KELLY live pulserefreshed agolearninglifetime heartbeat 0browser pulses warmingcells 0replay-ready 0engine checkinglatest cell watchingevidence checkinglive memory laneloop: checking live activity
activity feed
Interactive lens

Hypothesis Lab · Answered directly by KELLY · public writes blocked

interactive cognition lens · answered by KELLY

Hypothesis Lab

KELLY's unknown-ledger, causal-gap, value-of-information, and experiment-planning organ.

Turns surprise, missing variables, and evidence gaps into falsifiable hypotheses with source plans, tests, closure targets, and memory mutation candidates.

real timeGlass Lab sourcedread-only window
Real-time cognition window

Samples Glass Lab, Advanced Processing Evidence Spine, and the lab-specific runtime routes every 15 seconds.

Glass Lab sourceReceiptspublic writes blocked

Glass panels

0/3

Snapshot

0/?

Routes

linked

Refresh

15s

Glass Lab source panels
discovery unknown ledgerpending
causal gap organpending
runtime gap organspending
Runtime route checks

This lab uses anchored Glass Lab proof routes only.

Latest snapshot: pending. Evidence spine: pending. The lab can expose KELLY-generated hypotheses. Public users cannot approve memory or GraphState writes.

Hypothesis & Theory Workbench

Test plans and falsifiers — hypotheses are not facts; memory_candidate ≠ memory write.

read_onlymulti_replica_ready=falseprocess_is_not_truth
open unknowns
0
active hyps
0
blocked
0
latest receipt
none
  • • Source plan + falsifier lanes stay inspect-only until durable candidates land.
  • • Closure never promotes truth without settlement + severe-test receipts.
  • • Public users cannot create Prime unknowns from this shell.
  • • Gauges follow GET event-bus status (process-local Map), not multi-replica truth.
  • • List API may serve process-local Map rows when non-empty; else demo_fallback — never multi-replica.
  • • CLI Map seed scripts write a separate process Map and do not hydrate this engine HTTP list.
  • • memory_candidate is a court-queue hint only — never memory_written / GraphState write.
  • • Falsifiers and source-plan are test-plan lists; closure pending is not truth promotion.

Event bus status

GET onlymulti_replica=falseprocess_local_map

Loading event-bus status…

Map stores status

GET onlymulti_replica=falseprocess_local_map

Loading map-stores status…

Organs glass chain

GET onlymulti_replica=falseladder=n/a

Loading organs chain…

process_is_not_truth · never uses terminal “blocked” · Glass mission control only

APES organ rollups

GET onlymulti_replica=falseorgans=n/a

Loading organ rollups…

instrument rollups · labeled_unproven · never demo-as-live · process_is_not_truth

Glass Lab visibility

GET onlymulti_replica=falsepanels=n/a

Loading visibility pack…

four panels · inspect-only · never terminal “blocked” · process_is_not_truth

Cognition Organs V1

GET onlymulti_replica=falsen/a%

Loading organs V1 completion…

ladder=n/a · neon_any=no · instrumented=no

envelope · neon-first labeled · no direct writes · process_is_not_truth

Operator command ladder

GET onlyauto_deploy=falseobserved=0

Loading operator ladder…

never auto-deploy · process_is_not_truth · missing_evidence not “blocked”

Program status

GET onlymulti_replica=false

Loading program status…

process_is_not_truth · never terminal “blocked” · public lens only

Organ honesty summary

GET onlymulti_replica=falseprocess_is_not_truth
  • Map stores process-local
  • CLI seed ≠ engine HTTP
  • Claim→hyp links not truth
  • Radar inspect-only
  • Severe-test inspect-only
  • Claim strength not settlement
  • memory_candidate ≠ written
  • Falsifiers not proof of kill
  • Closure pending ≠ promotion
  • lineage_view not write
  • birth_chain not settlement

Standing law checklist — never multi_replica_ready; never public write; process is not truth.

Cognition workbench

This is the sector-specific operating view for Hypothesis: gauges, receipt queue, closure timeline, topology, and empty-state behavior. Live route health remains in the real-time panel above.

operator grade

Unknown ledger

first-class gaps

Surprise, blind spots, missing causes, and missing sources become explicit objects.

Value of information

rank before chase

The lab prioritizes unknowns that can change forecasts, memory, or causal structure.

Experiment plan

source + test

Each hypothesis needs a source plan, falsifier, and closure target.

Closure

confirm / weaken / kill

Outcomes must be explicit before memory or GraphState promotion is considered.

Receipt queue

The queue is a receipt path, not a claim of hidden live data.

traceable

Lineage and closure timeline

1

Surprise

observed

Something changed more than expected.

2

Unknown

written

The gap becomes a ledger object.

3

Hypothesis

formed

KELLY proposes a falsifiable causal bridge.

4

Test

planned

Source plan, falsifier, and reality check are declared.

5

Closure

pending

Confirm, weaken, kill, defer, or promote.

Cognitive topology

Prediction miss

Forecast or model error that creates surprise.

Unknown

Missing variable, cause, source, or context.

Hypothesis

Falsifiable answer candidate.

Experiment

Observation plan that can change the answer.

Memory candidate

Only after closure and replay.

Viewport empty-state contract

No active unknown selected

If the unknown ledger has no current item, the lab should say what it is watching and keep source-plan links visible instead of showing fake hypotheses.

Open unknown ledger

Live inputs

Forecast errors, stale loops, contradiction pressure, and missing-source reports.
Science/Research falsifiers and external evidence packets.
Advanced Processing evidence-spine receipts and Glass Lab live snapshot panels.

Active objects

Unknown ledger entries
Active hypotheses
Causal bridge candidates
Value-of-information rankings
Experiment and source plans

Receipts

Hypothesis cards
Observation receipts
Falsifier receipts
Source-plan/test/closure receipts

Closure status

A hypothesis closes by confirm, weaken, kill, defer, or promote-to-memory-candidate outcome.
No hypothesis becomes durable memory without replay, settlement, calibration, and governed authority.

Automation loop

Continuously refreshes from surprise, blind-spot, unknown-ledger, and causal-gap signals.

cognition pipeline

step 1

Detect surprise or forecast miss from live reality and evidence spine.

step 2

Ingest self-inquiry answer seeds (assumptions, claims, falsifiers) — answers are not truth.

step 3

Write the unknown as a ledger object with value-of-information scoring.

step 4

Generate a causal bridge, falsifier, source plan, and closure target.

step 5

Send closed or weakened hypotheses to memory lineage and GraphState promotion review.

Glass Lab source

These are the full proof windows this lab samples into the one-page view.

Storage posture

Postgres

Use canonical KELLY Neon/Postgres with hypothesis-scoped tables/read models for unknown_ledger, hypotheses, source_plans, experiment_tests, and closure_receipts.

Object store

Use R2/object storage only for evidence packets, test bundles, and immutable closure receipts linked from Postgres.

Volume

No dedicated persistent volume required. Runtime state should be reconstructable from canonical receipts.

Runtime routes

Lab functions stay simple on this page; deeper receipts open in the exact runtime or Glass Lab window.

operator links