In the video “NEXT AI,” Matthew Berman discusses the evolution of AI beyond traditional models, focusing on a new approach called Reinforcement Pre-Training (RPT). This method aims to enhance AI capabilities without relying on curated datasets, potentially revolutionizing how AI learns and operates. Berman explains the underlying mathematics and concepts, emphasizing the importance of understanding these advancements for the future of AI development.

Discover AI
Not Applicable
June 15, 2025
PT40M58S