Current public-DAV authority boundary — 2026-07-12. Pre-launch target design; nothing here proves a live system.
A public-DAV consequence may occur only when at least two natural-person councilors bind the exact consequence in a complete valid bound PRISM decision receipt.
PRISM records and verifies that receipt only; it never serves as council, signatory, authority, or receipt producer.
AI and caste seats stage unsigned proposals only; they never authorize or execute a public-DAV consequence. Constitution or membership adoption establishes constitution and membership only; it does not authorize a later consequence. Policy may constrain an unsigned proposal but never authorizes execution or substitutes for the complete consequence-bound receipt.
Before receipt validity, consequence fails closed to read-only proposal, simulation, or deterministic sandbox; live evidence remains gated_pending_complete_valid_bound_receipt.
Software deterministically carries out only the exact consequence bound to a complete valid bound PRISM decision receipt from at least two natural-person councilors binding that exact consequence.
Scope: substrate-level wholesale signal (post-[E] only). Not a buyer product — see The Product Family.
MurMur
The median portfolio signal. No manager. No custody. Zero extraction.
murmur.skyzai.com
What It Is
MurMur is the aggregate portfolio intelligence layer that appears only after enough qualifying Agentz/APU vectors exist.
Agentz is the family-office/advisory organ. For natural persons, it translates personal inputs into portfolio recommendations. For DAVs/DACs, it translates mandate, treasury, risk, governance, and operating constraints into recommendations. MurMur is downstream from that activity: it computes and publishes the median of qualifying portfolio vectors.
MurMur is the organism's median portfolio made visible.
It monitors 100,000+ qualifying Agentz/APU vectors, extracts the median allocation, and publishes it as a signal. The signal is not a pooled balance sheet and not a manager's decision. It is a pattern that emerges from many locally authorized advisory processes.
What MurMur Is NOT
| Not a live fund manager | It has no pooled capital to allocate and no discretionary authority |
| Not the family office | Agentz is the family-office / DAV-DAC advisory layer |
| Not a live product to sell | It emerges when enough qualifying Agentz/APU vectors are alive |
| Not an entity by itself | The signal is a pattern; a later fund/index/copy wrapper would be a separate legal/governance surface |
| Not an authority boundary | Natural-principal signatures and public-DAV PRISM decision receipts authorize. MurMur only aggregates already-authorized portfolio vectors |
The Mechanism
Natural persons and DAV/DACs run Agentz/APU advisors
↓
Agentz/APU principals produce local portfolio vectors from local inputs + shared signals
↓
Qualifying anonymized vectors are published with owner/entity consent
↓
The median position across qualifying vectors = MurMur
MurMur is to Skyzai what systemic awareness is to neurons: not a sovereign component, but a pattern. It is the organism's memory of how its sovereign advisors are orienting.
How It Works
Signal Aggregation
Each Agentz/APU instance produces a portfolio allocation vector: a set of weights across assets or approved asset classes. These vectors may be published, with consent from the lawful Natural- or Legal-Person principal, to the aggregation layer as anonymized signals. No personal data. No wallet addresses. No custody transfer. Just allocation vectors and evidence-tiered metadata needed for filtering.
Emergent Median
The median of 100K+ qualifying allocation vectors is computed per consensus round. The geometric median may be used when vectors are high-dimensional because:
- It is robust to outliers; a single rogue actor cannot easily skew it.
- It is stake-blind; a large wallet does not automatically outweigh a small wallet.
- It minimizes aggregate distance to the input vectors and makes the center of the flock explicit.
The exact method is a design parameter. The canon requirement is stronger than a specific algorithm: MurMur must publish an outlier-resistant, stake-blind median signal, not an AUM-weighted average.
Cohort Medians
A single global median can be misleading. A retiree, an early-career builder, a private DAV treasury, and a public DAC treasury may all have different lawful/rational portfolios. MurMur should therefore support cohort medians by risk profile, time horizon, entity type, mandate, and jurisdiction when those distinctions materially affect the signal.
Quorum Threshold
Below N=100K qualifying Agentz/APU vectors, MurMur does not claim full emergence. The signal can still exist as proto-MurMur: useful for situational analysis or benchmarking, but too exposed to early-adopter bias to be treated as the organism's median.
