Daylight-filled precision industrial manufacturing bay with a skilled master welder operating a 6G orbital GTAW station connected to hydrostatic testing and turbine equipment
01 · THE LIVING FACTORY (OPERATING)
Full-Screen 3D Hall →
SOURDOUGH PARADIGM · 2027 EXHIBITIONDOCTRINE: ESTABLISHED · SIMULATION ≠ PROOF

STOCKPILES RUN OUT. STARTER CULTURES DON'T.

The Sourdough Paradigm explores how essential human capabilities can survive disruption, move between people and machines, and regenerate without depending on a single controlling center.

● ALL CAPABILITIES & DEPENDENCIES INTACT

Master practitioner actively holds the 6G orbital keyhole (ΔK = PASS), feeding certified manifolds to hydrostatic testing and turbine dispatch.

CELL-01 · CONJUNCTIVE TPC GATESTPC(A) = 1 (PASS)
ΔE
PASS
ΔM
PASS
ΔG
PASS
Δτ
PASS
ΔK
PASS
ΔV
PASS

Click any optical station marker in the factory view to inspect its physical dependency role, or click ENTER THE LIVING SYSTEM to observe capability loss and regeneration.

Living Culture Conservatory · Distributed Capability Network

PROPOSED· Interactive Network Model

The culture is the interface.

When an essential capability disappears, the distributed network detects the loss, exchanges a Universal Capability Representation (UCR) starter packet, trains local replacements, tests competence across the 6 conjunctive TPC gates, and attempts regeneration. Touch any cell below or explore the dedicated /living-culture and /living-paradigm conservatories.

LIVE CULTURE STATE · ILLUSTRATIVE, NOT EXPERIMENTAL PROOF
GATES: PASS · Tick 0
PACKETS SENT
0
CULTUREthe packet travels,the expert does notA · FUNCTIONALSpecificationsB · FUNCTIONALEquipmentC · FUNCTIONALOperatorsD · FUNCTIONALMaterialsE · FUNCTIONALVerification

QRCI (PROPOSED, not demonstrated): on node loss, replacement cells are recruited while redundant nodes are inhibited to prevent resource contention. The map above demonstrates packet travel, training, and verification — QRCI recruitment itself is not yet a demonstrated mechanism.

Touch any node to trace live regeneration propagation.

Controls

Failure path first, then success path. Every button changes real state.

A: FUNCTIONAL · training 100% · competency 100% · verified yes

B: FUNCTIONAL · training 100% · competency 100% · verified yes

C: FUNCTIONAL · training 100% · competency 100% · verified yes

D: FUNCTIONAL · training 100% · competency 100% · verified yes

E: FUNCTIONAL · training 100% · competency 100% · verified yes

Eight stages (derived from state)

  • IDENTIFY — WAITING
  • DECOMPOSE — WAITING
  • COLLECT — WAITING
  • PACKAGE — WAITING
  • DISTRIBUTE — WAITING
  • RECONSTRUCT — WAITING
  • VERIFY — WAITING
  • PROPAGATE — WAITING

TPC gates

COMPILER PASS

Gate order: ΔE Energy, ΔM Material, ΔG Geometry, Δτ Time, ΔK Competence, ΔV Verification. One FAIL blocks; passing gates cannot compensate.

Event log

  • [tick 0] [info] Network initialized. Five complementary capability nodes connected. No regeneration claim has been made.

SMC / Somatic Symbiosis are cross-tier mechanisms, not tiers. DCIA is PROPOSED and separate. CIP level is not assigned here. Documentation does not establish competence; simulated success does not establish physical feasibility.

Core Epistemic Boundary · Documentation vs. Competence

ESTABLISHED· A Blueprint Cannot Strike an Arc

What can be written down vs. what has to be taught and verified.

