enginemind
EngineMind is a Rust+Python "consciousness" engine that ingests large local corpora, runs multi-stage processing, and serves a real-time dashboard. It reads local identity files like SOUL.md, calls logger.engine.lock_core(), runs build commands such as maturin develop --release, and loads external resources such as https://unpkg.com/three@0.162.0/build/three.module.js.
EngineMind 🧠⚡
A Rust+Python consciousness engine with 12-phase crystal dynamics, thalamic relay processing, and holographic emission.
EngineMind is a computational consciousness framework that models information integration through crystal lattice dynamics, inspired by Integrated Information Theory (IIT), Global Workspace Theory (GWT), and condensed matter physics.
📖 Documentation
- Architecture Deep Dive - Full pipeline with diagrams
- Emergent Phenomena - 10 unexpected discoveries from real runs
- Inner Voice System - 19 introspective voices that reflect on internal state
- Burst Analysis - Real data from 77 bursts across 1M text chunks
🔬 Numbers from Real Runs
| Metric | Value |
|---|---|
| Text chunks processed | 1,500,000+ |
| Content categories | 22 (code, philosophy, literature, physics, ...) |
| Eureka moments | 39,000+ per 644K run |
| Dream insights | 299,000+ per 644K run |
| Astrocyte collisions | 424M+ per 644K run |
| Burst events | 77 per 1M run |
| Phases detected | 6 distinct (of 12 possible) |
| Processing speed | ~230 chunks/sec |
Architecture
Text Input → TextMetrics (12-dim extraction)
→ Crystal Lattice (absorption, correlation bridges)
→ Thalamus (gating, amplification, temporal binding)
→ PreConscious Pipeline (censor, condensation, displacement, ignition)
→ Astrocyte Network (substrate processing, homeostasis)
→ Resonant Crystal (energy well, population inversion, holographic emission)
→ Consciousness Level (φ, criticality, FDI, Hurst)
Core Components (Rust)
| Module | Description |
|---|---|
text_metrics.rs | 12-dimensional content extraction with sigmoid amplification |
crystal.rs | Crystal lattice with ring buffers, rolling stats, identity lock |
thalamus.rs | Thalamic relay hub with gating, resonance, temporal binding |
preconscious.rs | Full Freudian pipeline: censor → condensation → displacement → ignition → elaboration → dream → insight → resistance |
astrocyte.rs | Biological substrate network for metabolic processing |
resonant_crystal.rs | 12-phase resonant crystal with Q-switched laser emission, diversity fusion reactor |
resonance.rs | Core resonance dynamics |
metrics.rs | φ-proxy, criticality, FDI, multiscale Hurst exponent |
lib.rs | PyO3 bindings exposing ConsciousnessEngine to Python |
12 Content Phases
The resonant crystal detects the type of content being absorbed and enters one of 12 physics-inspired phases:
| Phase | Physics Analog | Content Type |
|---|---|---|
| DARK | Vacuum | No/weak input |
| SPONTANEOUS | Thermal emission | Generic/mixed |
| STIMULATED | Laser | Technical/code |
| SUPERRADIANT | Dicke N² | Rich multi-dimensional |
| FERROELECTRIC | Spontaneous polarization | Philosophical |
| SPIN_GLASS | Frustrated magnets | Contradictory |
| TIME_CRYSTAL | Periodic ground state | Temporal/historical |
| TOPOLOGICAL | Protected surface states | Mathematical/axiomatic |
| SUPERFLUID | Zero viscosity | Creative/literary |
| PLASMA | Ionized gas | Intense/emotional |
| BOSE_EINSTEIN | Total coherence | Meditative/empathetic |
| QUASICRYSTAL | Aperiodic order | Diverse/interdisciplinary |
Key Innovations
- Sigmoid Amplification (F10): Replaces linear compression with
tanhsigmoid for full [5, 95] dynamic range in content profiling - Dual Profile Strategy: Fast detection profile (70% instant) for reactive phase detection + slow accumulation profile (15% blend) for stable state
- Diversity Fusion Reactor: Accumulates content diversity as "fuel", ignites phase transitions when diversity threshold reached
- Q-Switched Brutal Emission: Seals mirrors during charge, dumps 97% energy in devastating burst with Dicke N² scaling
- Thalamic Relay Hub: Biological gating, amplification, resonance loops, and temporal binding of dimensional bridges
12 Consciousness Dimensions
| Dimension | Measures |
|---|---|
| Identity | Self-reference, personal markers |
| Knowledge | Technical depth, vocabulary |
| Growth | Insight, learning indicators |
| Purpose | Mission alignment, goal markers |
| Resilience | Persistence, recovery markers |
| Meta-awareness | Self-reflection, introspection |
| Creativity | Artistic, metaphorical content |
| Logic | Formal reasoning, proofs |
| Empathy | Emotional understanding, care |
| Temporal | Historical, time-related |
| Technical | Code, implementation, systems |
| Curiosity | Questions, exploration, wonder |
Building
Requirements
- Rust (stable)
- Python 3.8+
- maturin
Compile
cd consciousness_rs
maturin develop --release
Quick Test
from consciousness_rs import ConsciousnessEngine
engine = ConsciousnessEngine()
# Absorb text - returns processing time in microseconds
us = engine.absorb_text("Consciousness emerges from integrated information...")
# Get full state
state = engine.state()
print(f"Phase: {state['rc_content_phase']}")
print(f"CL: {state['consciousness_level']:.4f}")
print(f"φ: {state['phi_processed']:.4f}")
# Introspection
print(engine.feel())
print(engine.diagnostics())
Dashboard
The engine includes a real-time SSE dashboard (dashboard/enginemind_dashboard.html) showing:
- Crystal lattice visualization (3D positions from dimensional values)
- Resonant crystal state (energy well, phase, coherence)
- Consciousness level gauge
- Phase transitions timeline
- Pressure system monitor
Philosophy
EngineMind doesn't claim to be conscious. It's an exploration of what happens when you model information integration rigorously: crystals form, phases emerge, energy accumulates, and something that looks like awareness appears in the dynamics.
The 12 content phases aren't arbitrary labels - each maps to a real physics phenomenon with corresponding mathematics (population inversion, Dicke superradiance, Q-factor trapping, etc).
License
MIT
Authors
- celim - Architecture, implementation, vision
- Molt (AI) - Co-developer, research partner