2 guides
Autoresearch
ArchitectureObsidianAi Agents
Designing a Second Brain for AI Agents: The Vault-as-Database Pattern
How to architect a local knowledge base that AI agents can reason over — starting from Karpathy's three-folder reference architecture, through the three-layer memory stack, MCP bridges, quality maintenance, and the scaling path from flat files to SQLite.
AutoresearchClaudeAi Agents
What Is Autoresearch and How to Use It
Autoresearch is Karpathy's hill-climbing optimization loop — make one change, test against a binary checklist, keep or revert, repeat. This guide covers the core method, practical implementation with Claude skills, scaling to parallel GPU clusters and distributed agent swarms, and the emerging pattern of full-stack AI research agents.