The AI Field Guide
Fifteen cards on how AI actually works.
The AI Field Guide is a four-week Beyond Reason series, now collected as a free downloadable PDF and a permanent home on the web. Twenty-two pages. Fifteen Tufte-coded cards. Written for the people who have to approve the strategies, sign the contracts, and answer for the outcomes.
Each card stands alone. The strongest reading is sequential — the synthesis on Card 15 depends on the layers the earlier cards make legible.
If a colleague has time for exactly one card, send Card 04 (Context Assembly). It is the card most likely to change a buying decision. Start there →
See the journey
June 9 · Manifesto
How AI Actually Works
The anchor essay. The bird metaphor, the literacy gap, the three preview cards, and the four-week arc.
Read on Substack →June 16 · Part 1
What Happens When You Hit Send
The runtime of a single AI request. Cards 01–05: end-to-end, caching, tools, context assembly, context vs. memory.
Read on Substack →June 23 · Part 2
Where the Answers Come From
Training, retrieval, infrastructure. Cards 06–10: training run, inside training, RAG, data centers, prompt footprint.
Read on Substack →June 30 · Part 3
Why You Can Trust Some AI Systems
The trust layer. Cards 11–15: agents, permissions, routing, evaluation, the enterprise AI control plane.
Read on Substack →July 10 · The Artifact
Download the PDF
Twenty-two pages. Fifteen cards, six wrapping pages, full receipts. 8.5 × 14 inches. Free. CC BY 4.0.
Download ↓
The cards borrow from Edward Tufte without apology. The aesthetic is restrained because the substance is dense. None of these cards is the front of a brochure. All of them belong on a workbench.
If you find them useful, share them. If you find them wrong, tell me. The guide will need updating in eighteen months. The plumbing changes. The questions do not.
— Eric