Context Engineering and Agentic Search with Chroma CEO

Explore the critical concept of "context rot" and its impact on AI systems in this 45-minute podcast episode featuring Jeff Huber, CEO and cofounder of Chroma, discussing context engineering and agentic search with host Demetrios Brinkmann. Discover how AI memory slowly decays over time, breaking even the most sophisticated models, and learn why most AI systems struggle to retain what matters most. Examine the hidden limitations of context windows, the chaos in current retrieval systems, and how Chroma is revolutionizing the entire retrieval stack for modern AI applications. Delve into the evolution of search eras, from traditional keyword-based approaches to semantic search and beyond, while understanding the specific memory challenges faced by AI agents during task execution. Investigate the limitations of current semantic search implementations and explore the concept of "context hygiene" - maintaining clean, relevant information flow in AI systems. Learn about Chroma's on-device functionality and their vision for building precision-focused AI systems that can reliably access and utilize contextual information. Gain insights into the practical challenges of ML model deployment and the infrastructure requirements for building robust retrieval systems that support next-generation AI applications.