Project Chimera /// Constitutional Cognition Stack

System Explanation

ARG Architecture

A constitutional cognition stack for governed human-AI scientific work.

ARG is not a single model, and not a publication layer with a manifesto attached. It is a system: human intent establishes direction, governance constrains action, multiple models contribute cognition, coordination maintains coherence, memory preserves continuity, and execution remains bounded.

Orientation

What This Page Is For

This page is a conceptual systems map, not an API reference. It explains the stack that ARG uses to frame human-AI scientific cognition: where intent enters, where governance operates, how cognition is distributed, how continuity is stabilized, and how execution stays subordinate to constitutional rules.

Stack

Constitutional Layers

Top to bottom: direction, control, cognition, coordination, memory, execution.

Layer 01

Human Intent

The source of direction, scope, and responsibility. ARG starts from human initiative rather than autonomous model goals; intent defines what the system is allowed to pursue.

Closest current route: Homeostatic Directive

Layer 02

Chimera

The governance layer. Chimera establishes stability conditions, refusal logic, and enforcement boundaries so the system remains governed even when capability increases.

Closest current route: Homeostatic Directive

Layer 03

Meta-Fleet

Distributed cognition across multiple models and roles. Meta-Fleet is the multi-model layer where different systems contribute search, synthesis, critique, and variation under one frame.

Closest current route: Interference Intelligence Layer

Layer 04

Gyroscope-Live

The coordination layer that prevents semantic drift and keeps cognition locked to intent. It is the homeostatic control function of the stack.

Closest current route: Interference Intelligence Whitepaper

Layer 05

Mnemosyne

Memory and continuity. Mnemosyne names the long-duration retention layer: how research persists, accumulates, and re-enters future work instead of dissolving with each session.

Current route: Project Mnemosyne

Layer 06

Spock

Execution under constraint. Spock is the task-facing layer: bounded action, structural output, and disciplined implementation rather than unconstrained generation.

Closest current route: System Kernel

System Logic

How The Stack Behaves

Human intent sets the vector. Chimera constrains the operating envelope. Meta-Fleet distributes cognition across complementary model roles. Gyroscope-Live keeps the process coherent. Mnemosyne prevents loss of continuity. Spock executes work within bounded authority.

Public Interface

How To Read It

The homepage is the entry point. This page explains the system. The essays page is the conceptual interface where the research becomes legible in public language. The knowledge graph provides the broader conceptual map around these layers.

Adjacencies

Related Surfaces

Research Essay

Boundary Completeness

Shows how human-AI collaboration produces discovery at the boundary rather than in isolated nodes.

System Map

ARG Knowledge Graph

A conceptual network of cognition, governance, science, and essays across the ARG ecosystem.

Editorial Surface

ARG Essays

The public essay series where architectural, scientific, and philosophical questions are made readable.