The field of dexterous manipulation is moving towards more efficient and effective methods for controlling prosthetic and robotic hands. Recent research has focused on developing imitation learning-based approaches, such as diffusion policies, to improve grasping and manipulation tasks. These methods have shown promising results in reducing the cognitive load on users and enabling prosthetic devices to operate in more unconstrained scenarios. Additionally, there is a growing interest in exploring non-anthropomorphic hand designs and morphology-aware policy learning to enhance policy sample efficiency and generalize over dynamic, kinematic, and limb configuration variations. Noteworthy papers include: HannesImitation, which presents an imitation learning-based method for controlling a prosthetic hand, and DiWA, which introduces a novel framework for fine-tuning diffusion policies using a world model. UniFucGrasp is also notable for establishing a universal functional grasp annotation strategy and dataset for multiple dexterous hand types. Furthermore, the Real-Time Iteration Scheme for Diffusion Policy has been proposed to accelerate inference in diffusion policies, and Efficient Morphology-Aware Policy Transfer to New Embodiments investigates combining morphology-aware pretraining with parameter efficient finetuning techniques.