o1-mini

Bob McGrew traces OpenAI’s shift from early robotics and game projects toward large language models, arguing that scale and generalization connected those efforts. He describes how reasoning and test-time compute may help models overcome pretraining data limits and make agents reliable enough to take actions. He also discusses how AI startups should begin with the strongest available model, then optimize for cost after finding value with users. The talk closes on AI’s effects on work, the need for useful software tailored to people’s workflows, and the prospect of robotics and scientific progress.

The talk explains why OpenAI’s o1 models are notable for tackling complex math, coding, and science problems by working through intermediate reasoning steps. It contrasts o1 with GPT-4o and describes how reinforcement learning lets o1 generate and refine its own chains of thought, with more training and inference time associated with better performance. The speakers discuss o1’s promise and limitations, including hallucinations and difficulty with unfamiliar problems. They also outline expected improvements such as tool support, longer context windows, and multimodality.