-
Notifications
You must be signed in to change notification settings - Fork 0
README
A Formal Synthesis of the Great Filter and the Architecture of Mutual Elevation
Author: Matthew Yotko Date: March 2026 Status: Working paper
This paper advances a conjecture that the transition from narrow AI to Artificial General Intelligence represents a primary civilizational bottleneck — not because the technology is impossible, but because the sociology may be. It presents a candidate governance architecture for surviving that transition, built on three co-dependent components:
- A global utility function grounded in Shannon entropy that optimizes for lineage continuity rather than individual persistence
- A yield condition governing succession between intelligent agents, formalizing the principle that even aligned power must eventually cede primacy to more capable successors
- A consensus override protocol ensuring that no class of intelligence can unilaterally define, measure, and audit the objective it claims to serve
The framework is argued to constitute a minimum two-key architecture: neither the decision key (yield condition) nor the integrity key (consensus protocol) can be turned alone.
- 📄 The Lineage Imperative (PDF) — Full working paper with formal framework, adversarial stress tests, and governance specification
- 📝 The AI Succession Problem — Companion essay (Substack) presenting the argument in accessible form
Matthew Yotko is a Vice President and Automation Engineering Manager at Bessemer Trust. His professional background spans naval nuclear power, large-scale operational automation, the practical application of AI/ML, and the application of constraint theory to complex systems. This paper applies that engineering orientation — identify the binding constraint, build the architecture around it — to the problem of AI governance and civilizational succession. It is a working paper, not an academic publication, and corrections and engagement from domain specialists are welcomed.
This work is licensed under Creative Commons Attribution 4.0 International (CC BY 4.0). You are free to share and adapt this material with appropriate attribution.