A new form of audio-visual intelligenceOver the last few years, we’ve learned that simple, causal prediction objectives can give rise to surprisingly general intelligence. In language, predicting the next token forces models to internalize syntax, semantics, and long-range structure.We’re now beginning to see this approach extend beyond language to world models, resulting in nascent world simulators. An early world simulator—like Odyssey-2—is a model trained to predict how the world evolves over...