Self-Assembling Coherence
A swarm-based framework for studying AGI alignment through emergent coherence. Agents self-organize around shared attractors while entropy forces test system stability. The simulation below lets you explore how alignment force, chaos, and density interact in real time.
Attraction toward coherent center
Random perturbation magnitude
Number of agents in the swarm
Self-Assembling Coherence
AGI-SAC models alignment as an emergent property of multi-agent interaction rather than a top-down constraint. Each agent in the swarm follows simple local rules: move toward nearby agents (cohesion), avoid collisions (separation), and orient with neighbors (alignment).
The Alignment Force (α) controls how strongly agents are pulled toward the swarm's coherent center — the emergent attractor. Higher values produce tight, stable clusters. Lower values allow agents to drift and explore.
The Chaos/Entropy parameter (Ω) introduces random perturbations that test system resilience. When entropy exceeds alignment, the system transitions from STABLE to CRITICAL — a phase change analogous to alignment failure.
The key insight: robust alignment doesn't require perfect control. Systems that tolerate moderate entropy while maintaining coherence are more resilient than brittle, over-constrained ones.