V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planning Paper • 2506.09985 • Published Jun 11 • 28 • 2