The Physics of Governance: Thermodynamic Limits of Autonomous Agents
Fuente:
Zenodo
Salvato in:
| Autore principale: | |
|---|---|
| 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 |