Tomaso Poggio: Brains, Minds, and Machines

Lex introduces Tommaso Poggio, an MIT professor and director of the Center for Brains, Minds, and Machines, cited over 100,000 times, whose work profoundly shaped our understanding of intelligence in biological and artificial neural networks. He advised highly impactful figures including DeepMind's Demis Hassabis, Mobileye's Amnon Shashua, and the Allen Institute's Christof Koch, and this conversation is part of the MIT AGI course and the AI podcast.

Recalling his boyhood fascination with relativity and Einstein as a hero, Poggio marvels that Einstein reached deep physical truths—how time depends on space and speed—purely through a thought experiment, the power of the mind. He believes such breakthroughs are in principle available to all of us, drawing a lesson from Einstein being the worst of five Zurich PhD students, the only one without an academic post, who went to the patent office: an anti-conformist who didn't think traditionally. Don't buy when everyone's buying, in the stock market and in science.

Staying with physics, Lex asks the out-of-the-box question whether time travel is possible. Poggio says in science you never say no, but it's very likely impossible to travel back in time—we might travel forward by freezing ourselves or near-light-speed spacecraft. Asked if he still holds the childhood dream of engineering intelligence able to make such leaps, he turns to the nearer dream.

Poggio believes machines will, in the short and midterm, help us think better, and sees no reason they couldn't become more intelligent than us—though intelligence, consciousness, and love must be disentangled. He calls intelligence the greatest problem in science, greater than the origin of life or universe. As a teenager he reasoned that solving intelligence was the key that unlocks all other problems—an amplifier like an Einstein ten times faster—but it soon became a genuine focus because the deepest question is who we are, using our own brain as the very tool of science.

Asked how far we can get engineering intelligence without understanding the biological brain, Poggio notes we solved flight without much bird biology—though the Wright brothers did observe birds. For intelligence it's a bet: we'll surely reach a machine indistinguishable from his secretary, but whether we get there with or without brain knowledge is an educated wager people decide by background.