4 ms·Next-token prediction and video diffusion in computer vision and robotics1 points by kaonwarb 2y ago