Your Second Brain Is a Graveyard. Make It Agent Memory.

2025-02-06 · Source: Decoding AI Magazine · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Robotics & Autonomous Systems · Depth: Intermediate, long

Summary

The article introduces an "AI Research OS," an LLM-maintained wiki designed to serve as an agent memory layer, addressing the limitations of traditional "second brains" and context window-bound LLM harnesses like Codex or NotebookLM. Developed by Louis-François Bouchard and the author, this system converts personal notes (e.g., 10,994 notes across Obsidian, Readwise, Notion, Google Drive) and web content into a dynamic, queryable knowledge base. The architecture evolved from static `research.md` outputs to a living LLM wiki, inspired by Andrej Karpathy's concept and Google's Open Knowledge Format. It operates without complex databases, utilizing plain files organized into `raw`, `wiki`, and `index` folders, enabling a human-inspectable knowledge graph viewable in Obsidian. The system supports progressive disclosure for querying and expands incrementally through new data ingestion, deep research rounds, or user interactions, while maintaining project-specific scope via the PARA method to preserve the integrity of the original second brain. The open-source implementation is available as a Claude Code plugin.

Key takeaway

For AI Engineers struggling with inert personal research notes, consider implementing an LLM wiki as your agent's memory layer. This approach transforms static "second brain" content into a dynamic, queryable knowledge base, enhancing agent performance for research, coding, and content creation. You should clone the open-source AI Research OS workshop to integrate your existing notes and web sources, enabling agents to access and evolve project-specific knowledge without complex database infrastructure.

Key insights

An LLM wiki acts as a dynamic agent memory, transforming static notes into a queryable, evolving knowledge base.

Principles

• LLM-maintained wikis offer persistent, interlinked knowledge.
• Project-scoped wikis balance utility and cost.
• Plain files can form a powerful, inspectable knowledge base.

Method

The deep research algorithm uses an orchestrator and sub-agents for query rounds, ranking sources, and scraping top-K results. This feeds into an LLM wiki that generates entities, concepts, and comparisons.

In practice

• Integrate personal notes (Obsidian, Readwise) with web sources.
• Use a file-based index for agent retrieval.
• Scope wikis to projects using the PARA method.

Topics

• LLM Wiki
• Agent Memory
• Personal Knowledge Management
• Deep Research Algorithms
• Open Knowledge Format
• Obsidian Integration

Code references

• iusztinpaul/ai-research-os-workshop

Best for: AI Engineer, Machine Learning Engineer, AI Architect

Related on AIssential

• Wiki Code Memory: Giving My Coding Agent a Real Memory (and Cutting Token Costs While I’m At It) — Combining episodic and semantic memory layers creates a cost-effective, persistently knowledgeable coding agent.
• Building my own LLM-Wiki Research Team — Treat notes like code and compile them into a persistent, compounding markdown wiki using an LLM Agent to avoid rediscovering insights.
• Papers You Should Know About — AI research is advancing LLM efficiency and agent autonomy through novel architectures, self-improvement mechanisms, and real-world applications.
• Your Second Brain Doesn’t Need RAG. It Needs a Map — The "second brain" benefits from structured markdown maps and long LLM context windows more than RAG.
• Give Your AI Unlimited Updated Context — An LLM-maintained personal wiki creates a persistent, compounding knowledge artifact, eliminating repetitive context setup.
• Karpathy Named It. I Built One on My Notes. — A private LLM knowledge base can outperform RAG for personal research using CLIs and filesystem-based memory.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Decoding AI Magazine.