Physicist: Physics Proves AI Is Inherently Evil!

The partner of the This is World channel is IQM: https://resonance.meetiqm.com/sign-up ⸻ IQM Quantum Computers is pioneering the future of superconducting quantum computers. With over $600 million raised, including the largest quantum raise outside of the US, IQM is the market leader in on-premises and cloud-accessible systems. IQM quantum computers empower the most innovative HPC centres, universities, enterprises, and researchers to push the boundaries of science and technology. To foster and grow the quantum ecosystem, we want to invite you to join IQM's quantum cloud platform, IQM Resonance. In it, you will find exclusive learning resources and courses on quantum topics and get access to our best quantum computers. Creating an account is free and no credit card is required. Explore Anthony Aguirre’s forecasting platform, Metaculus: https://www.metaculus.com/ What happens when AI stops being a tool… and becomes an autonomous agent? In this episode, physicist and AI-risk researcher Anthony Aguirre explains why the race toward AGI (Artificial General Intelligence) may be fundamentally misaligned with human interests. We talk about the real difference between “useful AI” and “autonomy,” why “alignment” might not exist in the way people imagine, and how the economics of replacement pushes companies toward systems that can do everything humans can do. Aguirre also dives into a deeper framework: why “most things an advanced AI can do are bad,” how this connects to control theory and thermodynamics (entropy), and why the control problem gets harder as systems become bigger, faster, and more capable than their human controllers. 00:00 — Why AGI may be against human interests 03:50 — AI tool vs AGI: autonomy changes everything 07:20 — “Why build something that can do everything a human can do?” 12:40 — Autonomy is a bug, not a feature 14:10 — Does AI alignment exist? 15:40 — Digital copy after death: “It would look like you… but it wouldn’t be you” 17:00 — Hinton’s idea: humans as the “second” intelligence 19:50 — The next 5–10 years: loss of control? 21:40 — Progress vs hype 25:40 — Consciousness: no definition, but huge consequences 29:50 — Free will, determinism, and physics 41:30 — Aguirre’s framework lecture: entropy + control problem Don't forget to subscribe to our channel and turn on notifications so you won't miss any of our future episodes ►    / @thisistheworldofficial