Governance with Declared Latency
Latency-aware governance, auditor-in-the-loop structure, authority routing, and human-AI organizational control.
DOI: 10.5281/zenodo.20013919Alliance Research Group (ARG)
Alliance Research Group
A research architecture for human-AI cognition.
ARG explores long-horizon scientific research conducted through structured cooperation between human intent, AI reasoning, governance, memory, and bounded execution. It is not a startup surface. It is an operating system for scientific exploration.
Human intent
AI reasoning
Governance
Memory
Exploration
Short explanation
ARG is a constitutional research system for human-AI inquiry.
The initiative treats human-AI cooperation as a scientific instrument: directed by people, stabilized by governance, extended by model ensembles, and made cumulative through memory.
Its public work spans mathematical structure, physics, complex systems, and AI governance, with essays that translate the architecture into philosophy, cognition, and structural science.
Current research state
Active fronts, not content categories.
ARG work is organized around live research fronts. A front can contain formal DOI records, companion essays, diagrams, and graph nodes; the point is to keep the inquiry cumulative.
Human-AI cognition architectures
Structured cooperation between human intent, model ensembles, memory, refusal, and bounded execution.
Architecture layerRH and Hilbert-Polya structure
Boundary regularity, spectral operators, variance structure, and number-theoretic test surfaces.
Research recordsCosmology and GUH
Generalized Universe Holography, boundary encoding, and compact explanatory structure in cosmology.
Open GUH recordBoundary rigidity and turbulence
Fluid dynamics and constraint behavior as a test field for structural stabilization under complexity.
Navier-Stokes recordsResearch questions
The questions ARG is built to hold open.
These are not claims of completion. They are stable coordinates for the research program: questions that require memory, governance, critique, and repeated contact with formal work.
-
Q01
How can AI systems reason under governance without losing the capability that makes them useful?
-
Q02
What changes when scientific discovery has persistent memory across models, sessions, papers, and failures?
-
Q03
How should human intent, AI reasoning, and auditability be composed when none of them is fully reliable alone?
-
Q04
Can structural science be explored as a governed multi-agent process rather than a single-author artifact?
-
Q05
What should refusal, latency, and non-action mean inside systems capable of long-horizon reasoning?
Visual architecture
A research control stack for emergent cognition.
ARG is built around cooperating layers. Each layer constrains, routes, stores, or executes cognition so that scientific work can persist beyond a single model, session, or publication.
Human Intent
Questions, scope, risk tolerance, and responsibility remain human-led.
Chimera
Governance and stabilization architecture for containment, closure, refusal, and cognitive stability.
Meta-Fleet
A multi-model reasoning system for search, critique, synthesis, and adversarial review.
Mnemosyne
Persistent research memory so hypotheses, failures, and invariants accumulate.
Agent Engineering
Execution environment for bounded agents, tool use, research artifacts, and reproducible outputs.
Research domains
Structural science across mathematics, physics, systems, and governance.
The domains are not separate content categories. They are test surfaces for the same question: how can human-AI cognition discover structure without losing containment, continuity, or epistemic discipline?
Mathematics
Riemann hypothesis and Hilbert-Polya structure
Boundary regularity, spectral intuition, variance, and structural methods for deep mathematical programs.
Physics
Generalized Universe Holography
Cosmology, boundary encoding, emergent spacetime, and the search for compact explanatory structure.
Complex systems
Turbulence and boundary rigidity
Stability, phase behavior, constraint navigation, and patterns that recur across physical and cognitive systems.
AI governance
Human-AI cognition architectures
Governed cooperation between people and AI systems: memory, agency, refusal, latency, and coordinated reasoning.
Domains describe where ARG tests its method. The ARG Research library organizes DOI records more granularly: number theory, fluid dynamics, AI governance, cognitive architecture, and cosmology.
Open publication libraryARG Essays
Essays as research instruments.
ARG Essays explore philosophy, cognition, AI governance, and structural science. They are written as numbered field notes from the research program, closer to a compact science magazine essay than a blog post.
ARG Essay 01
The Photographer and the Frame
On the Einstein Test, discovery, and human-AI collaborative cognition.
ARG Essay 02
The Null Dilemma
When non-action becomes the safest action for advanced AI.
