Not a content marketing factory. Posts when there is something worth saying, written by the operators doing the work, grounded in what we actually run on.
Most AI vendors will not tell you what their agents can and cannot do. We wrote it down: authorization, behavior, and change control. Here is what is in it and why we built it instead of buying one off the shelf.
A walkthrough of the actual stack. Not a demo reel, a working day. What we automate, what we deliberately do not, and why.
Anthropic changed their billing terms with less than 24 hours notice. Here is what that should tell you about building on any single AI provider.
How a simple structured registry eliminated context confusion across 30+ projects and made my AI assistant actually useful.
How a simple wrapper script eliminated 90% of AI command failures. A practical case study in building AI-friendly tooling.
Anchor pieces in draft, each grounded in real work. They open when we have the material to back them.
Drafting
The majority of large companies have a Chief AI Officer. A fraction of their employees actually use AI. The gap is the business. What the same numbers look like at sub-100-person scale, and what to do about it.
The open source work and project activity at Hexaxia Labs is the canonical place to follow what we are building. Reach out if you want to talk about any of it.