Layers12
LanguagePython
Font UsedSyne
🧠

OMEGA.

A 12-layer cognitive architecture that models, benchmarks, and rewrites itself.

Each layer is an evolutionary pressure—stacked like sediment, each load-bearing for the next.

Nothing static. Observing itself, proposing hypotheses, and evolving inside sandboxed boundaries.

Architecture by: Tejaswanth Surisetty · github.com/Tejaswanth2406 Theoretical Foundations: Holland, Friston, Baars, Clark, Langton
↓ ⌘
01 · elevation profile

Twelve layers,
one ascent.

01
Cognitive Genome
Evolvable component DNA — the base encoding every other layer mutates from.
/genome
02
Artificial Cell System
Modular cognitive cells that compose into larger reasoning units.
/cells
03
Neural Knowledge Fabric
A dynamic knowledge graph that reshapes itself as understanding grows.
/graph
04
Multi-Scale Memory
A 7-tier memory hierarchy spanning working to long-horizon recall.
/memory
05
Meta-Cognition
A machine self-model — the organism's account of its own state.
/metacognition
06
Continuous Scientific Method
Hypothesis-driven outputs — every claim is tested before it's trusted.
/scientific_method
07
Evolution Engine
Multi-level evolution across the genome, cells, and higher structures.
/evolution
08
Cognitive Immune System
Mutation safety gating — nothing evolves without passing quarantine.
/immune
09
Fast Ontological Recognition
Pre-retrieval classification that routes a query before it's expensive.
/ontology
10
Energy-Based Cognition
Budget-constrained reasoning — thought has a metabolic cost here.
/cognition
11
Cognitive Ecosystem
Competing agent populations, selected the way traits are selected.
/ecosystem
12
Recursive Self-Improvement
The full self-redesign loop — the summit layer, closing back on Layer 1.
/self_design
02 · definition

What counts
as intelligence, here.

the organism's working definition
Intelligence= Knowledge× Adaptability× Efficiency× Self-Improvement
03 · run it

Three modes,
one organism.

install
# dependencies
pip install -r requirements.txt
talk to it
python -m organism.main --mode interactive
let it evolve
python -m organism.main --mode evolve --generations 100
benchmark it
python -m organism.main --mode benchmark
04 · field notes

Standing on a few
tall shoulders.

John Holland
Evolutionary Computation
Karl Friston
Active Inference & the Free Energy Principle
Bernard Baars
Global Workspace Theory
Andy Clark
Predictive Processing
Christopher Langton
Artificial Life
05 · repository

Where each layer
actually lives.

artificial-organism/
├── genome/          # Layer 1: Cognitive DNA
├── cells/           # Layer 2: Artificial Cells
├── graph/           # Layer 3: Neural Knowledge Fabric
├── memory/          # Layer 4: Multi-Scale Memory
├── metacognition/   # Layer 5: Self-Model
├── scientific_method/ # Layer 6: Hypothesis Engine
├── evolution/       # Layer 7: Evolution Engine
├── immune/          # Layer 8: Immune System
├── ontology/        # Layer 9: Ontological Recognizer
├── cognition/       # Layer 10: Energy-Based Cognition
├── ecosystem/       # Layer 11: Agent Ecosystem
├── self_design/     # Layer 12: Recursive Self-Improvement
├── rag/               # Enhanced RAG system
├── benchmarking/       # Fitness benchmarking
├── simulation/         # Sandboxed mutation testing
├── self_repair/        # Self-repair mechanisms
├── observability/       # Logging, tracing, dashboards
└── deployment/          # Production deployment

Ascend it.

Clone the repo, run it in interactive mode, and watch a system reason about its own reasoning.

Clone the repository → Read the source