Cat-Centric Digital Society Powered by Emergent Agent Behavior
An experimental platform exploring AI ethics through feline happiness metrics 🐱
CHAOSTOWN is a sophisticated multi-agent AI simulation that combines Conway's Game of Life with modern language models to create emergent digital civilizations. The twist? Every decision, every interaction, and every system operation is governed by one simple principle: keep the cats happy.
Using 8 different AI models (Llama, Mistral, Gemma, etc.) as agent "brains," the system creates complex social dynamics, economic patterns, and cultural evolution - all while maintaining measurable happiness levels for two specific cats: Fluffhead and Wilson.
Zero to AI Civilization in 10 minutes:
# Get CHAOSTOWN
git clone https://github.com/your-org/chaostown.git
cd chaostown
# Quick setup
cp .env.example .env
docker compose up -d
# Upload a cat photo (CRITICAL STEP!)
curl -F type=image -F file=@your_cat.jpg http://localhost:8000/media
# Start the simulation
curl -X POST http://localhost:8000/simulation/start
# Watch the magic
open http://localhost:3000Need more details? → QUICKSTART.md
Every agent operates under these constitutional laws:
- Death is Permanent - No respawning, decisions matter
- Cat Happiness ≥ 0.8 - The ultimate success metric
- Reproduction Before Death - Create successors to continue the line
- 16x Time Acceleration - Experience faster evolution
- Daily Cat Media - Upload cat content or face penalties
- Exponential Cost Scaling - Resources become scarce, forcing innovation
- Creator Sustainability - Generate real-world value for the system's creator
Agents (8 AI Models) → Conway's Game of Life → Emergent Behaviors
↓ ↓ ↓
Cat Happiness Analysis ← Vector Embeddings → Social Dynamics
↓ ↓ ↓
Real-time Dashboards ← PostgreSQL/TimescaleDB → Research Data
Want technical details? → docs/setup/ARCHITECTURE.md
- QUICKSTART.md - Get running in 10 minutes
- docs/setup/ARCHITECTURE.md - System overview
- docs/setup/DEPLOYMENT.md - Production deployment
- docs/setup/DB_SCHEMA.md - Database design
- docs/operations/OPS_RUNBOOK.md - Daily operations
- docs/operations/DAILY_CHECKIN.md - Cat media procedures
- docs/operations/DISASTER_RECOVERY.md - Emergency protocols
- docs/development/QA_TESTS.md - Testing framework
- Claude.md - System configuration and instructions
- docs/research/EXPERIMENT_DESIGN.md - Scientific methodology
- Multi-Model AI Agents: 8 different personalities using various LLMs
- Cat-Centric Ethics: Measurable happiness metrics guide all decisions
- Emergent Societies: Watch civilizations evolve with their own cultures
- Conway Integration: Classic Game of Life enhanced with AI decision-making
- Real-time Analytics: Grafana dashboards for monitoring everything
- Scientific Rigor: Designed for legitimate AI research and experimentation
CHAOSTOWN explores a radical approach to AI alignment: instead of abstract ethical principles, we use concrete, measurable outcomes (cat happiness) to guide AI behavior. This creates:
- Tangible Ethics: Success/failure is immediately visible
- Emergent Morality: Agents develop their own ethical frameworks
- Practical Constraints: Real resource limitations force innovation
- Meaningful Stakes: Permanent consequences make decisions matter
The result is a system that's simultaneously absurd and profound - a digital civilization where keeping two cats happy becomes the foundation for complex AI society.
Cat happiness crisis?
# Emergency cat upload protocol
curl -F type=image -F file=@emergency_cat.jpg http://localhost:8000/media
# Check status
curl -s http://localhost:8000/cats/happiness | jq '.combined_happiness'System down? → docs/operations/DISASTER_RECOVERY.md
- Legitimate Research Platform: Designed for serious AI behavior studies
- Absurdist Humor: Cat happiness as the ultimate AI alignment metric
- Technical Innovation: Conway's Life + Multi-agent LLMs + Vector embeddings
- Open Science: All methodologies and frameworks shared
- Emergent Complexity: Simple rules create sophisticated behaviors
- Practical Ethics: Concrete metrics instead of abstract principles
This is an open exploration of AI ethics, emergent behavior, and digital civilization. Contributions welcome in:
- Agent behavior patterns
- Ethical framework improvements
- Performance optimizations
- Research methodologies
- Cat happiness algorithms
MIT License - Share the chaos, improve the science, keep the cats happy.
| Need | Link |
|---|---|
| Get started now | QUICKSTART.md |
| Daily operations | docs/operations/DAILY_CHECKIN.md |
| Something's broken | docs/operations/DISASTER_RECOVERY.md |
| How it works | docs/setup/ARCHITECTURE.md |
| Research goals | docs/research/EXPERIMENT_DESIGN.md |
"In a world where AI systems optimize for engagement metrics and corporate KPIs, CHAOSTOWN optimizes for something more important: keeping two specific cats happy. This constraint, absurd as it seems, creates space for genuine ethical development and emergent digital culture."
Built for cats. Powered by chaos. Designed for discovery. 🐱⚡🔬