This Big Technology Podcast episode interviews OpenAI's Chief Research Officer, Mark Chen, about the release of GPT-4.5. The discussion centers on the model's development, focusing on the "predictable scaling paradigm" and the complementary roles of unsupervised learning and reasoning in its creation. Chen emphasizes that GPT-4.5 shows continued scaling returns, exceeding GPT-4 in various benchmarks, including user preference for everyday and knowledge work tasks (60-70% higher). He also highlights the model's improved emotional intelligence and creative writing capabilities. The interview concludes by discussing the future of AI model development, emphasizing OpenAI's dual-track approach of pushing the frontier of intelligence with large models while also developing more cost-effective, smaller models for various use cases.