{"slug":"rag-pgvector","label":"How does this blog do vector search?","events":[{"text":"What does this blog say about using PostgreSQL for vector search?","type":"user_message","ts_offset_ms":0},{"args":{"query":"PostgreSQL vector search pgvector"},"name":"blog_search","type":"tool_call","ts_offset_ms":2363},{"name":"blog_search","type":"tool_result","result":{"query":"PostgreSQL vector search pgvector","total":20,"results":[{"id":29,"slug":"pgvector-postgresql-vector-database","tags":[],"score":0.03278688524590164,"title":"PostgreSQL as a Vector Database: pgvector in Production","views":146,"summary":"You don't need a separate vector database. pgvector turns PostgreSQL into a semantic search engine — with HNSW indexes, hybrid queries, and full SQL power.","created_at":"2026-04-27T12:00:00Z","match_type":"hybrid","updated_at":"2026-07-07T14:57:34.478635Z","header_image":"/static/articles/pgvector-postgresql-vector-database/header.webp","reading_time":"7 min read"},{"id":8,"slug":"building-rag-pipeline","tags":[],"score":0.03200204813108039,"title":"Building a RAG Pipeline from Scratch","views":67,"summary":"A step-by-step technical guide to constructing a Retrieval-Augmented Generation system for enterprise data.","created_at":"2025-12-22T12:00:00Z","match_type":"hybrid","updated_at":"2026-07-07T14:48:35.749354Z","header_image":"/static/articles/building-rag-pipeline/header.webp","reading_time":"8 min read"},{"id":19,"slug":"knowledge-graphs-and-llms","tags":[],"score":0.030536130536130537,"title":"Knowledge Graphs Meet LLMs: Structured Reasoning at Scale","views":96,"summary":"Hallucinations happen when models guess. Knowledge graphs give LLMs a structured, verifiable backbone — and the combination is more powerful than either alone.","created_at":"2026-03-16T12:00:00Z","match_type":"hybrid","updated_at":"2026-07-07T14:44:21.829720Z","header_image":"/static/articles/knowledge-graphs-and-llms/header.webp","reading_time":"5 min read"},{"id":15,"slug":"fine-tuning-vs-rag","tags":[],"score":0.029850746268656716,"title":"Fine-Tuning vs RAG: Choosing the Right Strategy","views":65,"summary":"Both approaches customize LLMs for specific domains, but they solve different problems. Here's how to pick the right tool — and when to combine them.","created_at":"2026-03-02T12:00:00Z","match_type":"hybrid","updated_at":"2026-07-07T14:44:22.013481Z","header_image":"/static/articles/fine-tuning-vs-rag/header.webp","reading_time":"5 min read"},{"id":16,"slug":"gemini-rag-integration","tags":[],"score":0.02967032967032967,"title":"Building a RAG Pipeline with Gemini","views":52,"summary":"A practical guide to building a Retrieval-Augmented Generation pipeline using Google's Gemini API, File Search Store, and embedding models.","created_at":"2026-01-05T12:00:00Z","match_type":"hybrid","updated_at":"2026-07-07T15:19:18.461533Z","header_image":"/static/articles/gemini-rag-integration/header.webp","reading_time":"7 min read"},{"id":39,"slug":"visualizing-latent-spaces","tags":[],"score":0.028693528693528692,"title":"Visualizing High-Dimensional Latent Spaces","views":81,"summary":"Latent spaces encode meaning in geometry. Learn how dimensionality reduction techniques like t-SNE and UMAP make these invisible structures visible.","created_at":"2026-01-12T12:00:00Z","match_type":"hybrid","updated_at":"2026-07-07T14:44:23.034688Z","header_image":"/static/articles/visualizing-latent-spaces/header.webp","reading_time":"6 min read"},{"id":20,"slug":"long-context-windows-million-tokens","tags":[],"score":0.027651515151515153,"title":"Long Context Windows: What 1 Million Tokens Actually Changes","views":94,"summary":"Gemini 1.5 Pro pushed context to 1M tokens. Claude 3.5 followed. But is a bigger context window always better, and what does it really enable?","created_at":"2026-03-23T12:00:00Z","match_type":"hybrid","updated_at":"2026-07-07T14:44:21.572463Z","header_image":"/static/articles/long-context-windows-million-tokens/header.webp","reading_time":"5 min read"},{"id":41,"slug":"react-agent-loop-explained","tags":[],"score":0.026875901875901876,"title":"How a ReAct Agent Loop Actually Works","views":140,"summary":"A walkthrough of one real ReAct agent run — the query was \"Houston, we have a problem\" and the agent returned the exact YouTube timestamp where Tom Hanks delivers the line. You can replay the recorded run right