{"href":"https://api.simplecast.com/oembed?url=https%3A%2F%2Fai-a16z.simplecast.com%2Fepisodes%2Fbuilding-self-accelerating-ai-with-mirendil-rt3KPYBk","width":444,"version":"1.0","type":"rich","title":"Building Self-Accelerating AI with Mirendil","thumbnail_width":300,"thumbnail_url":"https://image.simplecastcdn.com/images/89450696-5007-4713-8168-eea085e49626/80ad2223-d8fa-4bb2-9645-8e2c1168a372/ai-a16z-podcast-cover-art-2.jpg","thumbnail_height":300,"provider_url":"https://simplecast.com","provider_name":"Simplecast","html":"<iframe src=\"https://player.simplecast.com/0e476d5b-0145-4529-ba37-159a9e705e69\" height=\"200\" width=\"100%\" title=\"Building Self-Accelerating AI with Mirendil\" frameborder=\"0\" scrolling=\"no\"></iframe>","height":200,"description":"Matt Bornstein speaks with Mirendil cofounders Behnam Neyshabur and Harsh Mehta about their vision for building self-accelerating AI.\nAfter leading research efforts at Google and Anthropic, the founders started Mirendil around a simple question: what happens when AI systems can meaningfully contribute to their own development? Rather than focusing solely on AI as a tool for productivity, they argue that the most important application may be accelerating scientific and technological progress itself.\nThe conversation explores AI research, scaling laws, automated engineering, scientific discovery, and the challenges of building systems that can improve over time. They discuss the future of AI-assisted research, why they believe scientific progress remains bottlenecked by intelligence, and how more capable AI systems could help unlock advances across medicine, engineering, and the natural sciences.\n"}