Building a GTM Knowledge Graph Inspired by Karpathy's LLM Wiki
EXM7777 · x · 2026-08-19
Inspired by Andrej Karpathy's knowledge graph in Obsidian, the author shares how to build a Go-To-Market (GTM) system using AI agents for a one-person business.
Steps
- Vault Structure: One vault, folders per business domain (offer, buyers, market map, etc.).
- Note Format: One fact per note (50-150 words) with a speakable title and 4 frontmatter keys.
- Graph Connectivity: Link notes to folder hubs, siblings, and cross-folder notes to create a connected web.
- Rulings Folder: Store corrections as dated lines; agents read them before drafting to make corrections permanent.
- Workflow: Claude Code reads hubs, greps rulings, and drafts content.
More from coding & agent
- Holistic Evaluation of Memory Substrates in Long-Horizon Agents — Wei-Chieh Huang · 2026-08-19
- Fable One-Shots Old Minecraft Mod Port to Newer Version — repligate · 2026-08-19
- Is Agent Continuity About Preserving State or Coherent Trajectories? — National_Actuator_89 · 2026-08-19
- Databricks competition winners share lessons: parsing unsolved, build on coding agents — EchoShao8899 · 2026-08-19
- Langfuse Deep Dive: How This Open Source LLM Observability Platform Solves Production "Magic" — 大模型之路 · 2026-08-19
- GitHub Outage Inspires Durable Objects-based Git Forge — tobowers · 2026-08-19