Jeff Huber of Chroma joins us to talk about what actually matters in vector databases in 2025, why “modern search for AI” is different, and how to ship systems that don’t rot as context grows.

Full show notes: https://www.latent.space/p/chroma

Full Video Episode

Timestamps

00:00 Introductions00:48 Why Build Chroma02:55 Information Retrieval vs. Search04:29 Staying Focused in a Competitive AI Market08:08 Building Chroma Cloud12:15 Context Engineering and the Problems with RAG16:11 Context Rot21:49 Prioritizing Context Quality27:02 Code Indexing and Retrieval Strategies32:04 Chunk Rewriting and Query Optimization for Code34:07 Transformer Architecture Evolution and Retrieval Systems38:06 Memory as a Benefit of Context Engineering40:13 Structuring AI Memory and Offline Compaction45:46 Lessons from Previous Startups and Building with Purpose47:32 Religion and Values in Silicon Valley50:18 Company Culture, Design, and Brand Consistency52:36 Hiring at Chroma: Designers, Researchers, and Engineers

Get full access to Latent.Space at www.latent.space/subscribe