"We’re creating beings, not tools" | Geoffrey Hinton & Joel Hellermark

There should be more to AI than intelligence. Nobel Prize winner Geoffrey Hinton believes that better AI requires more than faster tasks and greater profits. As we move from being the only intelligent species to coexisting with artificial minds, we face a choice: build systems that are merely powerful, or ones that genuinely care about human flourishing. From fast‑learning neural architectures to training data ethics, his conversation with Joel Hellermark, Sana’s CEO, offers fascinating insight to anyone wondering whether we have any control over AI, and if so, how to wield it. Timestamps: 0:00 Ever-faster loops of intelligence: evolution, development, learning, and AI 6:08 The “third revolution” after Copernicus and Darwin: creating new beings 7:44 How self-play and internal inconsistencies can let LLMs improve with little new data 11:02 The danger of a profit-driven race for smarter AI and designing beings that care about us 14:18 Hinton’s architectural bet: fast-changing weights and brain-like synapses Recorded live at Sana AI Summit 2026, New York, May 21st, 2026. Subscribe for more insights on AI and the future of technology.