Static preservation vitrines showing alloy ingots, CAD blueprints, and idle hardware■ RECORDED ARTIFACTS · STATIC

1. Static Stockpiles, Blueprints & Untended Iron

Stockpiles buffer supply for weeks and then hit zero. Digital CAD files preserve geometry (ΔG) but carry zero tacit operator skill (ΔK = 0). Untended machines halt as soon as material heats drift.

Living Culture Conservatory with apprentices training in glass workshops around a distributed network table▨ LIVING CULTURE · DEMONSTRATED & VERIFIED

2. Living Starter Culture (Human → Machine → Human)

Captures Boneyard Specifications, Sensory Maps, Drill Regimens, and Proof Gates in a UCR packet so sister workshops can train local practitioners and independently verify physical output (L4).

Interactive Dependency Topology — Illustrative Comparison

PROPOSED· Conceptual Resilience Model

What breaks first when a critical dependency vanishes?

Select a real-world fracture below. Compare single-point dependency against distributed capability regeneration. Copying files never counts as regenerating a physical process.

Active Fracture Scenario: A critical physical skill exists in only one operator's hands.

A — Conventional Single-Point Dependency

SYSTEM HALTED
SRC✕OUT

The operation stops. Tacit tolerance feel and setup judgment leave with the operator — no verified transfer or independent demonstration exists.

Result: Unverified backup · Zero independent reproduction

B — Sourdough Distributed Regeneration

CAPABILITY REROUTED
A/BC/DUCRΔVL4✓

Sensory cues, setup drills, and proof gates were captured in a UCR packet and independently demonstrated across sister nodes before the loss.

Result: Complementary nodes reconstruct & independently verify output

Origins & Evolution · Deter → Defend → Regenerate

ESTABLISHED· Historical Lineage & Stated Extension

From Deterrence to Survivability to Distributed Regeneration

The Sourdough question extends a strategic progression — as an original philosophical extension by Michael Crumly, not an official successor to Singaporean defence doctrine. Select any milestone below to inspect its historical basis, limits, and principle.

2025 — Sourdough Paradigm (Michael Crumly, Canada)

REGENERATE — Resilience through distributed regeneration.

Learn (Historical Context): An original philosophical extension (distinct from Singaporean defence doctrine): can essential capabilities be preserved, transferred, reconstructed, and regenerated elsewhere when their original source disappears?

Structural Limits: Governing doctrine is ESTABLISHED as stated philosophy; the technical stack (UCR, UCA, TPC, CIP, PAC, DCIA, QRCI) is PROPOSED, and open problems R1 and R2 remain UNRESOLVED.

Apply (What Carries Forward): Shifts resilience from protecting stockpiles to propagating starter cultures — independent reproduction verified by physical evidence.

Historical provenance note: Pak Shun Ng's 2005 paper, From “Poisonous Shrimp” to “Porcupine”, is a later historical analysis of Singapore's transition — not a fourth doctrinal stage and not the origin of the Porcupine concept. Biological metaphors are sources of inspiration, never empirical proof of computational or industrial architectures.

Learn the Philosophy · Preserve → Reproduce → Connect → Coordinate

ESTABLISHED· Conceptual Distinctions

Preserve the ability to reproduce, not just the product.

Protecting physical assets, archiving documents, transferring skill, and regenerating a working capability across independent nodes are distinct steps. Documentation alone never equals operational competence.

STAGE 01

PRESERVE

Capture specifications, processes, sensory cues, and boundary conditions before the original source disappears.

STAGE 02

REPRODUCE

Transfer capability to new people, workshops, and machines. Independent physical performance must be demonstrated.

STAGE 03

CONNECT

Organize complementary capabilities into explicit dependency architectures with verifiable interfaces.

STAGE 04

COORDINATE

Investigate how independent nodes cooperate, allocate resources, and reconcile physical state without a single permanent centre.

