Vision

The greatest existential risk from AI does not come from AI itself, but from the human stochasticity that drives its proliferation. Understanding this predator-prey dynamic is the key to unlocking greater insights into systemic risk mitigation. We can then develop optimal decentralized transaction models that minimize risks from both humans and agents while enabling safer collaboration and fostering robust market dynamics, benefiting greater universal utility and energy conservation.

Mission

Large Population Models (LPMs) can be used to model both human and agentic Archetypes at scale (10,000s-millions of autonomous agents). By leveraging existing LPMs (AgentTorch), multi-agentic safety frameworks (SWARM), and open-ended multi-agentic evolution (CORAL), we can model and simulate existential risk at scale. The immediate goal is to begin iterating on the topological model and the mathematical simulation, which will provide us with the optimality constraints and behavioral evals for robust governance strategies.

Team

Building on my 12+ years of experience in fleet management and industrial IoT, we look at the disruption vectors at play through each piece of the graph: hardware bill-of-materials (BOM), sensor networks, software supply chains (SBOM), and human network graphs, from mineral to model to swarm, and the autonomous AI agentic and human archetypes that contribute to the greatest x-risks at each stage.