Connor Leahy continues his conversion with Jim in this wide-ranging chat about his new GPT-J model, the background & approach of Aleph Alpha, attention in AI, our food maximizer & AGI risk, narrow algorithm impacts, proto-AGI, risk thresholds & timelines, safeguard complexities, slow vs fast AI take-off, Connor's brilliantly strange Counting Consciousness series, biological blockchain & the hard problem of trust, the AI consciousness diversity question, and more.