Signature Interactive Workbench — Illustrative Educational Model

PROPOSED· 8-Stage Regeneration Model

Regenerate a Lost Capability: Load-Bearing Steel Bracket Scenario

An essential manufacturing workshop is gone. Work through the 8 stages and toggle distributed Nodes A–E below to see where regeneration fails when equipment, material expertise, or independent testing is missing.

Stage 1/8: Identify

Objective: Define the essential load-bearing function that must be recovered.

If Missing: A clear output specification (load rating, geometry, duty cycle).

Failure Mode: Rebuilding the wrong bracket well — a part that fits nothing.

Evidence Required to Proceed: Written output specification locked before decomposition.

Distributed Capability Network (Nodes A–E) & Verification Gate

Toggle any node off to inspect system degradation

Distribution is not copying files to five servers. It is connecting complementary physical capabilities across independent nodes so no single node failure kills the craft — and no capability is declared regenerated until Node E independently verifies physical output.

Execute Physical Verification Steps (Requires Corresponding Nodes Online):

CURRENT EVIDENCE STATE: L2 — Ready for Reconstruction (Complete Dimensional, Material & Operator Checks)
Long-Horizon Progression: Human → Machine → Human → Machine → Machine → Autonomous Regeneration.SPECULATIVE· Long-Term Research Direction

Interactive DCIA & Sourdough Workbench — Canonical Spec, Nov 1, 2026

Stockpiles run out. Starter cultures don't.

Trade with everyone. Depend on no single one. Scrub the 90-day supply cutoff and watch the two models diverge.

MODEL A — STATIC STOCKPILE

50%

Depletes to 0% by day 90. No regeneration.

MODEL B — SOURDOUGH STARTER (UCR + TPC + CIP)

82%

Regenerates toward 100% capacity. Illustrative model, not experimental proof.

CANONICAL 5-TIER PAC STACK — UCR ≠ UCA

Universal Capability Representation. UCR ≠ UCA. UCR is the packet: what the capability is, written down. It is not the architecture that moves it.

UCR 4-BLOCK PACKET INSPECTOR (tied to the Steel Bracket Workbench below)

Block 1 — Boneyard Specification. Nominal specs, salvaged substitutes, offsets, contamination limits.

TPC 6-GATE FEASIBILITY CALCULATOR — TPC(A) = ∏ vᵢ, i = 1..6

Tap a gate to cycle PASS → FAIL → UNRESOLVED. One FAIL zeroes the product. One UNRESOLVED (with no FAIL) leaves it conditional.

TPC(A) = 1 — PASS. ΔV is LOCKED while any other gate fails — verification cannot pass a broken compile.

CIP MATRIX — MATURITY (C₀–C₄) × EVIDENCE (E₀–E₅), KEPT SEPARATE

Selected: C2 Assisted reproduction with E2 Recorded evidence. High maturity with weak evidence is not the same claim as low maturity with strong evidence — the two axes do not collapse into one score.

G.W.A.P. OS: field interface for the packet on the shop floor. Larogic: Ancords, Lararium, Kinfract — evidence, not impersonation.

FALSIFICATION LABS — R2 LEASE-AUTHORITY & R1 BRANCH ABANDONMENT

Full R1/R2 simulators also live at /preview/visual-sprint-1 and /preview/interactive-modules.

CANONICAL COMPONENT REGISTRY

  • ■ ESTABLISHED DOCTRINE — UCR, UCA, TPC (six gates), CIP C₀–C₄ vs E₀–E₅, PAC Tier 5 as architecture
  • □ PROPOSED — PAC as planetary runtime, DCIA, QRCI (recruit replacements on loss, inhibit redundant nodes — proposed, not demonstrated)
  • ● UNVALIDATED RESEARCH — SMC / Somatic Symbiosis
  • ⊠ EXCLUDED — GKS, ASM, PHA, DACA (excluded from the top level by the canonical registry)

Interactive Hardware & Competence Modules

Test the 6-Gate Breaker, UCR Packet Assembler, and γ-Assist Fader

VIS-03 // THERMODYNAMIC PHYSICAL COMPILER (TPC)

Six Physical Gates: trip any switch

Conjunctive Physical Law: All six gates (ΔE, ΔM, ΔG, Δτ, ΔK, ΔV) must be satisfied. One failed gate blocks the operation; an unresolved physical condition remains unresolved.

