AI researchers Yann LeCun and Fei-Fei Li argued that world models represent a more viable technical pathway to Artificial General Intelligence than large language models