Visual Language Models for Robotics