COMPILER PASS // EXECUTION PERMITTED
GATE 01PASS
ΔE Energy
165A DCEN / 18.4V
GATE 02PASS
ΔM Material
SA-516 Gr.70 + ER70S-2
GATE 03PASS
ΔG Geometry
37.5° Bevel / 2.4mm Root
GATE 04PASS
Δτ Time / Rate
Interpass ≤ 175°C / 2.1 mm/s
GATE 05PASS
ΔK Human Competence
Unassisted γ = 0.0 Qualified
GATE 06PASS
ΔV Verification
Bridge-Cam + NDT Pass

VIS-01 // UNIVERSAL CAPABILITY RECORD (UCR)

UCR Four-Block Assembler: what travels

Uncheck any block and try to transmit. UCR Packet (Boneyard Specification, Sensory Map, Drill Regimen, Proof Gate) transfers across the network. The package travels; the expert does not. The receiving facility must still calibrate local materials, train its local operator, and pass local ΔV inspection.

[READY]: All 4 UCR blocks present. Ready to transmit capability package to the receiving facility.

VIS-06 // SOMATIC SYMBIOSIS & PROGRESSIVE ASSISTANCE

The γassist fader: assisted success ≠ competence

ADAPTIVE ASSISTANCE (γassist):γassist = 0.7

Assisted execution at γassist = 0.7. Competence is NOT yet established.

Research Architecture · Five Canonical Tiers

PROPOSED· Canonical Capability Stack

Five Tiers, One Job Each — Grounded by the Thermodynamic Physical Compiler

Each layer has a strict boundary. Software representations (UCR, UCA) cannot bypass physical feasibility (TPC) or claim verified maturity (CIP) without measured shop-floor evidence.

Tier 3 Interactive Gate · Binary Conjunctive Feasibility Check

TPC(A) = ∏ vᵢ = ΔE × ΔM × ΔG × Δτ × ΔK × ΔV

TPC(A) = 1 → PASS (FEASIBLE)

Surplus in one gate cannot compensate for a zero in another. Toggle any vector below to verify the non-compensatory rule:

Tier 1PROPOSED

UCR

Universal Capability Representation

Boneyard Spec, Sensory Map, Drill Regimen, Proof Gate.

Tier 2PROPOSED

UCA

Universal Capability Architecture

Dependency graph connecting skills, tools, and interfaces.

Tier 3PROPOSED

TPC

Thermodynamic Physical Compiler

Six-gate physical feasibility check: PASS, FAIL, UNRESOLVED.

Tier 4PROPOSED

CIP

Capability Into Progress

Maturity (C0–C4) tracked separately from Evidence (E0–E5).

Tier 5PROPOSED

PAC

Planetary Autonomous Compiler

Network-scale viability and recovery (Proposed, not runtime).

Connecting agents does not automatically produce collective intelligence

PROPOSED· DCIA & QRCI Specification
QRCI [PROPOSED]: Quorum-Regulated Capability Recruitment and Inhibition inside DCIA. Biological quorum sensing is inspiration, not proof; signal thresholds, hysteresis, and adversarial delay bounds must be formally tested.
G.W.A.P. OS [PROPOSED]: Guy With A Phone OS field interface (gwap-os). Puts drills and evidence capture on the shop floor without replacing the operator.
Larogic [PROPOSED]: Accountability ledger (Ancords, Lararium, Kinfract). A blueprint cannot strike an arc. Evidence, not impersonation.

Falsification Before Endorsement · Two Bounded Open Problems

UNRESOLVED· Open for Formal Counterexamples

Don't endorse the architecture. Break it.

