A decade ago, S. Aglioti, J. F. X. DeSouza, and M. A. Goodale (1995) published an experiment that has had a big influence on the way that visual information is thought to control human behavior. Their findings have often been simplified as suggesting that action is immune to perceptual illusions. The current authors critically analyze the 4 steps involved in this simplification and argue that research during the last 10 years has shown that the validity of 3 of the 4 steps is doubtful. They conclude that this experiment cannot be regarded as firm support for the 2-visual-systems hypothesis (i.e., that the ventral stream is for perception and the dorsal stream is for visually guided actions).