Independent reconstruction
Functional specification, frozen tests, controls, acceptance gates, and reporting templates.
Open independent replication invitation
Rebuild it.
Try to break it.
Tell us what survives.
We are inviting programmers, complexity researchers, network scientists, artificial-life researchers, and constructive skeptics to independently reproduce an adaptive decentralized architecture.
ZIP · source, protocol, gates, templates, verifier, and archived reference implementation
The central question
Can a system coordinate, learn, remember, repair, generate novelty, and reorganize itself without installing a permanent sovereign subsystem—and can it recognize when no intervention or novelty is needed?
Why we are opening it
Encouraging internally.
Unproven independently.
Our internal simulations produced encouraging results under growth, shock, recovery, community formation, migration, shared memory, fading mentorship, agency development, and longer-run viability tests. They also produced failures and bottlenecks that changed the architecture.
Now we want the work rebuilt by people who did not write the original implementation. Skeptical and adversarial attempts are especially welcome, including different programming languages and mathematical formulations. A useful result may be a success, a partial replication, a clear failure, or the discovery that one of our definitions is underspecified.
Clean-room structure
Held-back verification
Archived implementation and retained summaries, examined only after code and predictions are frozen.
Minimum expectations
Complete reporting.
Favorable or not.
- 01Register design, seeds, metrics, and predictions before viewing reference outputs.
- 02Run every required control and ablation—not only favorable cases.
- 03Preserve local harms and failures instead of hiding them inside averages.
- 04Publish code, dependency versions, raw outputs, analysis, and deviations.
- 05Keep demonstrated results separate from interpretation and metaphor.
Core tests
Where the architecture must answer.
Restraint in a Quiet World where nothing requires correction
Recovery after acute and repeated disturbances
Memory with provenance versus current-state-only control
Contextual influence versus permanent authority
Fading, absent, and permanent mentorship
Optional versus compulsory novelty and genesis
Structured versus randomized connectivity
Differentiation preserved while coherence is maintained
≠
The scientific boundary
Finite computational experiments—not proof of consciousness, biology, metaphysics, physical law, or a universal social design.
We are separating what the experiments demonstrated from what they merely invite us to investigate.
Your turn
Beautiful metaphors are easy.
We are looking for behavior that survives reconstruction.
Please return your preregistration, repository link, result report, and raw-output archive. Criticism is welcome; complete reporting is required.
Download replication packet ↓SHA-256: 2f1ad4aec1e7a4fc5c48422971f0b1ad3e7dbca2730d122e37327075ce87bac1
🌱 💎