Terry Winograd developed SHRDLU, a computer program capable of participating in conversations about a made-up world, in the late 1960s. This program was able to respond to commands in English and “understand” questions about its world. Despite its impressive performance, Winograd soon abandoned this line of research in favor of pursuing work devoted to the interaction of computers and people. In 2006, Winograd saw statistical language understanding, neural networks and machine learning as new developments that could potentially lead to human-like AI. However, LLMs still have to overcome their instability, lack of interpretability, and vulnerability to automated adversarial attacks before they can become “superintelligence.”
