Learn

The ideas behind sovereign memory, in plain language.

What organizational memory is, what a Domain Language Model does, how context is assembled, and why the standard underneath is free.

01Organizational memory for AI: what it is, and how to build oneEvery model you use should reason from what your business actually knows. That takes a memory the business owns, not one the model keeps.Read
02What is a Domain Language Model (DLM)?Not a fine-tuned model. A stateless, interchangeable AI reading and writing to a sovereign repository your organization owns.Read
03Institutional memory and large language modelsAn LLM reasons over what it is given. Most organizations have never written down what they know in a form a model can read.Read
04Context engineering for business: assemble context before the model callThe model is only as good as what it’s handed. Context engineering is deciding, by address rather than by search, what that is.Read
05AI data sovereignty for companies: contractual, or architectural?Sovereignty is usually argued at the compute layer. For a business, the layer that matters is its knowledge, and who holds it.Read
06Universal Cognitive Architecture (UCA), explainedA free classification standard that gives every piece of business knowledge a permanent address any AI can read. The Dewey Decimal System for business knowledge.Read
07AI knowledge management without lock-inSwitching models should be a choice in a picker, not a migration. That takes knowledge kept outside the model, in plain text you can take with you.Read