Since the early years of the modern discipline, nothing in underwater archaeology has evolved as dramatically as the technology for site and landscape recording. Photogrammetry, Photo-modeling, SLAM, and various acoustic imaging systems have all been touted as the ‘next big thing’ in digital mapping. Yet as much as archaeologists are eager to trade the laborious work of manual recording for the promises of the latest gadgets, we have yet to find a site-mapping technology with enough clear advantages for it to be widely adopted. Issues of cost, accuracy, and post-processing time are usually paramount. The capability to translate points and images into archaeologically useful data and diagrams is also a concern. In this paper I discuss the experience of using the Pladypos cognitive diving robot built by the University of Zagreb to map the ancient port of Caesarea, and offer some ideas about the future of robotics in underwater archaeology.