roles.
4. The Law of Defined Teleology
To govern a system, not merely restrict it, you must define its purpose.
Ways you can participate
SAFi is an open-source project. However you work, there is a way in:
- Run it: Explore the GitHub repository, run the live demo, or hire us to get started.
- Read it: The research preprint, the mathematical specification, and the GitHub documentation.
- Challenge it: If you are a classical thinker, a Thomistic philosopher, or are simply intrigued by the project, do not shy away, I want to hear your thoughts about it.
Get in touch using the form below.