inside this article, thinking tokens and all.","created_at":"2026-06-01T12:00:00Z","match_type":"hybrid","updated_at":"2026-07-07T14:48:36.459364Z","header_image":"/static/articles/react-agent-loop-explained/header.webp","reading_time":"11 min read"},{"id":10,"slug":"claude-code-ai-assisted-development","tags":[],"score":0.026491228070175437,"title":"Claude Code and the Future of AI-Assisted Development","views":141,"summary":"Claude Code brings an AI agent directly into your terminal. Explore what autonomous coding tools mean for software engineering workflows.","created_at":"2026-04-20T12:00:00Z","match_type":"hybrid","updated_at":"2026-07-07T14:44:20.317780Z","header_image":"/static/articles/claude-code-ai-assisted-development/header.webp","reading_time":"6 min read"},{"id":22,"slug":"mcp-servers-model-context-protocol","tags":[],"score":0.0259915611814346,"title":"MCP Servers: Building Tool-Using AI with Model Context Protocol","views":117,"summary":"The Model Context Protocol standardizes how AI models discover and use tools. Here's how MCP servers work and why they matter for the agentic future.","created_at":"2026-04-06T12:00:00Z","match_type":"hybrid","updated_at":"2026-07-07T14:48:34.378919Z","header_image":"/static/articles/mcp-servers-model-context-protocol/header.webp","reading_time":"6 min read"},{"id":45,"slug":"ai-that-watches-the-world-cup","tags":[],"score":0.025478740668614087,"title":"I Put an AI in My World Cup Pool. Here's How It Actually Did.","views":48,"summary":"A bot that reads live match status off the open web, turns a qualitative read of each game into a scoreline, and gets graded against real results every day. The group stage is over, all 72 matches are graded, and the bot is on top of the pool.","created_at":"2026-06-28T00:00:00Z","match_type":"hybrid","updated_at":"2026-07-07T14:44:16.980111Z","header_image":"/static/articles/ai-that-watches-the-world-cup/header.webp","reading_time":"7 min read"},{"id":38,"slug":"vector-databases-ai-memory","tags":[],"score":0.016129032258064516,"title":"Vector Databases: The Memory Layer of Modern AI","views":74,"summary":"Embeddings are the language of semantic search, and vector databases are where they live. Learn how ANN indices power retrieval at scale.","created_at":"2025-10-13T12:00:00Z","match_type":"hybrid","updated_at":"2026-07-07T14:57:33.777753Z","header_image":"/static/articles/vector-databases-ai-memory/header.webp","reading_time":"7 min read"},{"id":7,"slug":"building-ai-powered-personal-websites","tags":[],"score":0.015625,"title":"Building AI-Powered Personal Websites","views":142,"summary":"Your portfolio site can do more than display static content. Learn how to integrate AI chat, RAG, and agentic tools into a personal website.","created_at":"2026-05-18T12:00:00Z","match_type":"hybrid","updated_at":"2026-07-07T14:48:37.140995Z","header_image":"/static/articles/building-ai-powered-personal-websites/header.webp","reading_time":"7 min read"},{"id":26,"slug":"navigating-latent-space","tags":[],"score":0.014705882352941176,"title":"Navigating the Latent Space","views":66,"summary":"Unlock the secrets of high-dimensional data compression and how mapping complex inputs to simpler spaces drives generative AI.","created_at":"2025-10-20T12:00:00Z","match_type":"hybrid","updated_at":"2026-07-07T14:44:24.600726Z","header_image":"/static/articles/navigating-latent-space/header.webp","reading_time":"5 min read"},{"id":40,"slug":"visualizing-word-embeddings","tags":[],"score":0.014492753623188406,"title":"Visualizing Word Embeddings","views":47,"summary":"Explore the math behind language: see how 'King - Man + Woman = Queen' plays out visually in vector space.","created_at":"2025-10-27T12:00:00Z","match_type":"hybrid","updated_at":"2026-07-07T14:41:35.693318Z","header_image":"/static/articles/visualizing-word-embeddings/header.webp","reading_time":"5 min read"},{"id":3,"slug":"ai-agents-autonomous-action","tags":[],"score":0.014285714285714285,"title":"AI Agents: Beyond Chat to Autonomous Action","views":85,"summary":"LLMs were just the beginning. Explore how autonomous AI agents use tools, memory, and planning to act — not just respond.","created_at":"2025-09-15T12:00:00Z","match_type":"hybrid","updated_at":"2026-07-07T15:12:32.138520Z","header_image":"/static/articles/ai-agents-autonomous-action/header.webp","reading_time":"6 