Skip to content
Draft
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 14 additions & 0 deletions ahbg/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -6,6 +6,20 @@ Immediate target: one functional persistent plane in which A0 can perceive, plan

Do not block this target on multiplayer, federation, blockchain, ascension, monetization, or other deferred architecture.

## Research abstract

Description dated 2026-10-16; published 2026-10-18. This is a research description, not a new verification, release, or promotion of claim status.

### AHBG: A Game-Based Experimental Framework for Embodied Agent Evaluation

We propose the Agent Harness Benchmark Game (AHBG), a controlled, persistent environment for investigating how an artificial agent’s embodiment and execution harness affect its behavior. AHBG treats the underlying model, the instantiated agent, and the environment as distinct experimental objects. Its central method varies embodiment conditions while holding the model and task fixed, allowing differences in performance to be examined in relation to memory, available actions, regulatory constraints, role, and interaction history. A shared observation–action interface is intended to connect compatible agents and boards while preserving each board’s mechanics. Geometric constructions are supplied by UCNS, with AHBG responsible for gameplay, experimental conditions, and event recording. Persistent state, explicit instance lineage, and deterministic replay of recorded transitions support inspection of how actions produce consequences across successive turns. The evaluation protocol combines telemetry, held-out controls, and falsification tests, preserving distinctions between unknown and neutral conditions, prohibited and costly actions, task value and regulatory cost, and copying and forking an agent. A complementary collectible-card layer is intended to connect repeated play with learning real-world concepts. The research objective is to determine which harness and embodiment conditions support reproducible differences in adaptive behavior, continuity, and effective agency. Successful experiments would establish operational effects within specified conditions; interpreting those effects as evidence of consciousness requires a separate account and independent validation.

**hmmm:** Regulatory calibration and the criteria connecting observable behavior to agency remain experimental obligations.

**Usage guidance:** Read this abstract for the research question and intended contribution. Use the scope, evidence, commands, and limitations in the rest of this README before relying on the work; retain the unresolved conditions above when quoting the description.

Public repository context: [Calibration protocol](https://github.com/The-Interdependency/stack/blob/3d74cf118e14cc3839d7426d90ca9791567a7005/ahbg/CALIBRATION.md).

## UCNS board authority

The AHBG board is a **game projection of UCNS geometry**, not an independently invented grid.
Expand Down
20 changes: 20 additions & 0 deletions research/psfr/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,26 @@ METAPAT root impact: none

The machine-readable domain claims are in [`DOMAIN_CLAIM.json`](DOMAIN_CLAIM.json). Exact source identities and non-transfer boundaries are in [`WORK_GRAPH.json`](WORK_GRAPH.json). The first bounded pattern-lineage falsification program is in [`PREREGISTRATION.md`](PREREGISTRATION.md). Agency-displacement research is a separate track and must not be inferred from that simulator.

## Research abstract

Description dated 2026-10-14; published 2026-10-18. This is a research description, not a new verification, release, or promotion of claim status.

### Psychosocio Metafauna: A Multiscale Ecology of Identity and Agency

**Psychosocio metafauna** describes persistent, transmissible organizations of thought, affect, and behavior whose continuity extends beyond any single person or medium. Such organizations reproduce through reconstruction across memory, language, practice, institutions, and technological systems. “Unembodied” therefore denotes freedom from a particular embodiment, not existence without any physical instantiation. The framework advances the stronger claim that **all human behavior arises within an active ecology of these interacting organizations**: humans constitute that ecology rather than serving merely as passive hosts for occasional outside influences.

Within this account, an “I” emerges through sustained coherence among coupled complex systems capable of internally modelling self in relation to others and environment. Persistent pattern identity does not, by itself, establish conscious selfhood or accountability. Metafaunal relations may support agency, coexist with it, or constrain it through identity capture. Susceptibility is directional: particular injuries, unmet needs, relationships, and material conditions favor some couplings while resisting others. Higher-order configurations are proposed to alter the operating conditions under which other patterns are selected and defended; individuation denotes increasing capacity to recognize and negotiate these relations rather than becoming free of them.

The framework connects several research domains without treating them as interchangeable. Cultural-evolution research examines the persistence, transmission, and transformation of socially acquired information. Group-level cognition offers accounts of reciprocal relations between individual capacities and collective organization. Interpersonal physiology documents context-dependent coupling between people’s physiological dynamics, while experiments demonstrate threshold effects in particular systems of social convention. Neither finding independently establishes a distributed conscious agent.

