Search has always been a hard problem. Once people started writing things down, they also needed ways to find the right thing again. That challenge moved from libraries to databases, from databases to the web, and now from web search into AI systems that can retrieve, reason, and act.

In this session, I’ll walk through a brief history of search in computing and show why each generation solved a different problem. We’ll cover:

• How exact matching and Boolean search helped people query structured collections
• How TF IDF and BM25 made large document collections easier to rank
• How semantic search finds related meaning when the same words are missing
• How RAG helps turn retrieved context into grounded answers
• How agents use search, memory, tools, and verification to work through larger tasks

We’ll also look at recent examples from mathematical research, where agentic systems used theorem retrieval, literature search, and formal verification to make progress on open problems.

We’ll finish with Redis 8 and Redis Search as the low latency and high throughput layer for modern retrieval. You’ll see how to index application data, combine text search with filters, run vector similarity queries, and serve search results fast enough for real time user experiences and AI workflows.

Participants will leave with a practical mental model for keyword search, semantic search, RAG, and agentic retrieval, plus a Redis based path for adding efficient search to real applications.