Google DeepMind CEO Demis Hassabis: "We Will Have AGI By 2030"

Demis Hassabis, CEO of Google DeepMind and Nobel Prize winner in Chemistry, pushes back on claims that AGI has already arrived. In a rare interview at Davos, the man behind AlphaGo, AlphaFold, and AlphaZero explains why today's AI systems are nowhere near artificial general intelligence—and why turning AGI into a marketing term undermines the science. While Sam Altman declares OpenAI knows how to build AGI and is already aiming for superintelligence, Hassabis identifies the fundamental gaps that remain unsolved. Continual learning. Persistent memory. Long-term reasoning. The "goldfish brain" problem where models forget everything the moment you close a session. These aren't minor upgrades. They're breakthroughs that nobody has cracked. His definition of AGI sets a far higher bar than most. Not solving math problems—inventing entirely new fields of mathematics. Not generating art—creating genres that never existed. Einstein-level theory creation. Picasso-level artistic invention. Physical intelligence that matches elite athletes. By that standard, current systems don't register. But DeepMind isn't just waiting. They're building world models—AI systems that understand physics, causality, and can simulate reality itself. Hassabis believes these may be the missing ingredient, enabling the kind of long-horizon planning that humans do effortlessly but AI still cannot. His timeline: 5 to 10 years. His vision: AI that discovers room-temperature superconductors, new energy sources, and scientific breakthroughs hiding within the laws of physics that human minds simply haven't found. Source:    • Google DeepMind CEO Demis Hassabis: AI's N...   Google DeepMind CEO Demis Hassabis: "We Will Have AGI By 2030" #agi #ai #artificialintelligence