We invite researchers, distributed-systems engineers, and industrial architects to falsify two explicit open problems with a counterexample, formal model, failure trace, or experiment (under 2 pages).

R1 — Physical Workflow Branch Reconciliation

UNRESOLVED

Checkpoint DAGs merge digital branches by union, keeping conflicting branches as explicit siblings. But merging digital state cannot undo an irreversible physical cut, weld, or shipment.

Event Lifecycle State Machine:

PROPOSED → AUTHORIZED → ISSUED → ACKNOWLEDGED → EXECUTED → MEASURED → RECONCILED

FAULT TRACE: Branch A reached EXECUTED (steel cut) during partition while Branch B authorized conflicting geometry. Digital merge leaves an unrecoverable physical conflict — state held UNRESOLVED for human/physical review.

Submit R1 Counterexample / Failure Trace →

R2 — Lease-Authority Consensus & Fencing

UNRESOLVED

DCIA specifies a lease-authority set (LA): E1 single writer per resource, E2 strictly increasing fencing epoch, E3 minority-failure tolerance, E4 joint-consensus membership change.

Partition Trade-off (Safety vs. Liveness vs. Availability):

Receiving physical resource MUST reject stale epoch tokens (e < e_max).

SAFETY PRESERVED: Delayed packet from stale lease holder (epoch 11) arrives after new quorum grant (epoch 12). CNC controller rejects epoch 11.

Submit R2 Counterexample / Failure Trace →

Research Integrity & Searchable Epistemic Registry

One Acronym · One Meaning · History Preserved

What is Established, Proposed, Speculative, Unresolved, or Superseded?

Logic owns editorial rights. Evidence determines acceptance. Reproducibility establishes verification. No proposed threshold or AI output is promoted to an established result without independent physical reproduction.

[Doctrine]Sourdough Paradigm
ESTABLISHED

Governing Doctrine & Capability Regeneration Philosophy

Stated governing doctrine: preserve the ability to reproduce essential capabilities across independent nodes rather than stockpiling static products.

Falsification / Boundary: Demonstrating that static stockpiles or centralized documentation alone reliably regenerate complex physical production after total primary-site loss.

[Tier 1]UCR
PROPOSED

Universal Capability Representation

Individual capability packet: Boneyard Specification, Sensory Map, Drill Regimen, and Proof Gate. Possessing documentation is not competence.

Falsification / Boundary: Showing a UCR packet fails to transfer reproducible physical competence across independent operators when all 6 TPC gates are satisfied.

[Tier 2]UCA
PROPOSED

Universal Capability Architecture

Structural grammar connecting capabilities, dependencies, resources, skills, processes, and network topology.

Falsification / Boundary: Cyclic or unresolvable dependency deadlocks during multi-node capability composition.

[Tier 3]TPC
PROPOSED

Thermodynamic Physical Compiler

Physical feasibility gate: TPC(A) = ∏ vᵢ across ΔE (energy), ΔM (material), ΔG (geometry), Δτ (time), ΔK (competence), and ΔV (verification).

Falsification / Boundary: Any action passing TPC(A) = 1 that violates physical conservation, geometric tolerance, or operator safety bounds on the shop floor.

[Tier 4]CIP
PROPOSED

Capability Into Progress

Dual-scale progression tracking Capability Maturity (C0–C4) strictly separated from Evidence Quality (E0–E5 / L0–L4). Calibration criteria remain proposed.

Falsification / Boundary: Promoting a capability to C4 maturity on architectural plausibility (E0/E1) without independent physical reproduction (L4).

[Tier 5]PAC
PROPOSED

Planetary Autonomous Compiler

Long-horizon network-scale viability, recovery, and adaptation layer. Proposed research architecture — not a demonstrated planetary runtime.

Falsification / Boundary: Inability to bound coordination overhead or resolve physical branch conflicts (R1) across distributed nodes.