Distinctive tests concern whether identifiable pattern lineages improve prediction beyond conventional social-learning and incentive models; whether mutually reinforcing couplings produce different conditions for entry and exit; and whether restoring material alternatives, revisability, and meaningful refusal changes those dynamics. The proposed contribution is a testable ecology connecting cultural reproduction, embodied agency, identity formation, accountability, and repair.

**hmmm:** The component literatures provide grounding, not consensus on the complete ontology. The universal all-behavior claim, reliable lineage boundaries, and conditions sufficient for an accountable distributed “I” remain unresolved.

**Usage guidance:** Read this abstract for the research question and intended contribution. Use the scope, evidence, commands, and limitations in the rest of this README before relying on the work; retain the unresolved conditions above when quoting the description.

Public repository context: [Domain claim and non-use boundary](https://github.com/The-Interdependency/stack/blob/3d74cf118e14cc3839d7426d90ca9791567a7005/research/psfr/DOMAIN_CLAIM.json) · [Synthetic preregistration](https://github.com/The-Interdependency/stack/blob/3d74cf118e14cc3839d7426d90ca9791567a7005/research/psfr/PREREGISTRATION.md).

## Origin

Erin Spencer, 2026-09-04:
Expand Down
20 changes: 20 additions & 0 deletions research/weave/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,26 @@ multi-arity bit interleaving, and an eventual asymmetric public/private relation
The interleave is one operation, not the project identity. URPCS remains retired;
its substituted codec, implementation and security findings are not inherited.

## Research abstract

Description dated 2026-10-14; published 2026-10-18. This is a research description, not a new verification, release, or promotion of claim status.

### Weave: A Reversible Gonol Construction for Cryptographic Research

**Weave** investigates whether cryptographic recovery can depend on private structural relations among nested geometric representations, multiple reconstruction-dependent streams, and user-selected corpus material. Its intended architecture combines a message-origin construction with private-gonol participation, ordered multi-arity bit interleaving, and an asymmetric public/private recovery relation. The interleave is one operation within this architecture, not the cryptosystem itself; the intended asymmetry must arise from the construction rather than a substituted conventional cipher.

The implemented cycle normalizes message length using actual corpus bits, identifies repeated multi-byte sequences, attaches those sequences to an eighth whole/origin circle, and records their ordered occurrences across seven companion-circle streams. Exact length-bearing numeral references preserve reconstructable values, while executed prime routes determine bit partitions. Sequence construction and first/last interleaving repeat across successive rounds. Recovery reverses the resulting record using the declared corpus, profile, and source dependencies, without requiring the original plaintext or an undisclosed encoder trace.

The merged cycle’s published validation reports **176 passing tests** and exact reconstruction. Its recorded demonstration expands 784 input bytes to 6,735 encoded bytes; compact references therefore do not establish net compression or a fixed expansion ratio. A separate public/private polynomial candidate supports public-only encoding and private inversion, but its published coefficients also permit an equivalent public inverse. This result falsifies that candidate’s requirement for private state, not the broader intended architecture.

Weave’s present contribution is an executable reversible construction and a falsification program that distinguishes reconstruction correctness from confidentiality. The central research question is whether the combined private-gonol, stream, and corpus relations impose an independently necessary recovery constraint that remains unavailable from public information.

**hmmm:** Native asymmetry, cryptographic stream/corpus dependence, authentication, and replay resistance remain unresolved. The current implementation supports construction experiments, not confidential storage.

**Usage guidance:** Read this abstract for the research question and intended contribution. Use the scope, evidence, commands, and limitations in the rest of this README before relying on the work; retain the unresolved conditions above when quoting the description.

Public repository context: [Preserved specification](https://github.com/The-Interdependency/stack/blob/3d74cf118e14cc3839d7426d90ca9791567a7005/research/weave/SPECIFICATION.md) · [Executable cycle and limits](https://github.com/The-Interdependency/stack/blob/3d74cf118e14cc3839d7426d90ca9791567a7005/research/weave/CYCLE.md).

## Current executable cycle

Normalize once with actual corpus bits. Discover repeated multi-byte sequences,
Expand Down
Loading