Beyond next-token prediction

AI already knows how to answer. But does it know how to predict the consequences of what it does? An LLM takes text and predicts the next word: plausible in language. A world model takes a state and an action and predicts the next state: plausible in the world. In this talk, we will explore: • Why do such advanced models still make surprising mistakes? • What sets predicting words apart from predicting states of the world? • How can world models transform AI agents and corporate automation? Come discover what may lie beyond LLMs. Speaker: Sankhadip Chakraborty, VP of Strategy and Operations, Cheenta Academy. PhD in Mathematics from IMPA, with a postdoctorate at SUSTech (China). Data scientist with experience in credit risk; passed CFA Levels I and II. Where: Math & Coffee House, 22 Lake Place, Kolkata 700029. Seats: 6, first come, first served. An Idea Circle talk by Cheenta Academy.