Neuruh — Adaptive Intelligence Research Program

Latent Cross-Domain Emergence and Case Zero, Built Runnable

The hypothesis document as instruments: the nine-stage LCDE mechanism runner, the eight predictions, the six competing hypotheses with the observations that discriminate between them, the ten-dimension scoring worksheet, the domain ablation board, the five research ruling classes, and the pilot checklist.

7 pages — read in full9-stage mechanism 8 predictions6 competing hypotheses 10 scoring dimensions5 ruling classes

§1-2The Hypothesis and Its Nine-Stage Mechanism

LCDE v0.1 — When an adaptive agent develops deep representations across multiple structurally distinct domains before encountering a future target domain, recurring relational invariants may become encoded as latent schemas. When a novel target creates sufficient structural resonance and productive polarity, the schemas can recombine and reorganize into an integrated capability requiring less target-specific learning and producing broader transfer than equivalent target exposure without the prior structure.

Run the mechanism. Each stage is necessary in the hypothesis; remove one and the chain is the control condition, not the phenomenon.

Idle. Case interpretation (§3): Jeramie may have approached AI not as a blank-slate software learner but carrying deeply encoded schemas from physical markets, real estate, operations, defense, self-modeling, content, and IP packaging — with AI as the enabling medium in which dormant structures could be externalized and recombined. Plausible formulation. Not evidence.

§4The Eight Predictions

If LCDE is real, these eight things should be observable. Click each: what it claims, and the section of the study that tests it.

Select a prediction.

§5Competing Hypotheses — The Discriminator

Six rival explanations, each making its own prediction. The study is only meaningful where LCDE and a rival predict different observations. Click a rival.

Select a competing hypothesis.

§7Scoring Framework Worksheet — Ten Dimensions

Each dimension asks one question. Score a hypothetical result set and read the profile. This is the hypothesis document's qualitative framework; the quantified version (PP, FR, F1, SAS, gates) lives in the Case Zero protocol demo.

§8Domain Ablations — Expected Loss

Remove a domain and, if the hypothesis is correct, a specific capability family weakens. Toggle domains off.

§9Research Rulings — Five Classes

This document's ruling vocabulary. Set a result pattern and read which ruling it earns. (The later full protocol uses a stricter four-class vocabulary: Supported, Mixed, Unsupported, Protocol Failure.)

Select a ruling class.

§10Pilot Checklist

Nine steps to a minimum viable test. In the source document these are numbered 19-27, continuing the study-step list; they are renumbered 1-9 here for use. Step 1 has a hard dependency: the collector runs on Jeramie's local drive, and per the Thread Ledger its execution is still required there.

0 of 9 complete (browser checklist only).