1. What AGI Actually Requires (A Structural Definition) In open discussions, “AGI” is often described as: a very large model, a universal problem solver, a human‑level agent, a system based on subjective experience. These definitions contradict each other and do not provide an engineering criterion. A structural definition of AGI: AGI = a system with a stable vertical cognitive architecture capable of generating, evaluating, and refining its own direction (S1), constraints (S2), knowledge (S3), and honest integration (S4), and capable of completing a full reasoning cycle (S1–S11) without collapse. This definition does not depend on: model size, training data, biological analogies, philosophical assumptions. It depends only on structure . 2. Why Modern AI Systems Cannot Be AGI LLMs and agent frameworks lack key elements of vertical cognition: Missing S1 — Direction Models do not generate their own goals. Missing S2 — Values and Constraints No internal priorities or risk boundaries.…