About this talk
This talk explores the concept of AI-native databases, advocating for a unified system rather than traditional distributed architectures that complicate data consistency and security. The speaker discusses Oracle AI Database and presents four key ideas: converged architecture to eliminate data movement issues, JSON relational duality for seamless data representation, trusted AI that validates model outputs, and end-to-end security that enforces access control directly within the database. By challenging current practices, this presentation highlights the potential for better efficiency and safety in AI applications.
