Modeling personality as adversarial equilibrium
A computational framework for representing a person's decision-making as competing internal drives that stake evidence and find equilibrium. Not a profile. Not a summary. A dynamic system that is the personality, in computational form.
Ideas are atomic. Position is relational.
The observation "Lives paycheck to paycheck at 42" has no inherent meaning. Its meaning emerges from what you're optimizing for:
- SURVIVAL sees: financial risk, precarity
- MEANING sees: sacrifice for purpose, choosing mission over money
- AUTONOMY sees: rejecting salary slavery, freedom at cost
- COMFORT sees: unsustainable intensity, warning sign
Same fact. Different positions. The World Model captures this by letting the same observation appear across multiple value trees with different polarities.
Seven universal human tendencies act as agents competing for influence:
| Tendency | Question It Asks |
|---|---|
| SURVIVAL | "Is this safe? Do I have enough?" |
| STATUS | "Am I respected? Do I matter to others?" |
| MEANING | "Does this matter? Will it outlast me?" |
| CONNECTION | "Am I known? Do I belong?" |
| AUTONOMY | "Am I free? Can I choose?" |
| COMFORT | "Is this pleasant? Can I sustain this?" |
| CURIOSITY | "Do I understand? What's there to learn?" |
Each agent:
- PROPOSES claims about what matters
- STAKES observations as evidence (PRO or CON)
- WINS or LOSES debates based on evidence strength
- GAINS or LOSES allocation based on outcomes
The equilibrium that emerges—which tendencies dominate, where they conflict, how they resolve—is the personality.
First training run on 165 observations:
| Metric | Value |
|---|---|
| Accuracy | 27.3% |
| Baseline | 7.1% (random chance = 1/7) |
| P-value | 0.001 |
The model predicts which tendency "owns" an observation 4x better than chance.
from world_model import create_world_model, Arena
# Load observations about a person
model = create_world_model("Person", "observations.json")
# Run adversarial debate
arena = Arena()
trees, result = arena.run_full_debate(
observations=model.observations,
agents=model.agents,
)
print(f"Winner: {result.winner}")
print(f"Allocations: {model.agents}")from world_model.dynamics import Trainer, TrainConfig
config = TrainConfig(
max_epochs=5,
convergence_threshold=0.01,
validation_split=0.2,
)
trainer = Trainer(config)
history, result = trainer.train(
observations=model.observations,
agents=model.agents,
)
print(f"Validation accuracy: {history.validation_results[-1].accuracy:.1%}")world_model/
├── models/ # Core data structures
│ ├── observation.py # Atomic facts (~280 bytes)
│ ├── agent.py # Tendencies with allocations
│ └── tree.py # Value hierarchies + weight propagation
├── dynamics/ # Adversarial competition
│ ├── arena.py # Debate orchestration
│ └── trainer.py # ML-style training loop
├── extraction/ # Extract observations from text
├── staking/ # Evidence staking mechanisms
└── storage/ # JSON and Firestore persistence
api/ # FastAPI service
docs/ # Full documentation
This architecture bets that:
- Binary distinction + recursion = sufficient to model meaning
- Seven tendencies = comprehensive but tractable agent set
- Competition = produces coherent personality from plurality
- Same structure = applies to anyone (swap observations + allocations)
- More useful = than unstructured text at scale
If the bet pays off, this is a general architecture for modeling minds.
The world model captures the mind—values and decision-making. A complete digital twin adds the body—voice, face, mannerisms.
┌─────────────────────────────────────────────────────────────┐
│ COMPLETE DIGITAL TWIN │
├────────────────────────────┬────────────────────────────────┤
│ WORLD MODEL │ EMBODIMENT MODEL │
│ (Mind) ✓ │ (Body) ○ Future │
├────────────────────────────┼────────────────────────────────┤
│ What they'd decide │ What they look/sound like │
│ Why they'd decide it │ How they move/gesture │
│ Internal tensions │ Mannerisms and style │
├────────────────────────────┴────────────────────────────────┤
│ EXPRESSION BRIDGE │
│ Tendency activation → Physical manifestation │
│ MEANING at 0.8 → animated gestures, faster speech │
│ SURVIVAL at 0.7 → tense voice, guarded posture │
└─────────────────────────────────────────────────────────────┘
See docs/future-digital-twin.md for the full vision.
This system is infrastructure for After Me—trustless estate planning with posthumous digital continuity.
The diary in After Me isn't just a journal. It's a training corpus. Enough video entries create the raw material for a digital double that:
- Reasons from your values (world model)
- Looks and sounds like you (embodiment model)
- Carries cryptographic attestation via embedded weight hashes—proving content derives from your authentic model, not an impersonation
This inverts the deepfake problem: instead of detecting fakes, you prove authenticity. Content without your embedded signature is suspect by default.
- Architecture — System design and components
- Core Concepts — Key ideas and terminology
- Training Guide — ML-style training with convergence detection
- API Reference — Module documentation
- Future: Digital Twin — Vision for complete representation
A person isn't monolithic. They're a coalition of drives finding equilibrium. The structure of that equilibrium IS the personality.
Traditional profiles describe a person. This system is the person's decision-making structure—or at least, a computationally faithful approximation that can speak for them when they no longer can.
Not immortality. But continuity.
MIT