Pulsar

Fei-Fei Li traces how ImageNet helped shift computer vision toward data-driven machine learning and how the ImageNet Challenge brought data, GPUs, and neural networks together in the AlexNet breakthrough. She describes the progression from recognizing objects to captioning scenes, then argues that spatial intelligence: understanding, generating, and reasoning about 3D worlds, is a fundamental frontier for AI. At World Labs, she and her team are working on world models, which she sees as useful across creative fields, robotics, and virtual worlds. She also reflects on entrepreneurship, fearless research, graduate study, and the value of a healthy mix of open and closed approaches to AI.