The idea that visual data processing by AI alone could suffice for a robotic hand to gently grasp an apple is a dangerous assumption. It’s like believing a cook can make a good dish just by looking at the ingredients, without knowing how to cut or mix; tactile sensation is the weak link here. In Portugal, our expert hands know that without very precise pressure sensors and haptic feedback, a Festo could turn the apple into applesauce. For such a delicate task, knowledge of the physical world far exceeds simple visual recognition, which is only an insufficient prerequisite.