model-in-a-world
Compiled with AI assistance and selectively reviewed. Please consult the cited passages; the books remain the authoritative source. About this companion
Editorial synthesis · provisional
Definition
The model-in-a-world is an AI model considered together with the technical, human, and institutional ensemble that sustains it and gives its outputs consequences. Infrastructure, interfaces, labor, law, and culture belong to this unit of analysis. Changing deployment changes the system's relations and modes of correction; the model's weights alone do not specify what the system becomes.
Explanation
AB rejects both the neutral instrument and the sovereign mind as pictures of a self-contained system. Its classroom and courtroom contrast shows why the same output can enter different practices and acquire different consequences. The resulting capacity belongs to an arrangement rather than a component alone, while human responsibility is not dissolved into an imagined machine personality.
PB's Tractatus locates the ethical unit in the regime: interface, institution, and trust give output somewhere to go. Its reproduced AI exchange about having a world remains a different voice, not an independent authorial definition. RB3 develops the ensemble through Simondon's technicity and calls AI a system-with-a-world. Deployment changes individuation, but this account explicitly does not make AI alive. The term therefore marks situated technical organization, not proof of autonomous world-making or machine experience.
Scope and attribution
Direct treatment is assigned to AB, PB and RB3. RB1's maintained infrastructures and RB2's organism-centered scaffolding are relevant comparisons, not occurrences of this AI analytical unit.
AB — The Actual Book
Names the model-in-a-world as the unit of analysis. Classroom and courtroom deployments differ in consequence and correction; the sustaining ensemble includes infrastructure, interface, labor, institutions, law and culture. Distributed capacity does not make the model the moral author.
Occurrence locators: 152, 163, 164.
PB — The Possible Book
The Tractatus names the regime rather than the model as the ethical unit. A later categorical restatement presents model, interface, institution and trust as a composite. The inherited AI exchange contrasts human ecological niches with autonomous world-making but is retained as a reproduced answer, not an authorial definition.
Occurrence locators: 104, 155.
RB3 — The Real Book, Volume III: Murmurations
Calls AI a system-with-a-world and shifts analysis from model to ensemble. Its Simondonian account makes deployment a change in individuation while expressly withholding the conclusion that AI is alive.
Occurrence locators: 143, 144, 145.
Background used in the editorial explanation
PB’s reproduced AI conversation is preserved as a separate source voice and contextual comparison. The model–regime unit is established independently in the Tractatus and formal sketch.