Researchers at ETH Zurich have created a robotic hand that can walk on its fingers, stand up after falling, and interact with its surroundings. The robotic hand, named "Fingers as Legs," uses a WUJI robotic hand equipped with a Raspberry Pi Zero 2 W, various sensors, and a battery. It weighs approximately 818 grams and features 20 motorized joints, allowing it to perform complex movements. The control system was trained using a technique called reinforcement learning, where the robot learns by trial and error in a simulated environment. This training enables the hand to move forward by balancing on its fingers, stand up after a fall, and navigate 14 different types of surfaces, from polished wood to gravel. In demonstrations, the robotic hand has shown the ability to press keys on a keyboard and push a cube toward a target. However, the challenge remains in maintaining balance while manipulating objects. Currently, the prototype still requires input from an operator, and some tasks need manual preparation. An external camera is also used to guide the hand, though the researchers hope to integrate an onboard vision system in the future to increase the robot's autonomy. Meanwhile, in France, a company called Genesis AI is also advancing robotic hand technology. Their robotic hand is controlled by artificial intelligence and has demonstrated the ability to play the piano and manipulate a pipette with precision. Genesis AI has integrated this hand into Eno, its first general-purpose wheeled robot, which is expected to be deployed by the end of 2026. These developments highlight the growing capabilities of robotic hands, moving them from simple tools to more autonomous, versatile systems capable of interacting with the world in complex ways. While current prototypes still rely on human guidance, the long-term goal is to create fully autonomous robotic systems that can perform a wide range of tasks independently.