PLATFORM / SYSTEM ARCHITECTURE

One machine. Search cannot certify itself.

Kairos Core is the proposed cross-domain product layer that generalizes the strongest research-control ideas already developed in the Heatmap Strategy Lab and XG Engine. Candidate generation explores. Evidence governs advancement.

AUTHORITY MODEL / 02

Deployment is a monitored state, not a finish line.

System
Kairos Core
Status
Specified / schematic
Customer data
None shown
01 Proposed 02 Researching 03 Historically supported 04 Forward observation 05 Decision-eligible 06 Approved 07 Deployed 08 Monitored 09 Rejected 10 Retest 11 Degraded 12 Suspended 13 Retired
01 / 08 Proposed

A bounded claim enters with intended use, source, and stop conditions.

AUTHORITY / NONE
AUTHORITY INCREASES ONLY THROUGH EXPLICIT STATE

This is a specified Kairos Core governance model. It does not represent live customers, deployed models, or operational telemetry.

A claim progresses from proposed through research, historical support, forward observation, decision eligibility, approval, deployment, and monitoring. Rejection and underpowered evidence can return to retesting. Monitoring can degrade, suspend, or retire a deployed system.

  1. 01

    Source World

    Authorized data with provenance, policy labels, and effective-time / available-time / ingest-time semantics.

  2. 02

    Evidence Membrane

    Canonicalize what the system may know before candidate generation begins.

  3. 03

    DecisionSpec

    Compile the target, information set, objective, constraints, baselines, evidence requirements, authority, and stop conditions.

  4. 04

    Evolutionary Research Engine

    Generate, mutate, recombine, cull, revive, and preserve diverse candidate systems across representations.

    NULL COMPASSnegative knowledgeREFEREEfalsificationQD ARCHIVEdiversity memory
  5. 05

    Forward Observer

    Freeze survivors and accumulate prospective evidence. Historical success cannot silently substitute for forward behavior.

  6. 06

    Decision State Machine

    Separate finding outcome, evidence maturity, and advancement authority across reject, retest, observe, advance, suspend, and retire.

  7. 07

    Action Gateway

    Recommendations and external execution remain separate permissions. Consequential action is explicitly authorized.

MODEL ROUTING / BOUNDARY CONDITIONS

No foundation-model provider is the evidence authority.

LOCAL_ONLY

Sensitive work stays local.

Customer-controlled or Kairos Node inference for data and tasks that must not leave the local trust boundary.

PRIVATE_ALLOWED

Approved private infrastructure.

Local models plus approved private-cloud services inside the organization’s policy boundary.

FRONTIER_ALLOWED

Frontier capability, selectively routed.

Permitted providers may support approved tasks without becoming the system of record or evidence authority.