Why Data Access is Bottlenecking AI Development w/ Protege CEO Bobby Samuels The internet has been scraped. Protege is building the infrastructure layer for the high-quality, real-world data models need to continue advancing. CEO Bobby Samuels sat down with a16z's Daisy Wolf and Eva Steinman to discuss the myth that we've run out of data for AI, how Protege connects healthcare systems and other data holders with the major AI labs, and the advantages of real-world data over synthetic data, and more. 0:00 Introduction 1:27 The data scarcity myth: Are we really running out of data? 2:59 Building data networks: Unlocking proprietary sources 5:39 Compliance & ethics: The legal reality of AI data 9:56 Real-world vs. synthetic data: What actually works? 11:00 Healthcare data infrastructure: Multimodal patient data explained 16:15 Is data the real bottleneck in AI? 19:57 Why build a marketplace for AI data? 24:10 The most exciting applications in healthcare AI 28:50 Beyond healthcare: Expanding into video, audio & more 31:19 What’s next: Agentic workflows, world models & biology @BobbySamuels @daisydwolf