The latest from RA-H
Insights on knowledge management, AI-powered research, and the future of local-first thinking.
Own Your Data: Organise It Into a Database for AI Agents
As more work moves through agents, the durable asset is the context they read from. That context should live in a local database you own.
Published on Apr 25, 2026
Build Your Own AI Memory System
A practical walkthrough of setting up RA-H OS: a local SQLite knowledge graph that Claude Code, Codex, Cursor, and other agents can read from and write back to through MCP.
Published on Apr 21, 2026
Karpathy's LLM Knowledge Base Is Right. It Should Be a Database.
Karpathy's markdown knowledge-base architecture points at the right future: LLMs manipulating knowledge. The stronger backbone is a local SQLite graph.
Published on Apr 9, 2026
Don't Build a Second Brain. Build a Database.
Traditional PKM is useful, but agent-era knowledge work needs a local structured database that agents can query, update, and connect.
Published on Apr 8, 2026
How I Built a Team of AI Agents to Run My Life
A practical breakdown of the Zeu-style operating system: one orchestrator, project agents, requirement documents, skills, and a local RA-H memory layer.
Published on Mar 24, 2026
What's New: Light Mode + Bundled Skills
A couple of updates shipped over the past few weeks — light mode is here, and new installs now come with a set of pre-built skills to get you started faster.
Published on Mar 18, 2026
Structure is Good: How I Built a Wiki-Base (Context Engine) for the Latent Space Community
The file-vs-graph debate misses the point. There is only the context window — and structure is what makes filling it efficiently possible.
Published on Mar 14, 2026
How I Get More From AI Than Most People
The real difference is not the model, the prompt, or the app. It is whether you own a structured context graph that your agents can read from and write back to.
Published on Mar 5, 2026
Build Your Own Knowledge Base (Open Source)
A step-by-step walkthrough of how to set up your own local knowledge graph using RA-H's open source tools — clone, install, and start building your external context in minutes.
Published on Feb 16, 2026
Why Obsidian and Markdown is the wrong approach for personal knowledge management
If agents are going to manage your knowledge, you should be building systems for agents, not humans. Markdown is the wrong foundation. SQLite is the better backbone.
Published on Jan 9, 2026
Why Your AI Intuition Is Broken (And How to Fix It)
We default to treating language models like people, but that's a mistake. Here's how to develop the right intuition for working with AI systems.
Published on Dec 16, 2025
Why Your Context Is So Important
Your context is becoming one of the most valuable assets in how you interact with AI. Here's why it matters, why you should protect it, and how.
Published on Dec 8, 2025
How We Should Think About Human/AI Collaboration
The way we understand and work with AI is becoming more important than domain expertise. Here's how to build richer interactions with language models.
Published on Dec 6, 2025
RA-H Now Connects to External Chat Apps
Connect Claude, ChatGPT, Cursor, and other MCP-enabled tools directly to your local RA-H knowledge base for seamless research integration.
Published on Dec 5, 2025
Externalise Your Thinking
Building your own structured, linkable and evolving knowledge base is the best way to externalise and improve your thinking and research.
Published on Dec 4, 2025
We Just Shipped Real-Time Voice to RA-H
Why real-time voice changes everything for knowledge work - and how the phonological loop makes you a better thinker when you talk instead of type.
Published on Nov 26, 2025
You Should Be Talking to AI, Not Typing
Why real-time voice interfaces are the cognitive enhancement we've been overlooking - and how they fundamentally change how we think and learn.
Published on Nov 18, 2025
Some Thoughts on Continual Learning
Why the way humans learn is weird, how AI tools are making us cognitively lazy, and how RA-H supports natural learning processes instead of replacing them.
Published on Nov 8, 2025
Zettelkasten in the Age of AI
Why the traditional Zettelkasten method becomes more valuable, not less, when AI can seemingly think for us — and how to avoid the vendor lock-in trap.
Published on Oct 28, 2025
The Case for Owning and Growing Your Own Context
Why context ownership is the most important thing you can do in the AI era - and how to stop bleeding your valuable insights to vendor platforms.
Published on Oct 15, 2025