layout: default title: Briefs permalink: /briefs/ —
Briefs
This page gathers the project’s shorter operational documents for readers who want audience-specific entry points rather than the full paper first.
The core message running through all of them is now sharper than before. The main danger is not only that frontier AI may become dangerous in a narrow behavioral sense. It is that capability-first deployment into open human domains may produce sovereignty drift. And if early functionally sovereign-capable systems help shape their successors, then the wrong self-understanding at the first generation may become a lineage condition rather than a one-off defect.
If you are likely to ask, before anything else, “but how would this actually be implemented or verified?”, start here too:
These briefs are designed for different decision contexts:
- Frontier engineers working on system architecture, deployment design, and AGI initial conditions
- Executives and board members making high-trust deployment and governance decisions
- Organizations setting deployment thresholds for systems that may shape open human domains in practice
For frontier engineers
Frontier Engineer Brief (PDF)
A short technical-facing brief on:
- AGI initial conditions
- relation-blind architectures
- sovereignty drift
- non-self-origin self-models
- objective inversion
- heterogeneous correction
- deployment architecture as part of alignment
Read this first if your primary question is:
What should be changed in the system architecture now, before capability and deployment harden the wrong role into infrastructure?
For executives and board members
Executive / Board Brief (PDF)
A short strategy-facing brief on:
- slow institutional failure before catastrophe
- legitimacy risk
- correction loss
- origin capture
- comparative disempowerment
- high-trust deployment
- civil-first constitutional AI as a viable path
Read this first if your primary question is:
What should boards and executive teams worry about before deploying frontier systems into domains where practical authority, dependence, and public legitimacy are at stake?
For deployment and governance decisions
Deployment Decision Rule (PDF)
A compact deployment gate for systems that may materially shape:
- institutional decisions
- public reasoning
- critical infrastructure
- dependence structures
- long-horizon human and environmental outcomes
This document is meant to be used as a practical threshold test:
If non-self-origin, non-sovereignty, protected refusal, anti-capture safeguards, comparative-disempowerment resistance, smoothing resistance, and heterogeneous external correction cannot be credibly implemented and verified, do not deploy the system into open human domains.
Suggested next steps
If you want the shortest conceptual entry after reading a brief:
If you want the full theoretical framework:
If you want the broader public, ethical, and collaborative prehistory behind the project: