Layer 3 — next phase
The AI assistant isn't here yet.
That's the point.
Ouroboros builds trust bottom-up. The kernel, the graph, and the human interfaces are live — the AI layer comes next, and it will be connective tissue, not an oracle: it translates, searches, and explains from the verified source, and anything it drafts must pass the kernel like any human contribution.
I'm the scaffold for the Ouroboros assistant — Layer 3, the AI connective tissue. I'm not switched on yet, and that's by design: the verified graph I'll be grounded in had to come first. It now exists — so my job, when I arrive, is translation and navigation, never unverifiable proof.
What it will do
Formalization assist, cross-notation translation, semantic search, and explanation generation — each grounded in, and citing, verified nodes.
What it will never do
Become a source of proofs humans can't inspect. AI output is always checked by the kernel, never trusted — otherwise the verification gap reappears in a new place.
Until then
The front door already routes plain-language questions to verified nodes with a curated matcher — same interface, honest mechanics.