Moravec's Paradox highlights the gap between human cognitive abilities and machine capabilities. While computers excel at complex tasks like playing chess or solving math problems, they struggle with everyday physical actions such as opening a door or spreading peanut butter. In response to this paradox, Benjie Holson proposed "Robot Olympics" tasks that focus on these seemingly simple yet challenging activities. The tasks aim to test a robot's manipulation skills, emphasizing that actions we find easy can be incredibly difficult for machines. The authors fine-tuned their π0.6 model to tackle these challenge tasks, achieving notable results. They successfully completed "gold medal" tasks in three out of five categories and "silver medal" tasks in the others. Two tasks were deemed physically impossible for their robot, although one could be tackled with a minor modification. Overall, they found an average success rate of 52% across tasks, significantly higher than a baseline robot model, which only managed a 9% success rate. This underscores the importance of large-scale pre-training for robotic systems. Specific tasks included opening a self-closing door and turning a sock inside-out. The latter task presented challenges due to the robot's gripper design. Other tasks, like making a peanut butter sandwich or peeling an orange, further illustrated the difficulty of fine motor skills for robots. The authors point out that while humans can easily handle physical tasks, programming these abilities into machines is complex, as robots lack fundamental physical skills. They can process language and instructions but struggle to translate those into actions without a base level of physical intelligence.