Episode Details

The "Thinking Revolution" is here, shifting AI from fast, intuitive guesses to the deliberate "System 2" reasoning of OpenAI’s o1 and the open-source disruptor DeepSeek-R1. We break down the architectural pivot of inference-time scaling, exploring how reinforcement learning and the "Aha moment" of machine logic are shattering the Chinchilla scaling laws. From the efficiency of GRPO to the distillation of PhD-level reasoning into lightweight models, dive into the new meritocracy of code where intelligence is becoming a global commodity.