Dr. Li recounts the development of ImageNet, which catalyzed the deep learning revolution by demonstrating the power of data-driven methods and convolutional neural networks. She discusses how this led to advancements like image captioning and generative models. Turning to present-day challenges, she emphasizes the importance of spatial intelligence for achieving AGI, explaining that modeling the 3D world is more complex than language understanding. In reflecting on her personal journey, she offers advice to young researchers, encouraging intellectual courage and interdisciplinary exploration while highlighting emerging opportunities in AI research and open-source collaboration.