Building Models That See the World - Eric Xing with Nicholas Thompson

If you want AI to drive your car or simulate a cell or control a robot, it had better be able to understand the world it’s operating within. Eric Xing is one of the leading researchers working on this kind of AI, known as ‘world models.’ Unlike LLMs, which are trained largely on text data and predict the next token, world models aim to understand and simulate physical environments using multimodal data. Xing found his way to the field after a long trek through prestigious institutions, studying cancer before transitioning to machine learning. Now, as the president of Mohamed bin Zayed University of Artificial Intelligence in the United Arab Emirates, he’s helping to train a generation of AI builders. In conversation with Nicholas Thompson, CEO of The Atlantic, Eric discusses how his unusual career trajectory prepared him to work in the AI industry, what people misunderstand about the US-China arms race, and how the war in Iran has impacted his work in the region. Produced independently of The Atlantic’s editorial staff. (0:00) Introduction (02:03) Lessons on survival from a mountaineering expedition gone wrong (04:48) Eric’s winding path from physics to biology to computer science (11:56) Balancing academia and business: raising $100M and founding an AI company (19:41) The limits of large language models (26:43) Using world models to simulate biology (36:30) Revolutionizing drug discovery with digital organisms (43:10) Why Eric refused to sign an AI safety letter (49:37) Debunking existential risk narratives: the difference between a blueprint and a weapon (57:40) Why Eric is a “true believer” in open-source AI (1:05:37) How the Iran War has impacted AI in the UAE (1:11:18) What Eric would do if he had an unlimited budget to invest in AI Learn more about your ad choices. Visit podcastchoices.com/adchoices