A California bill proposing safety measures for large AI systems passed a crucial vote, aiming to prevent potential risks like grid sabotage. Despite opposition, the legislation seeks to set safety rules for AI models over $100 million in data. Governor Newsom will decide its fate by end of September.