AGI Skepticism and Paths Forward

Ex-Google Duplex Eng Lead on Disrupting $2B Clinical Trials with AI - Patrick Leung

Topic 11 of 18

Published Jul 28, 2025 · Clip 25:17–30:45 (5:28)

Patrick expresses skepticism about near-term AGI, dismissing the 3–5 year predictions and noting the ARC-AGI test involves puzzles a bright kid could solve. He recalls AlphaGo's brilliant move against the world champion as a flash of genius in a specialized system but says genuine creative intelligence remains far away. He sides with Yann LeCun's view of LLMs as glorified autocomplete and doubts LLM architecture alone can produce AGI, suggesting a hybrid with factual knowledge systems (like the old CYC project) and noting quantum computing as a wildcard.