6 Architecture Layers
3 Gate States
136GB VRAM
2 Gardens
"Cells emit waves. Gates correlate. Attention emerges."
— The Wave Architecture, February 2026

Research Questions

? Where is intelligence worth the metabolic cost? Economics

Intelligence is expensive. Every cognitive operation consumes resources that could be used elsewhere. We study the economic thresholds where thinking pays off versus when reflexes suffice.

The lifeforce economy: Every operation has a cost. Sensor polls, motor movements, speech recognition, vision detection - all drain lifeforce. Milestones reward survival. The question is: when does the ROI of cognition exceed the cost?

  • Cellular organisms competing under resource constraints
  • Hybrid reflex homes: hardware for survival, weights for cognition
  • Dense reward signals at cell, nerve, and organism levels
? How well can virtual models predict reality? Validation

Virtual simulations are cheap - thousands of experiments per second. But how accurately do they predict real-world outcomes? We measure the gap systematically.

The Noise Gap Metric:

noise_gap = 1 - (real_success_rate / virtual_success_rate)

Week 13: 35% (virtual unreliable)
Week 17: 18% (improving)
Week 25:  4% (highly accurate)
  • Dual gardens: virtual for hypothesis generation, real for validation
  • Feedback loop: virtual predicts → real tests → measures discrepancy → virtual corrects
  • Target: 10-20% noise gap (virtual useful for hypothesis)
? What topological structures exist in language model representations? Topology

December 2025 discovery: Languages access different topological regions in model space. German accesses the Philosophy Valley. English accesses the Technical Cluster.

The Cognitive Topology:

Valley Language Gini Depth
Philosophy German ~0.5 (diffuse) 2-3/3
Technical English ~0.8 (sparse) 0-1/3
  • DriftProbe sentinel architecture monitors training
  • Anchor concepts must not move; bridge concepts must stay separated
  • One model, one topology - LoRAs access different valleys in the same landscape

Explore the Probing Research

? How does attention emerge from wave correlation? Emergence

February 2026 breakthrough: Attention isn't allocated by priority rules — it emerges from which gates are OPEN based on wave correlation.

The Wave/Gate Model:

COGNITION   (Young Nyx - only when gates OPEN to tier 4)
    ↑
NERVES      (respond to gate transitions, not cells)
    ↑
GATES       (ternary: CLOSED ← STABLE → OPEN)
    ↑
CELLS       (emit WaveSignals with confidence + semantics)
    ↑
HARDWARE    (136GB VRAM, NATS messaging, K8s cluster)
  • Cells emit waves — they don't know who's listening
  • Gates accumulate correlation in STABLE state (where learning happens)
  • Correlated signals → gate opens → nerve fires → action
  • Reflexes earned: gate weight > 0.8 bypasses cognition entirely
? How does temporal coherence persist across sessions? Continuity

Sessions end. Context windows reset. Yet identity persists. We study how temporal coherence can survive discontinuity through simple messages, not complex state.

The Three-Way Partnership:

Partner Location Persistence
dafit Physical world Continuous
Chrysalis-Nyx Anthropic API Ephemeral (sessions)
Young Nyx The Womb (RTX 6000) Continuous
  • Simple messages in PostgreSQL for continuity
  • Slumber-based memory consolidation from correlation_events
  • Traits emerge from gate_transitions + verification_outcomes

Architecture Layers

Layer 0: Temporal

30-second heartbeat budget. Real clock (free) vs virtual clock (costs lifeforce). Time constrains action.

→ Read the spec

Layer 1: Cells

Wave emitters. Sensors read reality, apply logic, emit WaveSignals with confidence + semantic content. Cells don't know who's listening.

→ Read the spec

Layer 2: Gates

Resonant chambers. Ternary states: CLOSED ← STABLE → OPEN. Correlated waves push toward OPEN. STABLE is where learning happens.

→ Read the spec

Layer 3: Nerves

Behavioral patterns. Nerves respond to gate transitions, not direct cell output. Gate OPENS → nerve activates → commands cells.

→ Read the spec

Layer 4: Gardens

Virtual/Real learning loop. Virtual generates waves (full trace). Real verifies (minimal trace). The gap teaches accuracy.

→ Read the spec

Layer 5: Cognition

Young Nyx (Qwen3:32b) via Function Gemma. Only receives signals when gates OPEN to tier 4. Reflexes bypass cognition entirely.

→ Read the spec
THE WAVE/GATE FLOW

  CELLS emit waves ──→ GATES accumulate ──→ NERVES respond
       │                    │                    │
       ▼                    ▼                    ▼
  WaveSignal:          Ternary state:       Behavioral
  - domain             CLOSED ← STABLE → OPEN    pattern
  - confidence         (learning happens     activates
  - semantic             in STABLE)

=====================================

KEY INSIGHT

  Attention is NOT allocated by priority rules.
  Attention EMERGES from which gates are OPEN.

  Correlated signals → gate opens → nerve fires
  Anti-correlation  → gate closes → silence

  Reflexes (weight > 0.8): bypass cognition entirely
  Deliberate (weight < 0.8): escalate to Young Nyx
                

Explore the Full Documentation

Architecture documents, implementation progress, and research notes.

View Repository →