The AI Ego: What a 1,600-year-old theory of consciousness reveals about the design risks hiding in agentic systems | Synapse