Personal knowledge graphs and AI

The idea

A well-structured personal knowledge graph becomes an input that AI can leverage.

You have hundreds of notes. Maybe thousands. They contain genuine insight — hard-won understanding accumulated over years of reading, thinking, and working. You also have access to AI systems more capable than anything that existed two years ago.

And yet, when you ask an AI to help you think, it mostly ignores everything you know. It gives you generic answers drawn from its training data, not answers grounded in your specific understanding, your particular gaps, your unique combination of domains.

An example

You ask an AI assistant 'What should I study next?' Without your knowledge graph, it gives you a generic learning path. With your graph — 300 linked notes, typed edges showing prerequisites and contradictions, clusters revealing your domains and gaps — the AI can trace your actual understanding, identify the thinnest bridge nodes, and recommend the specific concept that would connect your two strongest clusters. Same AI, same question, radically different answer. The graph did not make the AI smarter. It gave the AI something structured to reason about.

Where it goes wrong

Dumping raw, unstructured notes into an AI and expecting graph-quality reasoning. If your notes are a flat pile of text with no explicit links, the AI has nothing to traverse. It will do its best with semantic similarity — finding notes that use similar words — but it cannot reason about relationships that were never made explicit. The failure mode is treating AI as a substitute for graph construction rather than as a tool that operates on top of a graph you have already built.

Make it stick

Choose your five most developed knowledge clusters. For each, identify the hub node and two to three key edges. Write these out as structured triples. Then give this skeleton to an AI assistant and ask: 'What questions could I explore that would create new edges between these clusters?' Use the AI's suggestions as candidate edges — not as truth, but as hypotheses worth investigating. You are using the AI to extend your graph, not to replace your judgment about what belongs in it.