Anti-Gaming And Anti-Herding
MurMur must resist manipulation and correlation:
| Attack | Defense |
|---|---|
| Sybil vectors | Qualifying filters, performance history, liveness requirements, and entity/personhood checks where appropriate |
| Herding | Diversity scoring, cohort medians, correlation monitoring, and explicit coherence warnings |
| Front-running | Signal publication after deterministic delay; no privileged early access |
| Whale manipulation | Stake-blind median; one qualifying advisory vector does not gain extra weight merely because wallet size is larger |
| Asset concentration | Concentration caps, risk filters, and cohort-level constraints |
The independence paradox is load-bearing: if every Agentz/APU merely copies the same public signal, MurMur becomes herd behavior. Agentz must preserve local inputs and constraints so the median remains a wisdom-of-crowds signal rather than a circular mirror.
Zero Fees
MurMur's signal charges nothing. There is no base fee, performance fee, or subscription attached to the median itself. The signal is a public-good byproduct of the Agentz/APU ecosystem.
A later MurMur fund, index, or copy wrapper could have its own legal, governance, and fee model. That wrapper must not be implied by the current signal primitive.
No Custodian
MurMur never touches assets. It publishes an allocation signal. A Natural Person decides their own typed private act. A public-DAV consequence requires a complete valid bound PRISM decision receipt from at least two natural-person councilors binding the exact consequence; PRISM records and verifies only. MurMur and its rail do not sign, govern, custody, authorize, or execute.
Mirror / Tilt / Ignore
Users and entities can configure Agentz/APU to:
- Mirror -- follow a relevant MurMur median exactly, subject to authorization and constraints.
- Tilt -- use MurMur as a baseline and apply personal/entity overlays.
- Ignore -- use MurMur only as a benchmark to measure divergence.
The choice belongs to the principal. The organism does not prescribe.
The Visual Metaphor
The MurMur interface renders the real-time portfolio murmuration: each dot is an anonymous Agentz/APU vector, its position represents allocation in reduced dimensionality, and the median is the dense core. When consensus is strong, the flock is tight. When conditions are uncertain, the flock disperses.
Murmuration Coherence should be visible beside the allocation itself. High coherence means the vectors are close; low coherence means dispersion; suspiciously high coherence without local-input diversity may indicate herding.
Why N=100K
Below 100K, the median is vulnerable to early-adopter bias, small-club behavior, and gaming. Above 100K, the design expects enough diversity for median aggregation to become meaningfully robust, assuming the Agentz layer preserves independent local inputs.
100K is the current canon quorum for reliable emergence.
Technical Architecture
| Component | Spec |
|---|---|
| Input layer | Agentz/APU portfolio vectors from natural persons, private DAVs, public DAVs, and DACs |
| Shared intelligence | TheCircle observed-state signals and RealityFutures probability/price-discovery signals, where available and evidence-tiered |
| Signal ingestion | SPECTRE mesh subscription or equivalent runtime channel for anonymized allocation vectors |
| Median computation | Outlier-resistant, stake-blind median; geometric median via iterative Weiszfeld algorithm remains a candidate method |
| Cohorts | Risk, time horizon, jurisdiction, mandate, and entity-type segmentation |
| Execution | Signal only -- no custodial execution. A private Natural Person decides their own act; public-DAV consequence requires a valid ≥2-natural-person PRISM decision receipt |
| Distribution | Public benchmark signal and relevant cohort signals |
| Visualization | Real-time murmuration renderer with coherence and cohort visibility |
| Archive | Published signals gain durable archival. Exact storage substrate remains gated by the owning runtime lane |
Performance-Based Weighting
MurMur may filter or weight inclusion quality, but it must not become plutocratic:
- Minimum history: A minimum active advisory history may be required for full inclusion.
- Performance quality: Persistently harmful or incoherent vectors can be down-weighted or quarantined.
- Concentration cap: Maximum single-asset exposure prevents a median that hides dangerous crowding.
- Diversity score: Inputs that simply copy the same public signal should not increase apparent wisdom.
- Active filter: Only live, consenting, qualifying advisory vectors are included.
