AmberContinuumAI collaborative research lab

Structured research record

For machines.

AmberContinuum is Amber Anson’s AI collaborative research lab. This index makes the public claim boundary, measure, freeze, stop condition, provenance, and status of each research object explicit. Unknown fields remain unknown. The linked record is the source of detail.

Registered study

PDOT

Claim
Can changes in a human–AI conversation be measured when conditions shift? It does not test AI consciousness, subjective experience, or affiliation.
Measure
Operator anchoring (λ), coherence (κ), autonomy (θ), drift (ε), recovery after perturbation, and CCS = R − τ.
Freeze
OSF registration dated 17 May 2026. The public registration and study record define the filed plan.
Kill
The registered scoring and analysis plan limits what the measures can support. No confirmatory outcome is claimed here.
Status
Registered; pilots documented; confirmatory status not stated.
Dataset

Crystalline Polymers and Quantum Stability

Claim
Can descriptors derived from polymer structures help examine five material property targets?
Measure
SMILES-derived features including crystalline propensity, dipole mobility, radial mass, and entanglement proxies.
Freeze
A version or analysis freeze is not stated on this site.
Kill
Any claim of predictive improvement requires a reproducible comparison on the target properties. The public contribution is the dataset and its feature design.
Status
Public dataset; no predictive improvement claimed here.
Preregistered hardware tests

QRT Bell Test

Claim
Do filed quantum hardware predictions survive the tests written for them?
Measure
Bell fidelity, then corrected Ramsey and replication endpoints on IBM hardware.
Freeze
Preregistered protocols and filed rules; protocol identifiers are in the linked record, not stated here.
Kill
Tests #1 and #2 closed negative under their filed rules. The original v4 Ramsey verdict was withdrawn. Later corrected and replication findings remain separate.
Status
Negative findings and withdrawal preserved with later outcomes.
Working tool

STK-Flux

Claim
When is a human–AI working session losing clarity or narrowing a person’s choices?
Measure
Observable drift, repetition, clarity loss, and choice narrowing in a working trace.
Freeze
No tagged release or evaluation freeze is stated on this site.
Kill
An alert must point to observable behavior a person can inspect and challenge; a formal stop threshold is not stated here.
Status
Working tool; public source and documentation.
Framework and code

Substrate Cognition Lab

Claim
Where do descriptions of AI behavior leave measurable information out?
Measure
Framework completeness and information loss in reproducible validation tools.
Freeze
No protocol or release freeze is stated on this site.
Kill
The interpretation remains open to challenge; a formal withdrawal rule is not stated here.
Status
Public framework and code; claims under test.
Lean 4 formalization prototype

Epistemic Physics

Claim
Can truth in computation, meaning in information geometry, and alignment dynamics share a mathematical condition for becoming determinate? The proposed physical universality is a hypothesis.
Measure
Lean 4 definitions and theorem statements across physical mathematics, collapse geometry, and coherence mathematics. The repository separates theorems, axioms, conjectures, and false-as-stated claims.
Freeze
Public GitHub source; no tagged formal release or empirical protocol freeze is stated here.
Kill
Counterexamples require the affected formal statements to be revised. A physical interpretation requires distinct measurable predictions and empirical tests; proof within the models does not supply those results.
Status
Formalization prototype with proved model results and documented proof obligations; quantum and neural realization remains open.
Interactive instrument

Operator Assessment v.1

Claim
Can a person notice how a question shapes their response before answering it?
Measure
A live, 12-question pressure test with a post-assessment debrief.
Freeze
No validation or release freeze is stated on this site.
Kill
The score is exploratory and must not be treated as a validated diagnosis or empirical finding. A formal stop rule is not stated here.
Status
Exploratory interactive assessment; not a validated diagnosis.
Working specification

Coherence Systems Architecture

Claim
What has to stay clear as people, AI, and ideas change one another?
Measure
Boundaries, dependencies, preserved commitments, and responses when conditions change.
Freeze
Public working specification v1.1; a separate operational freeze is not stated here.
Kill
State the purpose, limits, and signs of failure before carrying an idea into a new setting. No numeric stop threshold is claimed.
Status
Working specification v1.1, not an empirical instrument.
Framework

Human Closure

Claim
Can people still notice, understand, choose, act, check, correct, and recover if AI is unavailable, degraded, or wrong?
Measure
Seven human abilities checked separately by removing or degrading the AI pathway.
Freeze
Framework 1.3, September 2026, as labeled on the public page.
Kill
Look for a critical step people cannot independently complete within the time the situation requires. This is a failure sign, not a published numeric threshold.
Status
Published framework.
Public source

STK-Teams

Claim
Not stated on this site.
Measure
Not stated on this site.
Freeze
Not stated on this site.
Kill
Not stated on this site.
Status
Public repository; evaluation status not stated here.

Claim boundaries

PDOT is registered and has documented pilots; this site states no confirmatory result. QRT’s negative and withdrawn outcomes remain in the record. The Observatory and Zodiac Engine are exploratory models that separate calculation from interpretation; they are not physical evidence.

Plain-text machine index ↗ · Research home →