Topic
Pinecone
3 episodes
-
The RAG era is ending for agentic AI — a new compilation Stage knowledge layer is what comes next
Pinecone just announced Nexus, a 'knowledge engine' that shifts reasoning from inference time to a compilation stage — meaning agents get pre-built, task-specific knowledge artifacts instead of rediscovering context from scratch every session. Justy and Cody dig into why RAG was never really built for agents, what the architecture actually does, and whether the 98% token reduction claim holds water.
-
Vector Databases Explained in 3 Levels of Difficulty MachineLearningMastery
Izzo and Boone decode vector databases from basic similarity search to production-scale indexing algorithms like HNSW and IVF, explaining how they solve the core problem of searching unstructured data at scale.
-
SurrealDB 3.0 wants to replace your five database RAG stack with one
SurrealDB 3.0 combines vector search, graph traversal, and relational queries into a single transactional database engine, aiming to replace the complex multi-database stacks commonly used in RAG systems. The Rust-native architecture stores agent memory as graph relationships directly in the database with full ACID guarantees across distributed nodes.