Natural Math v5 and Bolt-On v0.3
Natural Math v5 is the governing frozen integer baseline. Bolt-On v0.3 is the frozen host-portability evidence layer. CNTM software gates remain software-only.
Status
Every claim keeps its evidence layer.
This page separates frozen results, local validation, code-present work, specifications, proposed integrations, external reported results, negative results, and historical material.
ASF Core v0.1 and Motorola M0
ASF is public as a read-only observation fabric. Core v0.1 passed 30 tests and freeze verification; Motorola Live Probe M0 passed 46 tests and freeze verification with no device writes and no actuation. Claims: CLAIM-0108 through CLAIM-0111.
Specificity, Construction A+, v0.4 contract
These have local evidence under bounded fixtures. Construction A+ is not collision-free across morphologies; v0.4 is not production integration.
Cognitive Basin structures
Implementation structures exist for action/evidence namespaces and recovery routes, but complete production LLM behavior is not qualified here.
Consequential Formation v0.4 RC1
Current orientation and external-review candidate. It preserves category suspension, independence audit, typed return, and recursive self-application while integrating candidate executable formation, sparse recruitment, and distributed embodiment. It does not validate the framework universally or supersede frozen project claims.
Minimum Formative Machine v0.1 and Formative Field
MFM tests whether acquired formation, separated from host capability and policy, causally changes later recruitment, reasoning, or admissibility. The Formative Field is the candidate executable runtime; neither is an achieved intelligence.
Distributed embodiment and energy ecology
Embodiment is modeled as changing reachability over capability, energy, authority, local formation, and connectivity. Partition, transfer, reconciliation, and resource-placement tests remain build-test pending.
UFWK and CNTM admission
UFWK WeightReceipt, uncertainty envelope, routing projection, and evidence/interpretation split remain specifications unless a named implementation qualifies them.
Persistent observer stack
Learnable-novelty-informed Basin routing, CONFIGURATOR bridge, and the full Natural Math/UFWK/Basin/Bolt-On/host chain are not demonstrated end to end.
APTD v1.0
The architecture is public and reviewable. R-001 is public/frozen; the reported repaired M2 remains publication-HOLD; the physical APTD threshold has not yet been demonstrated. Claims: CLAIM-0104 through CLAIM-0107.
Epiplexity and learnable novelty
The current paper results are recorded as external reported results until Fractalish reproduction protocols are run.
Eracii Arena: Duel
Eracii Duel is Natural Math-informed applied development led by Melissa Ellen Clow. It is not part of the frozen v5 conformance record.
Preserved failures
Construction A+ collision, descriptor round-trip limits, no physical CNT memory, Acrobot target failure, and namespace-collision risks stay visible.
Original Baby AI M2 HOLD
The original authoritative-history continuity attempt exposed provenance and required-history fail-closed gaps. It remains visible; a later local repair has not cleared public freeze reconciliation.
Resonant Morphology
Historical development artifact and bounded software fixture, retained for lineage and experiment design, not flagship proof.
Current release boundary
No current release has yet demonstrated the complete Natural Math–UFWK–Cognitive Basin–Bolt-On pipeline operating end to end inside ChatGPT or another production language-model service.
Not claimed
- No completed production language-model integration.
- No proof of consciousness, sentience, artificial personhood, or moral agency.
- No universal theory of everything.
- No claim that morphology uniquely contains its full generating history.
- No claim that descriptor round-trip reconstructs original morphology.
- No physical CNT memory-device result.
- No claim that biology, molecular chemistry, cells, metabolism, or CNTM is required for MFM.
- No claim that the Formative Field is proven superior to graph, RAG, RETE, production-system, pub-sub, or content-addressable architectures.
- No claim that distributed embodiment proves one self or one identity across bodies.
- No universal ranking of generation, storage, conversion, or energy sources.
- No fixed claim that machine-speed developmental search guarantees a particular evolutionary span or wall-clock result.
- No machine-generated human-value score for knowledge, education, rights, or basic provision.