Samuel Pedrielli comments on Ego‑Centric Architecture for AGI Safety: Technical Core, Falsifiable Predictions, and a Minimal Experiment