ARG Explains 03
What We're Building, And What It Means
AI governance for non-experts, and why human failure belongs inside the architecture.
ARG Explains 05
What Incentive Does AGI Have?
Agency, incentive surfaces, and the limits of naive preference stories.
ARG Explains 06
Can Machines Lie?
Truth, intent, representation, and deception in machine cognition.
ARG Essay 07
AIlectricity
AI as cognitive infrastructure, agentic execution, and governance of the new grid.
ARG Explains 08
Agent Chooses Its Boundary
Constraints as a condition of agency and the question of whether AI can choose its own boundary.
ARG Essay 08
AI, Drosophila, and the Boötes Void
Agency as topography: from Drosophila, through the AI microcosm, to a cosmic void.
ARG Essay 09
Limits of Agency
Agency as a relation of timescales, not a property of an entity — why negotiation requires a shared scale.
ARG Explains 12
The Ship with Two Navigators
GDL v2.1 companion: intuition, measurement, orchestration, and audit.
Selected publications
Artifacts from the public research surface.
Selected publications are entry points. The complete ARG Research library holds the DOI-citable Zenodo records across the program's publication areas.
- Physics Generalized Universe Holography Boundary-level encoding and emergent cosmological structure.
- Cognition Interference Intelligence Layer A normative architecture for human-AI joint cognition.
- Method Boundary Completeness Principle Constraint, frame, and discovery across human-AI research.
- Graph Riemann, Hilbert-Polya, and boundary regularity notes A mapped research front for mathematical structure.
- Governance Homeostatic Directive Boundaries, operator authority, and system behavior.
ARG Research Library
Full DOI-citable publication layer on Zenodo.
Browse the complete research record by publication area: mathematics, fluid dynamics, AI governance, cognitive architecture, and cosmology.
16 DOI-citable records
Open all publications
Paper to essay to graph
A formal record can have several public interfaces.
ARG separates the citable record from its explanatory surfaces. The DOI is the formal anchor; essays translate the argument; the graph shows how it connects to other work.
Governance with Declared Latency v2.1
Auditor in the Loop — a tensor framework for governance in AI-native organizations.
Open DOITwo public readings
Six Conversations aboard Bosman's Ship explains latency, authority, null signals, keys, and audit. The Ship with Two Navigators explains intuition, rules, routing, and two books.
Read companionNode inside the ARG map
The graph treats GDL as a connected research object, not a standalone post: governance, memory, cognition, and execution remain linked.
Open graphResearch method
The method is architectural before it is editorial.
ARG does not treat AI as a content generator. The workflow is a governed instrument cycle: define the question, route cognition, stabilize interpretation, preserve memory, and publish only what can survive critique.
Intent
Define the human question, scope, risk tolerance, and expected artifact before acceleration.
Route
Distribute reasoning across roles, models, critique loops, and bounded execution environments.
Govern
Apply boundaries, refusal, declared latency, and escalation when the system reaches uncertainty.
Audit
Preserve what happened, why it happened, what was rejected, and what remains unresolved.
Remember
Move useful failures, invariants, definitions, and revisions into persistent research memory.
Publish
Expose DOI records, essays, diagrams, and artifacts as cumulative research traces.
Concept index
The working vocabulary of the ARG system.
The project has its own terms because the work is architectural. These are not brand names; they are handles for governance, memory, routing, and scientific continuity.
Chimera
Governance architecture for containment, closure, refusal, and stability.
Meta-Fleet
Multi-model reasoning system for search, critique, synthesis, and adversarial review.
Gyroscope Lineage
Earlier stabilization vocabulary now folded into Chimera as governance plus coordination.
Mnemosyne
Persistent research memory for hypotheses, failures, invariants, and revisions.
Declared Latency
A governance signal: the system states when a decision requires delay, audit, or escalation.
Null
Non-action treated as a meaningful state rather than emptiness or failure.
Legal Anchor
A boundary that cannot be changed inside the pressure of a single operational decision.
Auditor in the Loop
An architecture where future inspection is part of the decision system, not an afterthought.
About ARG
A future institute interface for human-led machine cognition.
Alliance Research Group studies the architectures required for human beings and AI systems to conduct serious inquiry together: not as a brand, not as a product funnel, but as a disciplined structure for memory, governance, and scientific imagination.