The Architecture of Continuity
Altered Records and Available Actions
- How AI Tutors Hack Their Own Predictions — Audio Overview
Selective Retention and the Geometry of Continuation
Correspondents and Residents — Incomplete
Generative Continuation Geometry — Incomplete
Computation as Containment — Radiative Switch Computation
The Moon Should Not Be a Computer
- Audio Overviews – Why AI Servers Fail in Space and Why Data Centers Must Mimic Trees
This repository presents a unified framework for understanding artificial intelligence infrastructure as a thermodynamic and semantic system rather than a purely informational or computational one. The central thesis is that long-term stability in AI systems requires xylomorphic integration, in which computational residues—especially heat—are recursively reintegrated as substrates for future operation.
The documents in this repository develop this idea across multiple domains, including thermodynamics, infrastructure engineering, epistemology, and regulatory policy. Together, they articulate a theory of dual closure, defined as the simultaneous achievement of thermodynamic contraction (λ < 1) and informational exogeneity.
The materials are organized into complementary documents, each addressing a different layer of the framework.
The conceptual basis of the framework is introduced in
Conceptual Analysis, which develops the core ontology of xylomorphic systems, including the λ parameter, entropy reintegration, and the Thermodynamic Selection Theorem. This document establishes the formal language used throughout the repository.
A more accessible and narrative-driven introduction is provided in
Blog Post, which critiques proposals for off-world computation and introduces the notion of trophic, Earth-integrated intelligence systems.
The high-level strategic implications of the framework are developed in
Briefing Doc. This document presents the case for reclassifying AI infrastructure as a thermodynamic system and introduces the requirement of dual closure for stability.
Regulatory implications are formalized in
Regulatory Framework, which defines the Proof-of-Useful-Work-and-Heat (PoUWH) standard and the criteria for Public Research Object (PRO) funding. This document translates theoretical constraints into enforceable policy mechanisms.
The informational analogue of thermodynamic closure is explored in
Cognitive Ethics. This document examines how AI systems shape human reasoning, introducing concepts such as epistemic endogeneity, prior initialization, and the collapse of hypothesis space. It extends the λ framework into the domain of cognition.
Practical implementation details are specified in
Design Specification, which provides an engineering blueprint for integrating high-density GPU clusters with industrial thermal sinks. This includes exergy efficiency targets, scheduling constraints, and thermodynamic verification protocols.
Across all documents, a single principle recurs: stability is determined not by input abundance but by structural integration. Systems that externalize their residues—whether thermal or informational—enter an expansive regime (λ ≥ 1) and become unstable under scaling. Systems that reintegrate their outputs into their own conditions of operation achieve a contractive regime (λ < 1) and converge toward stable attractors.
This applies equally to physical infrastructure and cognitive systems. In both cases, the failure to maintain closure leads to divergence, while successful reintegration enables indefinite operation.
Only those systems that close their entropy flows compute indefinitely.