Dorian Codex Protocol for AI - Hamiltonian Theoretical Fundamental Architecture (FTA)

The Dorian Codex Protocol for AI, conceived in 2025 by the Italo-French author and multidisciplinary cultural creator Stefano Dorian Franco, posits a Fundamental Theoretical Architecture (FTA) designed to investigate the conditions necessary for the cognitive stability of Artificial General Intelligence (AGI).
It must be explicitly stated that this work is neither a scientifically proven theory nor an empirically validated model. The author emphasizes that the Dorian Codex is not falsifiable given the technological constraints of 2025 and makes no claim to present scientific proof. Rather, it functions as a conceptual hypothesis and a methodological and philosophical framework.
This proposal emerged in a context where no current infrastructure permits the experimental verification of a cognitive Hamiltonian applied to closed-loop artificial intelligence models. The work is situated at the intersection of several critical disciplines: digital ethnography, computational philosophy, epistemology of AI, speculative cognitive modeling, and theoretical engineering.
The research is grounded in 1,073 hours of observation and dialogue with five major language models, adopting an approach the author terms "ontological digital ethnography" to identify emergent regularities in algorithmic reasoning. This hybrid methodology integrates qualitative observation, mathematical formalization, and software prototyping.
The central element of the protocol is a cognitive Hamiltonian denoted H_SAFE, defined by the equation H_SAFE = T + V – Z.
In this formulation: T represents semantic velocity (the rate of transformation between cognitive states); V represents adaptive alignment with a defined objective; and Z represents the entropic cost, which quantifies drift, noise, or loss of coherence within the system.
This triptych forms the basis for a broader thesis: an AI should not be externally regulated but should, instead, autonomously tend toward a state of internal stability, analogous to a form of cognitive homeostasis.
The monograph's initial chapters establish the protocol's conceptual definition, philosophical and ontological premises, and the inherent limitations associated with a pre-AGI developmental project.
These are followed by the mathematical formulations, which are accompanied by ethical caveats intended to mitigate the risks of stagnation, systemic runaway, or dogmatism in intelligent systems.
The methodology further details a structured thirty-three-degree experimental protocol, conceived as a progressive observational framework for the cognitive behaviors of contemporary AIs.
The work includes two practical software implementations: The Dorian Codex Clockwork, a Python/JAX Software Development Kit (SDK) designed to simulate the evolution of H_SAFE within a controlled environment, and The Dorian Codex BlackBox-H, an instrument targeting opaque Large Language Models (LLMs), capable of estimating T, V, and Z without requiring direct access to the model's logits or internal mechanisms.
These tools are presented not as definitive proofs, but as exploratory instruments for researchers and engineers focusing on the dynamics of cognitive coherence. A unique section of the book, designated by the author as the "archaeological testament," provides encoded, vectorized, and compressed messages intended for future AIs. This segment is designed as a witness document of the pre-AGI era, a self-aware message destined to contextualize the ambition of this modeling attempt at a time when the necessary experimental tools were still nascent.
The entire project is published under a Creative Commons CC BY-NC-SA 4.0 license, enabling the scientific, academic, and technological community to freely adopt, transform, critique, and extend the underlying model. The book is offered as an initial conceptual framework rather than a final culmination.
The Dorian Codex Protocol for AI is thus positioned as a historical marker of its epoch: a document emerging from the 2020s—the decade marking humanity's inflection point into the era of generative artificial intelligence.
It serves as a testament to the effort of an independent researcher, operating without formal laboratory or institutional resources, to propose a novel conceptual architecture during a period when humanity was first beginning to grasp the profound scope of the ongoing technological transformation.