NEXUS, not vector database
Typed, cited, human-editable entities beat an opaque cosine-similarity index.
What a vector DB actually is
Pinecone, Weaviate, Qdrant — beautiful engineering, and structurally the wrong abstraction for grounded intelligence. A vector is a compressed lossy fingerprint of text. When your top-k retrieves the wrong chunk, you cannot ask the vector why.
NEXUS is entities
In NEXUS, a fact is not a vector. It is a typed entity with explicit relations, provenance, version, and supersession, stored as human-readable DTL under git. You can grep it. You can pull-request a correction. You can diff yesterday against today.
Sub-five-millisecond lookup
The critique against typed substrates is that they are slow. Ours is not. NEXUS returns in under five milliseconds against ninety-thousand-plus entries, because retrieval is a compiled inverted index, not a similarity search.
Why we use different words
Where the industry says LLM, we say CLLM. Where it says vector database, we say NEXUS. Where it says Chain-of-Thought, we say PRISM proof-tree. Where it says RAG, we say LATTICE. Where it says trillion-token dataset, we say Smart Router Dataset. Different words because different architecture. Read the whitepaper.