Meaning is the edge

Nodes are facts. Edges are meaning.

A database stores facts. A knowledge graph stores facts and the relationships between them. That sounds like a small difference. It's the difference between a phone book and understanding. The phone book knows that Alan's number is 07700 900000. The graph knows that Alan called Sarah, Sarah works with Keith, Keith spoke at Highland Fling, and Highland Fling was organised by Alan. The facts are the same. The meaning is in the edges.

A fact is a node. Meaning is how you got there.

A database can't do that.

A database stores Alan's name, Jack's name, the conference date, the blog posts. Each fact is a row in a table. To find the connection between Alan going quiet and Jack going quiet, you'd need to know to look for it. You'd need to ask the right question. The database doesn't volunteer meaning. It answers queries.

A graph volunteers meaning. The edges exist whether you query them or not. The connection between Alan and Jack is already there — in the graph — because the edge was drawn when the relationship was recorded. You don't need to know to ask. You can discover by traversing.

Here's where it gets interesting.

The same graph can hold multiple world-views. Alan sees the Highland Fling as his conference — the thing he built, curated, ran. Jack sees it as a speaking gig — one event among many. Keith sees it as a platform — a career stepping stone. All three are true. All three coexist in the graph as different edges on the same nodes.

A database forces you to pick one interpretation — one schema, one truth. A graph holds them all. The meaning resolves by context: whose perspective are you viewing from? The graph doesn't have a single truth. It has edges, and the edges are valid from different vantage points.

This is what Intershapes does.

The surfacing layer — the thing that decides what to show you — is a router across a graph. It doesn't store one truth. It stores the edges between things, and when you ask a question, it traverses the graph from your context. Your perspective, your history, your task. The same graph, different path, different meaning.

It's not a search engine. A search engine finds nodes — pages that contain the words you typed. Intershapes finds edges — the relationships between things that make the answer meaningful. The difference between "here are ten pages about Alan" and "here's why Alan and Jack went quiet at the same time."

Trinity's routing head — 10,000 parameters — outperformed frontier models on decomposable tasks. Not because the router was smart. Because the router knew which specialist to ask. The intelligence wasn't in the routing. It was in the graph of specialists the router could reach.

Same principle. The edges — which specialist for which problem — are where the meaning lives. The nodes — the specialists themselves — are just capability. A router without a graph is a switch. A router with a graph is intelligence.

And it connects to everything else.

The graph is why sovereignty matters. If the meaning is in the edges, and the edges are your edges — your relationships, your context, your perspective — then the graph has to be yours. A cloud provider's graph is their meaning, not yours. Your knowledge graph, running on your infrastructure, holding your edges, resolving from your context — that's not just data sovereignty. That's meaning sovereignty.

The facts are public. The edges are personal. And the edges are where the intelligence lives.