While it’s not often acknowledged, much of the progress in technology is underpinned by the idea that “everything old is new again” and vector databases, a concept that is many years old, are having a moment in generative AI. Analyst James Curtis joins host Eric Hanselman to dig into why this powerful idea has become so valuable in large language models and other aspects of AI. While it’s possible to overlay vector functions on a traditional database for data access, it may not be advisable.