OpenAI Podcast

Math is one of the clearest ways to see how far AI has come in a short span. OpenAI researchers Sébastien Bubeck and Ernest Ryu join host Andrew Mayne to explain what changed and what it could mean for the future of research. They reflect on how Ernest used ChatGPT to help solve a 42-year-old open problem, the difference between deep literature search and original mathematical discovery, and what changes when AI can work over longer timelines. Chapters01:27 The surprising progress of AI’s math capabilities 03:01 Solving an open problem with ChatGPT06:57 How models went from basic math to research level11:32 Why math matters for AGI14:26 AI and the Erdős problems21:26 Building an automated researcher28:19 The role of humans as models improve33:52 Verifying proofs with AI36:00 The risk of shallow understanding41:19 Advice for learning math with ChatGPT

What does it take to build AI systems that can actually help scientists? Research lead Joy Jiao and product lead Yunyun Wang discuss how OpenAI is developing models for life sciences and what responsible deployment means in a field with real biosecurity stakes. They explore how AI is already improving research workflows and where it could lead in drug discovery and more autonomous labs — including why a future with less pipetting sounds pretty good to most scientists.Chapters0:39 Introducing the Life Sciences model series3:47 Joy’s path into life sciences5:00 Autonomous lab with Ginkgo Bioworks7:27 Yunyun’s path into life sciences8:12 OpenAI’s life sciences work9:48 Biorisk, access, and safeguards15:43 What models can do in the lab17:51 Building scientific infrastructure20:14 Why compute matters for science24:54 Where are we in 6-12 months?29:51 Scientific adoption and skepticism33:17 Advice for students and researchers40:27 Where are we in 10 years?

The more AI can do, the more we need to ask what it should and shouldn’t do. In this episode, OpenAI researcher Jason Wolfe joins host Andrew Mayne to talk about the Model Spec, the public framework that defines intended model behavior. They discuss how the Model Spec works in practice, including how the chain of command handles  conflicts between instructions, and how OpenAI evolves it based on feedback, real-world use, and new model capabilities.Chapters00:00 Introduction01:10 What is the Model Spec?03:55 How does the Model Spec work in practice?06:26 Transparency: Where to read the Model Spec & give feedback07:51 How did the Model Spec originate?10:02 How does the spec translate into model behavior?11:26 What is the hierarchy / chain of command?13:35 Handling edge cases like Santa Claus17:41 How does the Model Spec evolve over time?19:59 What happens when models disagree with the spec?22:05 How do smaller models follow the spec?23:16 Is chain-of-thought useful for alignment?24:16 Model Spec vs Anthropic’s Constitution26:28 What surprised you most?26:56 How do you define the scope of the spec?27:44 What is the future of the Model Spec?31:16 How should developers think about the spec?34:44 Asimov’s laws vs Model Spec37:16 Could AI write a Human Spec?

Healthcare systems around the world are under strain, and both patients and clinicians are feeling the impact. OpenAI's Head of Health Dr. Nate Gross and Karan Singhal, who leads Health AI Research, discuss how AI can help address the biggest challenges. They cover how OpenAI is training models to handle sensitive health questions in collaboration with physicians, and how that foundation is unlocking a new generation of tools for patients, clinicians, and healthcare systems.Chapters00:00:38 – Origins of Nate and Karan’s interest in AI and healthcare00:05:01 – Strategy for building AI tools for clinicians00:06:57 – How AI models are trained for health use cases00:10:15 – How OpenAI is able to score well on health evals00:14:21 – Key challenges deploying AI in healthcare00:21:05 – Collaboration with hospitals and healthcare systems00:23:05 – Practical everyday uses of AI health assistants00:26:43 – Biggest “wow” moment during development00:28:46 – Feedback from clinicians and early users

How should advertising work in an AI product? Asad Awan, one of the ad leads at OpenAI, walks through how the company is approaching this decision and why it’s testing ads in ChatGPT at all. He explains how ads are built to stay separate from the model response, keep conversations with ChatGPT private from advertisers, and give people control over their experience.Chapters00:00:29 — Mission and principles00:04:01 — Separation between ads and answers00:07:31 — Who will see ads00:08:52 — Internal input and decision-making process00:11:06 — Controls and how ads will work00:15:53 — Guardrails for sensitive conversations00:17:33 — Skepticism about ads00:20:26 — Helping small businesses00:24:13 — Future of ads

OpenAI CFO Sarah Friar and Khosla Ventures founder Vinod Khosla argue the greatest challenges in AI right now are keeping up with demand and making sure more people get the benefit. They unpack what's driving big investments in compute and why this moment is different from other technology cycles — with meaningful advances in health, agents, and robotics still ahead. Chapters00:00:00 — What’s the AI story of 2026?00:07:28 — AI in healthcare00:12:01 — Scaling compute to match revenue00:18:05 — Difference between now and dot-com bubble00:27:41 — Ads in ChatGPT00:30:05 — Will consumers have more than one AI subscription?00:36:41 — Winning in enterprise00:39:44 — How can startups succeed?00:44:05 — Robotics and beyond