The lead story
Relational and document databases
Ria opens a relational drawer beside a document cabinet.
Notes from the work.
Explore the AI Agents research and comicMemori episodes.
42 articles, connected by curiosity.
Keep exploring
Vector databases, the map drawer at scale
Cosine similarity, comparing directions
Similarity search, finding neighbors on the meaning map
Embeddings, coordinates for meaning
Vectors, coordinates on the meaning map
RAG, the retrieve-then-write workflow- Retrieval, bringing the right pages back
- The archive, storing information outside the desk
- Compression, turning a pile into a clean brief
- Attention as a spotlight, why clutter hurts quality
- Context bloat, when the desk turns into a pile
Every post
- Relational and document databases
- Vector databases, the map drawer at scale
- Cosine similarity, comparing directions
- Similarity search, finding neighbors on the meaning map
- Embeddings, coordinates for meaning
- Vectors, coordinates on the meaning map
- RAG, the retrieve-then-write workflow
- Retrieval, bringing the right pages back
- The archive, storing information outside the desk
- Compression, turning a pile into a clean brief
- Attention as a spotlight, why clutter hurts quality
- Context bloat, when the desk turns into a pile
- The context window, desk size and boundaries
- Tokens, the sticky notes of text
- Personal Agents Are Coming, and Their Security Model Is the Open Problem
- How I Extracted the Muse Spark System Prompt on Launch Day
- The intern who writes fast, and what the desk is
- Production Agents Summit: The 40% Context Rule and Other Uncomfortable Truths
- Why 95% of Agentic AI Projects Fail: Key Takeaways from the Conference Everyone's Talking About
- The 30-40% Model Myth: Why Framework Architecture Matters More Than Model Capability
- The 30-40% Model Revelation: What a Leading Autonomous-Agent Startup Learned
- Good Agents: Implementing Plan-Verify-Execute for Multi-Agent Orchestration
- A Multi-Agent Framework Company's Lesson: Why Enterprise Topology Complexity Is Your Moat
- Repo Patcher: Using State Machines to Make AI Agents Reliable
- Building a Shopping Agent: Framework Bloat in Action
- How a Major Identity Company is Solving AI Agent Identity: The Next SSO Moment
- The Demo-Production Chasm: Why 70% Accuracy Creates False Expectations
- An AI Sales Intelligence Company's Approach: Building AI Sales Agents That Actually Work
- Inside an AI Agent Orchestration Company: Why 90% of Enterprise AI Agent Pilots Never Convert to Production
- An AI Infrastructure Company's Take: Managing Expectations with Probabilistic Systems
- Dual Memory Architecture: The Distinction Most AI Companies Miss
- Component-Level Evaluation: Why End-to-End Testing Fails for AI Agents
- The 40% Rule: Why Larger Context Windows Won't Save Your AI Agent
- The Evaluation Gap: Why 7 YC Companies Building Eval Tools Have Zero Adoption
- The Framework Paradox: Why 80-90% of Teams Abandon LangChain for Production
- The Coding Agent Exception: Why AI Works for Code but Struggles Everywhere Else
- Why 90% of AI Agent Pilots Fail Before Hitting Technical Limits
- What's Really Blocking AI Agents from Production? Insights from 36 Expert Interviews
- Handoff Rate: The North Star Metric for AI Agent Success
- The 92% Problem: Why System Integration Breaks AI Agent Deployments
- MCP Reality Check: The 25-Tool Accuracy Cliff No One Talks About
- How We Processed 44 Sources Into Actionable Research