Sam Altman - Reinforcement Learning in o1 (via searchlore.ai)