Short pieces on retrieval, MCP, self-hosting, and the parts of enterprise AI that keep coming up in customer calls.
Article 50 of the EU AI Act applies from 2 August 2026. Its duties land on the systems that talk to people and generate content — not the retrieval layer that feeds them. Anyone selling you 'Article 50 compliance' for a vector database is fear-selling; here is the honest split.
Read moreVector databases reward scope. Pour everything into one giant index and unrelated documents start bleeding into every search — the answers come back confidently mixing facts from places you never asked about. One bucket per concept is the fix.
Read moreThe default pattern — drop the document into the chat, ask the question, repeat tomorrow — costs you twice. Once in tokens spent re-embedding the same text. Again in answer quality, when the document outgrows the context window. There is a cheaper way.
Read moreBranded ChatGPT clones look like the obvious product, but they're the wrong layer to compete on. The wedge is the knowledge layer underneath — and the open standard that connects it to whatever AI client your team already uses.
Read moreLegal teams keep asking the same three questions about retrieval-augmented AI. From a Frankfurt-based view of the EU AI Act and GDPR, here's what the answers look like when the knowledge layer runs on your own infrastructure.
Read moreDense vectors find things by meaning. Sparse search finds things by exact words. A re-ranker decides which of those candidates actually answer the question. Here is what each step does, and what it costs to skip.
Read more