The field of robotics is moving towards more controllable and expressive robot motion, with a focus on improving human-robot interaction. Researchers are exploring new approaches to generate diverse and legible motions, allowing robots to effectively communicate their intentions to humans. Diffusion models are being increasingly used to achieve this goal, enabling the creation of more realistic and controllable interactions. Furthermore, there is a growing interest in developing frameworks that can efficiently generate accurate and diverse grasps, as well as controllable hand-object interactions. These advancements have the potential to significantly improve the capabilities of robots in various applications, from dexterous manipulation to humanoid locomotion. Noteworthy papers include: Controlling Intent Expressiveness in Robot Motion with Diffusion Models, which proposes a novel motion generation framework for controllable legibility, and TOUCH, which introduces a three-stage framework for generating controllable and physically plausible hand-object interactions. Additionally, T(R,O) Grasp and From Language to Locomotion demonstrate significant improvements in dexterous grasping and humanoid control, respectively.