The field of interactive video generation and animation is increasingly focused on creating immersive and realistic experiences. Recent developments have centered around improving the controllability and coherence of generated videos, particularly in areas such as game video generation, human-scene interaction, and 3D animation. Notable advancements include the use of hybrid history-conditioned training strategies, joint video-pose diffusion models, and causal-aware reinforcement learning. These innovations have enabled significant improvements in visual fidelity, realism, and action controllability. Some particularly noteworthy papers in this regard include Hunyuan-GameCraft, which introduces a novel framework for high-dynamic interactive video generation in game environments, and GenHSI, which proposes a training-free method for controllable generation of long human-scene interaction videos. AnimaX is also noteworthy for its feed-forward 3D animation framework that bridges the motion priors of video diffusion models with the controllable structure of skeleton-based animation.