The Physics of Governance: Thermodynamic Limits of Autonomous Agents

Fuente: Zenodo
Salvato in:
Dettagli Bibliografici
Autore principale: Davis, Matthew A.
Natura: Recurso digital
Lingua:inglese
Pubblicazione: Zenodo 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866901619077021696
author Davis, Matthew A.
author_facet Davis, Matthew A.
contents <p>The deployment of autonomous agents in high-stakes enterprise environments is currently constrained by a fundamental misunderstanding of alignment. We propose that "Governance" is not a policy layer but a thermodynamic function. By modeling the agent as a closed system subject to the Second Law of Thermodynamics, we demonstrate that "Drift" (G_I) is equivalent to entropy and "Alignment" (G_E) requires a continuous expenditure of work (W_Gov). We derive the <em>Governance Chandrasekhar Limit</em>, proving that as the Context Window (V_Context) expands, the density of the system prompt must scale linearly to prevent identity collapse. We conclude by defining the "Three Laws of Governance Dynamics," offering a mathematical framework for the stability of autonomous systems.</p>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_18604941
institution Zenodo
language eng
publishDate 2026
publisher Zenodo
record_format zenodo
spellingShingle The Physics of Governance: Thermodynamic Limits of Autonomous Agents
Davis, Matthew A.
AI Alignment
Thermodynamic Drift
Context Window Collapse
Enterprise Autonomous Agents
Constitutional AI
Physics of Governance
<p>The deployment of autonomous agents in high-stakes enterprise environments is currently constrained by a fundamental misunderstanding of alignment. We propose that "Governance" is not a policy layer but a thermodynamic function. By modeling the agent as a closed system subject to the Second Law of Thermodynamics, we demonstrate that "Drift" (G_I) is equivalent to entropy and "Alignment" (G_E) requires a continuous expenditure of work (W_Gov). We derive the <em>Governance Chandrasekhar Limit</em>, proving that as the Context Window (V_Context) expands, the density of the system prompt must scale linearly to prevent identity collapse. We conclude by defining the "Three Laws of Governance Dynamics," offering a mathematical framework for the stability of autonomous systems.</p>
title The Physics of Governance: Thermodynamic Limits of Autonomous Agents
topic AI Alignment
Thermodynamic Drift
Context Window Collapse
Enterprise Autonomous Agents
Constitutional AI
Physics of Governance
url https://doi.org/10.5281/zenodo.18604941