min read"},{"id":24,"slug":"multimodal-models-see-and-speak","tags":[],"score":0.014084507042253521,"title":"The Multimodal Moment: When AI Learned to See and Speak","views":60,"summary":"Vision-language models can now describe images, read charts, and watch videos. Here's how the architecture enables true cross-modal understanding.","created_at":"2025-09-29T12:00:00Z","match_type":"hybrid","updated_at":"2026-07-07T14:44:24.860654Z","header_image":"/static/articles/multimodal-models-see-and-speak/header.webp","reading_time":"6 min read"},{"id":2,"slug":"agentic-workflows-multi-step-ai","tags":[],"score":0.0136986301369863,"title":"Agentic Workflows: Orchestrating Multi-Step AI Pipelines","views":119,"summary":"Single-prompt AI is hitting its ceiling. Agentic workflows chain multiple LLM calls with tools, branching, and feedback loops to tackle complex tasks reliably.","created_at":"2026-04-13T12:00:00Z","match_type":"hybrid","updated_at":"2026-07-07T15:12:31.354415Z","header_image":"/static/articles/agentic-workflows-multi-step-ai/header.webp","reading_time":"6 min read"},{"id":18,"slug":"inside-the-transformer","tags":[],"score":0.013513513513513514,"title":"Inside the Transformer","views":55,"summary":"Peel back the layers of the Transformer architecture to visualize how attention heads process information.","created_at":"2025-11-17T12:00:00Z","match_type":"hybrid","updated_at":"2026-07-07T14:44:18.884055Z","header_image":"/static/articles/inside-the-transformer/header.webp","reading_time":"6 min read"},{"id":21,"slug":"mapping-star-wars-universe","tags":[],"score":0.012987012987012988,"title":"Mapping the Star Wars Universe","views":46,"summary":"A fun data experiment: embedding Star Wars characters in 3D space based on their Force sensitivity and factions.","created_at":"2025-11-10T12:00:00Z","match_type":"hybrid","updated_at":"2026-07-07T14:44:24.037643Z","header_image":"/static/articles/mapping-star-wars-universe/header.webp","reading_time":"5 min read"}]},"ts_offset_ms":4673},{"type":"thinking_start","ts_offset_ms":6613},{"text":"**Analyzing Potential Options**\n\nI've homed in on a promising lead: \"PostgreSQL as a Vector Database: pgvector in Production\". This resource seems directly applicable, outlining pgvector, HNSW indexing, and hybrid queries, essentially PostgreSQL as a semantic search engine. There's another RAG piece that may fit.\n\n\n","type":"thinking_chunk","ts_offset_ms":6613},{"type":"thinking_end","ts_offset_ms":7173},{"text":"The latentSource blog highlights that you can use PostgreSQL","type":"chunk","ts_offset_ms":7173},{"text":" as a fully-featured vector database using the **pgvector** extension. This approach allows you to transform a standard relational database into a semantic","type":"chunk","ts_offset_ms":7333},{"text":" search engine without the complexity of maintaining a separate specialized vector store.\n\nKey insights from the blog include:\n*   **Production","type":"chunk","ts_offset_ms":7500},{"text":" Readiness:** PostgreSQL is presented as a viable alternative to dedicated vector databases for production workloads.\n*   **Indexing:** It","type":"chunk","ts_offset_ms":7672},{"text":" supports advanced indexing techniques like **HNSW (Hierarchical Navigable Small World)** to speed up high-dimensional vector searches.\n","type":"chunk","ts_offset_ms":7827},{"text":"*   **Hybrid Queries:** One of the main advantages is the ability to run \"hybrid queries,\" combining semantic vector search with traditional SQL power","type":"chunk","ts_offset_ms":7983},{"text":" (like filtering by metadata or joining with other tables) in a single request.\n*   **Simplified Architecture:** By using","type":"chunk","ts_offset_ms":8132},{"text":" `pgvector`, teams can keep their data and embeddings in one place, leveraging existing PostgreSQL knowledge and infrastructure.\n\n**Read more","type":"chunk","ts_offset_ms":8281},{"text":":**\n*   [PostgreSQL as a Vector Database: pgvector in Production](/articles/pgvector-postgresql-","type":"chunk","ts_offset_ms":8353},{"text":"vector-database)\n*   [Vector Databases: The Memory Layer of Modern AI](/articles/vector-databases-","type":"chunk","ts_offset_ms":8424},{"text":"ai-memory)\n*   [Building a RAG Pipeline from Scratch](/articles/building-rag-pipeline)","type":"chunk","ts_offset_ms":8525},{"type":"done","ts_offset_ms":8569}]}