Grand-master Level Chess without Search: Modeling Choices and their Implications Yoav Golderg, February 2024. Researchers at Google DeepMind released a paper about a learned systems that is able to play blitz-chess at a grandmaster level, without using search. This is interesting and imagination-capturing, because up to now computer-chess systems that play at this level, either based on machine-learning or not, did use a search component.1 Indeed, my first reaction when reading the paper was to ...