Every defence against prompt injection you have read about lives in the prompt, the orchestrator or a filter in front of the data. All of them are software an attacker can talk past. There is one boundary a clever prompt cannot argue with: the agent's process never had the key. 8DB puts clearance and policy inside key derivation, so a request outside policy comes back as ciphertext no matter how persuasive it was. Here is how to build agents on that boundary.
6 min readTested claim · Holds with the network disabled · Evaluator bundle availableRead article ↗
Models are scored for accuracy and stores for throughput. The layer that actually decides whether an agent's answer can be acted on, whether its data knows where each fact came from, how sure to be, and when to say it does not know, has no score at all. We are proposing an open, vendor-neutral benchmark for that layer, and we want the people who would score badly on it in the room.
5 min readOpen proposal · Four metric families · Fixtures and scoring public before any scoreRead article ↗
Long-running agents are being given memory: a store of what they saw, decided and learned, carried across sessions and users. That store is now the most sensitive database in the building, and it is being built on vector indexes with no notion of who may recall a memory, where it came from, or how to make it gone. 8DB makes recall a cryptographic fact rather than a filter, gives every memory its pedigree, and turns forgetting into something a third party can check.
5 min readEngine mechanisms · Recall boundary tested with the network disabled · Erasure certificates for isolated recordsRead article ↗
When an AI system makes a consequential call, someone will ask why. Today the answer is assembled after the fact from logs, traces and the memory of whoever was on shift. In 8DB the answer is a query: the source records, their confidence, the governance state at the time, and a proof a third party can check without trusting anyone who ran the system. A two-week reconstruction becomes a session.
Your model did not hallucinate. Your database did. Retrieval taught the industry that an answer can only be as honest as its store, and most stores cannot say how sure they are, where a fact came from, or the difference between unknown and absent. 8DB carries confidence and provenance on every fact, distinguishes eight kinds of nothing, and refuses to answer past its evidence, which changes what the model on top of it is allowed to say.
5 min readEngine mechanisms · Public-corpus run · Benchmarks not yet formal claimsRead article ↗
Sanctions screening asks one question of one name at a time. Load the entire public sanctions corpus into a store that treats it as a structure, and different questions become askable: which unlisted entity holds the network together, which subsidiaries of a designated parent were never listed, which nominee held fourteen assets for sixteen months and moved them all in one day. Here is what that looked like on 4.28 million public records, and why every finding arrived with a proof.
6 min readPublic-data campaign · Findings with provenance artifacts · Performance figures withheldRead article ↗