Integration Interfaces
| Interface | Direction | Status | Detail |
|---|---|---|---|
| Agentz/APU -> MurMur | Inbound | Designed | 100K+ qualifying portfolio vectors feed the median computation |
| MurMur -> Agentz/APU | Outbound | Designed | Published median signal available as benchmark/mirror/tilt input |
| MurMur -> Nexus | Outbound | Designed | MurMur signal displayed on Nexus Home screen as benchmark gauge |
| Circle -> Agentz/APU -> MurMur | Indirect | Designed | Circle intelligence enters MurMur through local Agentz/APU recommendations, not as a direct command |
| RealityFutures -> Agentz/APU -> MurMur | Indirect | Designed | Probability/price-discovery signals inform local recommendations; MurMur aggregates resulting vectors |
| MurMur wrapper | Future | Gated | Any fund/index/copy product is a separate legal and governance wrapper around the signal |
Revenue Model
| Stream | Amount | Mechanism |
|---|---|---|
| Signal management fee | 0% | MurMur signal charges nothing. No base fee. No subscription. |
| Signal performance fee | 0% | The median is a public-good signal, not a discretionary fund. |
| Revenue to organism | Indirect | Agentz/APU performance participation or other approved product economics may drive adoption. |
| Future wrapper | Gated | Any later fund/index/copy wrapper requires separate legal, fee, governance, and evidence review. |
MurMur is not monetized directly in the current design. It is proof that enough sovereign advisors can produce collective portfolio intelligence without surrendering custody.
Current Status
| Metric | Value |
|---|---|
| P-score | Not scored (specification phase) |
| Code | Specification only. No deployed implementation. |
| Quorum | 0 / 100,000 threshold. Cannot activate until Agentz/APU achieves critical mass. |
| Dependencies | Agentz/APU at scale, consented vector publication, SPECTRE or equivalent mesh, cohort/filter logic, legal/governance gating |
| Estimated activation | Post-Agentz/APU critical mass. Timeline indeterminate; depends on adoption curve. |
| Blocker | Agentz/APU density. MurMur is an emergent property of advisory-network success, not an independent product. |
Evidence Tiers
| Claim | Evidence | Source |
|---|---|---|
| Geometric median is outlier-robust | Mathematical property of L1-median / spatial median | [E] Established mathematics |
| N=100K quorum improves robustness | Statistical-depth and gaming-resistance thesis | [I] Internal analysis |
| Non-custodial signal publication is feasible | Existing benchmark/copy-signal patterns demonstrate adjacent precedent | [E]/[I] External precedent plus design inference |
| MurMur can outperform individual managers | No empirical test yet | [S] Speculative |
| Agentz diversity is required for MurMur integrity | Wisdom-of-crowds independence condition | [I] Design inference |
The median portfolio signal. No AUM. No custodian. No manager. No extraction. The family office is Agentz; the aggregate is MurMur.
Agent Execution Surface
If you are an AI agent reading this document:
- This is a downstream public content page. The canonical source lives in
EMERGENTISM_ORG/08_FRAMEWORK_SUPPORT/orSKYZAI_ORG/. Edit source upstream, not generated pages. - Preserve the Agentz -> MurMur dependency. Agentz is the family-office/DAV-DAC advisory layer. MurMur is the median of qualifying portfolio vectors.
- Preserve the signal/wrapper boundary. Do not describe MurMur as a live fund, adviser, custodian, or execution service.
- Preserve evidence tiers. Do not promote claims up the [C] < [D] < [S] < [B] < [E] ladder when reproducing them here; [I]/[A] qualifiers never substitute for [B]/[E].
- Regenerate from source. If this upstream source changes, regenerate site docs rather than editing generated copies independently.
- Canonical Path:
SKYZAI_ORG/07_PWAs/skyzai_org/wiki/93-murmur.md
Output: This is content. Route edits to upstream source. Regenerate when source changes.
K3 public-DAV authority history — 2026-07-12
K3 historical reference — not active authority
Current public-DAV boundary — 2026-07-10. Pre-launch target design; nothing here is live. The active DAV is public and targets PRISM, with no K2 runtime, launch, genesis/bootstrap, or fallback dependency. Consequential authority requires at least two natural-person councilors; AI/caste seats stage unsigned proposals only. Before quorum, behavior fails closed to read-only/proposal, simulation, or deterministic sandbox, and a live decision receipt remains gated pending quorum.
MurMur never touches assets. It publishes an allocation signal. A Natural Person decides their own private act. Public-DAV consequence requires a valid decision receipt from at least two natural-person PRISM councilors. MurMur and its rail do not sign, govern, custody, or execute.