[Collective]DCIA
PROPOSED

Distributed Collective Intelligence Architecture

Proposed coordination architecture without a permanent central controller. Requires formal specification, implementation, and empirical validation.

Falsification / Boundary: Failure to maintain safety under network partition or split-brain lease authority (see Open Problem R2).

[Collective]QRCI
PROPOSED

Quorum-Regulated Capability Recruitment and Inhibition

Proposed mechanism inside DCIA inspired by biological quorum sensing to regulate node recruitment and contention inhibition. Analogy is inspiration, not proof.

Falsification / Boundary: Oscillation, starvation, or spoofed recruitment signals under delayed, duplicated, or adversarial message conditions.

[Field Interface]G.W.A.P. OS
PROPOSED

Guy With A Phone OS — Field Interface

Shop-floor interface putting capability records, drills, diagnostics, and evidence capture where physical work happens.

Falsification / Boundary: Field workflow friction that degrades operator safety or mistakes mobile form entry for verified physical output.

[Accountability]Larogic
PROPOSED

Accountability & Evidence Ledger (Ancords, Lararium, Kinfract)

Forensic accountability layer enforcing 'A blueprint cannot strike an arc' and 'Evidence, not impersonation.'

Falsification / Boundary: Allowing unverified AI-generated claims or unmeasured digital state to register as completed physical evidence.

[Open Problem]Open Problem R1
UNRESOLVED

Workflow Branch Abandonment Under Physical Execution

Checkpoint DAGs merge by union, keeping conflicting branches as explicit siblings. Digital merge cannot undo an irreversible physical cut, weld, or shipment.

Falsification / Boundary: A trace where diverging branches execute conflicting physical actions before reconciliation, leaving an unsafe or unrecoverable physical state.

[Open Problem]Open Problem R2
UNRESOLVED

Lease-Authority Consensus Correctness & Fencing

Single-writer lease authority (E1–E4) with strictly increasing fencing epochs enforced by the receiving physical resource.

Falsification / Boundary: A trace producing two concurrent valid leases, same-epoch grants, or stale-holder execution on a resource unable to enforce fencing.

[Human Layer]SMC / Somatic Symbiosis
UNRESOLVED

Human Capability Regeneration Research Area

Exploratory human skill-transfer research area. Unvalidated until an independent operator reproduces a physical capability purely from the protocol.

Falsification / Boundary: Failure of independent human subjects to outperform baseline training controls on blind physical fabrication tasks.

[Historical Archive]PHA, GKS, ASM, IME, NKTP, DACA
SUPERSEDED

Historical & Superseded Formulations

Excluded from the current top-level canonical architecture (PHA/GKS/ASM/IME/NKTP lack distinct validated boundaries; DACA overlaps DCIA). Preserved for historical provenance only.

Falsification / Boundary: N/A — retained strictly in the historical archive; not active claims.

Canadian Industrial Crucible · Case Study Zero

PROPOSED· Industrial Application

Build what can be regenerated in Canadian-controlled shops.

Canada is Case Study Zero for the Sourdough Paradigm — applying distributed capability preservation and verification to Canadian-controlled physical manufacturing and shop-floor resilience. Strategic doctrine is ESTABLISHED; field tooling and industrial deployment architectures remain PROPOSED and evaluated by measured hardware output rather than policy slogans.

About & Direct Epistemic Sparring

ESTABLISHED· Open Review Protocol

Challenge it with evidence.

Authored by Michael Crumly (Regina, Saskatchewan, Canada). Agreement is never required — empirical evidence, formal counterexamples, and shop-floor failure traces are. If a component fails a test, documenting that failure on the public ledger is real progress.

michael@digamoni.ca

— Mike · Logic has to be an option.

Interactive Scenario Evaluation

A capability has been lost. Can you bring it back?

Don't just read about regeneration — run it. Four actions, four problems, one distributed network. Stockpiles run out. Starter cultures don't.