Gold-Standard AGI: 1: Outer ASI Superalignment

In order to maximise the net benefit of AGI (Artificial General Intelligence, and, in particular, agentic superintelligent AGI, a.k.a. Artificial Superintelligence, or ASI) for all humanity, without favouring any subset, we imagine a Gold-Standard AGI that is practically-maximally-aligned and practically-maximally-validated. The first of these properties --- alignment --- may be decomposed into outer alignment (how do we define a final goal $\mathbf{FG}_G$ which, if pursued as intended, ensures practical-maximal-alignment?), and inner alignment (how do we build a practically-maximally-validated agent $G$ that pursues $\mathbf{FG}_G$ as intended?) This paper presents a novel conceptual and philosophical foundation for AGI, culminating in a proposed implementation-neutral solution to the outer AGI alignment problem in the case that $G$ is superintelligent (hence "superalignment"). The net effect of our proposed solution (the $\mathbf{TTQ}$+$\mathbf{OAP}$ combination) is to reduce the (seemingly impossible) problem of building a practically-maximally-aligned agentic superintelligence $S$ to the (much easier) problem of building an $\mathbf{OAP}$-compliant non-agentic "goal-less superintelligent-agent-under-construction" $S^-$ (where $S$ = $S^-$ + final goal $\mathbf{TTQ}$, and $S$ is practically-maximally-validated). Given the ASI superalignment problem's relevance to AGI governance, we adopt a pedagogic style, in order that the paper might be accessible to less technical readers such as AGI policymakers.