If Press Pocket Guides
People, Products, and Ideas Are Not the Same Thing
Eidolon, Ergon, and Noema: Why AI Needs to Understand What Exists, What We Build, and What We Think
Describe what exists, what was designed, and what you believe separately. Connect them deliberately.
Give AI a folder of documents and it will summarize them fluently. It may also say that your customer owns the software, that a vendor's promise is a measured result, and that a hypothesis has been proven. A good summary can sit on a bad model of the world.
Forthcoming. The book is in review; its companion and tools are open now.
Your AI can read everything and still mistake a company for its product.
This Pocket Guide teaches one habit. Keep what exists, what was designed, and what you believe separate, then connect them with relationships you can defend and evidence you can inspect. One invented clinic carries the case from a confusing pile of documents to a decision, and six prompts were tested on it, with the mistakes reported.
It is short on purpose, it says its terms are a working model and not a standard, and it never claims a check ran that did not.
What you will be able to do
- Tell a person, a product, and an idea apart
- Split a phrase that names several things
- Keep a vendor’s claim apart from a measured result
- Connect records with relationships that have dates and sources
- Spot the category errors an AI summary hides
- Decide when the extra modeling is worth it