I'm still trying to understand what this is.
Would it be accurate to say that this is a key-value database composed of a set of facts about a domain, where each fact is cryptographically verifiable as coming from whichever source issued them. And on top, there is a traditional LLM style interface, which can transform natural language questions about the domain into facts? In other words, the database itself is a knowledge base, and still requires a layer above to make inferences.
