In this episode of the OpenAI Podcast, Andrew Mayne interviews Kevin Weil, head of OpenAI for Science, and Alex Lupsasca, an OpenAI research scientist and physics professor, about the impact of AI on scientific research. They discuss OpenAI's initiative to accelerate scientific discovery using advanced AI models like GPT-5, highlighting examples of AI assisting in mathematical proofs, literature searches, and physics problems. Lupsasca shares his experience of using GPT-5 to solve a complex physics problem related to black hole symmetries, which led to his involvement with OpenAI. The conversation also covers the importance of collaboration between humans and AI, the challenges of working with AI models at the frontier of their capabilities, and the potential for AI to revolutionize various scientific fields, including drug discovery and fusion energy. They also touch upon a forthcoming research paper detailing the current state of GPT-5 in science, and offer advice to students and researchers on leveraging AI in their work.
Sign in to continue reading, translating and more.
Continue