• Google DeepMind, the artificial intelligence research company, has developed a table tennis robot that can play the game at an amateur level. The robot, called "Tischtennis," was trained using reinforcement learning, a technique where the AI agent learns by trial and error, receiving rewards or penalties based on its actions.

• The robot's performance is described as "solidly amateur," meaning it can return shots and rally with a human opponent, but it is not at the level of professional or even highly skilled players. The article notes that the robot's movements are still a bit jerky and its timing is not as precise as a human player.

• The development of the Tischtennis robot is part of DeepMind's ongoing research into building AI systems that can master complex physical skills and interact with the real world. While the robot's abilities are still limited, the article suggests that this work could lead to advancements in areas such as robotics, sports analysis, and human-machine interaction.