Running a Tiny Chess Neural Network by Hand

Explore a fascinating conference talk from !!Con 2024 where Amédée d'Aboville investigates the concept of running chess neural networks manually. Discover the intriguing story of someone who attempted to challenge world champion Magnus Carlsen after just one month of chess training, using the unconventional approach of calculating neural networks by hand during the game. Learn about the practical limitations and possibilities of minimal neural networks in chess, particularly when restricted to pen, paper, or basic calculators. Gain insights into the stark contrast between these simplified networks and modern mega neural networks, while understanding their potential practical applications. The 13-minute presentation, delivered by a full-stack software engineer from Montréal with interests in culinary mushrooms, chess, and techno music, was recorded at the !!Con 2024 conference in Santa Cruz, CA.