Roadmap | Updated September 30, 2026
Built, underway, and not built yet.
The next closure target is not less imagination. It is an auditable boundary between proposal space and durable formed state.
Established
- Fractalish public site, field guides, documents, study guides, and collaboration hub.
- Public Fractalish repo.
- Public Cognitive Basin platform repo.
- Cognitive Basin / BasinLab technical release and implementation disclosure.
- Natural Math v5 frozen integer baseline and separately governed experimental branches.
- R2R v0.1.3 trust-anchor and control-plane hardening branch, with production trust enrollment still incomplete.
- CNTM realization materials.
- Motorola activation evidence and AIA materials.
- Fractalish AI public overview.
Near-term priorities
- Run NLM Boundary Discriminator 001 without defining learning in advance.
- Freeze a CRR-positive base, then vary candidate learning offices independently.
- Specify proposal-space and formed-state interfaces, receipts, authority bounds, and retraction behavior.
- Map the Fractalish v1 event, Basin, CRR, CF-Logic, and promotion lifecycle onto existing R2R offices.
- Demonstrate actual causal condition separately from certifiable historical provenance.
- Test the morphology stop rule rather than preserving learning, intelligence, or consciousness by definition.
- Keep probabilistic model capability and compare governance overhead against native controls.
- Additional sensor and device experimentation on the Motorola path.
- Better public contribution packaging for code, evidence, and serious critique.
Longer-term work
- Dedicated public repos for additional source collections if and when publication posture supports them.
- Matched tests of exploratory breadth, unauthorized persistence, revision quality, and lifecycle cost.
- Stronger morphology-vs-generation benchmarks.
- SymLan token criteria after descriptor stabilization.
- InfinitySight perception work and richer APTD device embodiment.
- Deeper native activation work beyond stock Android groundwork.
HOLD: NLM is not frozen; the Fractalish v1 closure TODO named by the v0.12 concordance was not present in the searched local corpus; and no current result establishes solved hallucination, learning, intelligence, or consciousness.