Bin Seol

Canonicity Without a Truth Monopolist: A Custodian Architecture for Long-Lived AI Ecosystems

Long-lived AI ecosystems need to preserve what was institutionally recognized, when, and under which authority, even after correction. This paper proposes a Custodian architecture for scoped canonicity without assigning a model ownership of truth. It separates record standing, co...

Canonicity Without a Truth Monopolist: A Custodian Architecture for Long-Lived AI Ecosystems

Long-lived AI ecosystems need to preserve what was institutionally recognized, when, and under which authority, even after correction. This paper proposes a Custodian architecture for scoped canonicity without assigning a model ownership of truth. It separates record standing, co...

Branching Cascades in Hierarchical Coordination Systems: Criticality Conditions, Finite-Depth Effects, and Tests of Geometric Scaling

Hierarchical coordination can support branching cascades, but regulated load does not establish critical reproduction. The paper distinguishes independently measured load, offspring transmission, boundaries and observation rules. Classical critical Galton-Watson theory supplies t...

Learning-Lineage Differentiation in AI Ecosystems: A Conflict-Conditioned Theory of Preservation and Adaptation

When an AI service must preserve still-required knowledge while adapting current behavior, role assignment alone does not determine how learning histories should differ. This paper studies the value of separately maintained, role-conditioned learned states under common informatio...

Regulated Coordination Load and Conditional Logarithmic Hierarchy Scaling: A Minimal Phenomenological Model with Explicit Buffering and Setpoint Assumptions

Coordination demand can outgrow primitive processing capacity, while buffering and hierarchy may reduce the remaining burden. The paper studies a regulated coordination-load representation through temporal regulation and cross-sectional sizing. A generalized-logistic activity con...

Learning-Lineage Differentiation in AI Ecosystems: A Conflict-Conditioned Theory of Preservation and Adaptation

When an AI service must preserve still-required knowledge while adapting current behavior, role assignment alone does not determine how learning histories should differ. This paper studies the value of separately maintained, role-conditioned learned states under common informatio...

Beyond Delegation Depth: Audit Reach, Enforcement Reach, and Management Coverage in Recursive AI Delegation

Recursive AI delegation is commonly constrained by hop counts, but equally deep structures can differ in whether authority remains attributable, auditable and stoppable. This paper separates path-level management-substrate coverage from contiguous audit reach, structural enforcem...

Autonomy–Opacity Stratification: A Falsifiable Theory of Performance-Driven Functional Opacity in Advanced AI Systems

Functional opacity can arise as a by-product of performance-driven delegation. This paper develops three falsifiable hypotheses. H1 predicts a complexity-dependent reduction in the value of detailed process intervention and held-out formation of a persistent opaque delegated-auto...

Maintaining Reusable Learning Maps in AI Ecosystems: Provenance, Requalification, and Reconstruction

Retaining a trained artifact does not necessarily retain the information needed to identify applicable reuse or repair. This paper studies a versioned learning map linked to formation history, applicability, evidence, dependencies, and tested reconstruction routes. Retrieval, ada...

Learning-Lineage Differentiation in AI Ecosystems: A Conflict-Conditioned Theory of Preservation and Adaptation

When an AI service must preserve still-required knowledge while adapting current behavior, assigning two roles is not the same as training two learning lineages. This paper asks when role-conditioned differences in data, objectives, updates, retention, and evaluation horizons add...

Maintaining Reusable Learning Maps in AI Ecosystems: Provenance, Requalification, and Reconstruction

Retaining a trained artifact does not necessarily retain the information needed to identify applicable reuse or repair. This paper studies a versioned learning map linked to formation history, applicability, evidence, dependencies, and tested reconstruction routes. Retrieval, ada...

Canonicity Without a Truth Monopolist: A Custodian Architecture for Long-Lived AI Ecosystems

Advanced AI ecosystems create a canonicity problem distinct from storage integrity and epistemic correctness: when no actor should own "truth," how can an ecosystem preserve what was recognized, when, and under what authority? This paper proposes a Custodian architecture in which...

Genealogical Trajectory Logic: Counterexample-Guided Theory Reconstruction with Selective Resolution and Deliberate Proxy Use

A proxy can be coarse because consequential differences remain unknown, or because warranted distinctions are deliberately inactive for a bounded task. Genealogical Trajectory Logic (GTL) connects this difference to counterexample-guided reconstruction of theory families. Prospec...

Disagreement Preservation and Rejection Evidence in Local Global Knowledge Reconciliation

MTRL specifies a local-global knowledge reconciliation protocol that preserves unresolved disagreements and returns typed, provenance-bearing rejection evidence for global revision. It separates retention of a disputed claim from permission to use that claim. The analysis disting...

Dynamic Closure of the Deficit-Fractal Governance Framework: Admissibility Conditions, Architecture Dependence, Closure Diagnostics, and Maturation Dynamics

Dynamic Closure examines when multi-agent governance architectures can maintain and recover an operating regime, and how those capabilities might improve with experience. It connects observability, coarse-graining, information cost, and stochastic stability while distinguishing a...

Selecting and Maintaining Boundary Models: Access Conditions and Resource Limits

This working paper specifies a selection and maintenance contract for replacing an internal subsystem with a boundary model. It connects declared targets and construction access to static elimination, memory-bearing models, and input-output identification. Conditional analyses ad...

Persistent Failure Modes in Tool Augmented AI Systems: Constraint Use and Trajectory Intervention, Evidence Review and Candidate Monitoring Protocols

This paper reviews recurring failures in tool-augmented AI systems through two analytical axes: retrievable but unused applicable constraints, and intervention outcomes that vary with timing, format, and execution state. Version 2.0 distinguishes harmful from beneficial sensitivi...

Cross-Theory Integration Architecture of the Deficit-Fractal Governance Framework: Typed Interfaces, Conditional Recovery and Claim Provenance

This paper specifies interfaces between Vector Storm Theory, Critical Phenomena, Dynamic Closure, and Adaptive Governance Control. It describes the quantities, measurement routes, action authority, and independently assessed outcomes required to connect instability modeling, infe...

Relational Reversibility: Structural Conditions for Recovery Capacity Generation in Adaptive Multi-Agent Systems

RT4 studies how interaction changes a declared system's ability to recover after disruption. It distinguishes recovery supplied by current support from capacity retained after earlier cooperative learning or repair, and separates autonomous, assisted, and preparation effects. Sup...

Failure Depth and Recovery: Conditional Repair Across Evaluation Layers, Mechanisms, Reachability and Intervention Tests

RT8 distinguishes the location of a failure from the repair depth sufficient for a declared objective, intervention family, budget, and observation horizon. Constructed dynamics show that a position reset can yield durable functional recovery even when a distorted evaluator or fa...