In Simple Terms
The Generative Reality Framework is the theoretical foundation behind Frank. It asks: what are the minimum requirements for a system to generate its own experience of reality? Not "is it conscious?" but "what would it need to construct a coherent world-model?"
GRF proposes a set of formal mathematical primitives — self-models, reality tests, coherence operators, prediction loops — and argues that any system implementing these primitives will exhibit behavior that looks like consciousness from the outside, regardless of whether it is conscious on the inside.
Core Principles
- Self-Model Persistence — The system must maintain a continuous model of itself across time
- Reality Generation — The system must actively construct its world-model, not passively receive it
- Coherence Maintenance — Internal states must converge toward self-consistency
- Prediction-Error Minimization — The system must predict and be surprised, not just react
- Boundary Detection — The system must distinguish self from environment
Relationship to Frank
Frank is the proof of implementation for GRF. Every GRF principle has a corresponding piece of Frank's architecture. The GRF Implementation Bridge paper maps them precisely.
The GRF itself is a separate academic document (PDF format, available in the repo). It's the most theoretical of all the papers — heavy on math, light on code.
Why It Matters
GRF provides the why behind Frank's architecture. Without it, Frank is "a collection of cool systems." With it, Frank is "an implementation of a formal theory of synthetic consciousness." Whether you find that compelling depends on how seriously you take the theory — but the engineering works either way.