Blog
Analyses and positions on current developments.
2026-07-21
How Do You Know an Agent Is Done?
An agent working on its own needs a point where someone decides: accepted, or back it goes. The obvious candidate — a green test run — demonstrably does not hold. Why that is, and why the fix is a question of structure rather than diligence.
2026-06-23
Memory Without a Cloud Paywall
An AI agent without memory starts every session from scratch, because new content crowds out the old. The common assumption is that anchoring and recall need the large cloud model. That is only half true. We split agent memory into two parts and show how far each gets locally: data-sovereign, in your own language, and without a per-query cost.
2026-06-13
When a Model Disappears Overnight
On 12 June 2026, a US directive forced Anthropic to disable Fable 5 and Mythos 5 worldwide. Why this turns the question of sovereignty over model and infrastructure — especially in cybersecurity — from theory into practice.
2026-06-05
Four People, One RAG Platform — and What They Actually Need
A multi-tenant RAG platform does not sell on features. It sells on the problems of real roles: the IT lead with no DevOps team, the knowledge officer who has to answer for what the bot says, the DevOps engineer at a regulated insurer, the CISO who has to sign off. Four people, four constraints — and what the platform does for each of them.
2026-05-28
Who Holds the Key? Three Decisions That Shape Multi-Tenant RAG Architecture
Anyone running a RAG service for multiple tenants has to settle three architecture questions early: how tenants are separated, who can read plaintext, and how users authenticate. We explain the options, the half-truths, and the trade-offs you can't model away.
2026-05-18
LLM Self-Hosting in the Mittelstand 2026: The Hardware Question
Why 2026 is the year self-hosting becomes economical for the Mittelstand — workstation class, inference NPUs, production servers, and when the switch pays off.