AI for Lunch

You use these things every day. Do you know what happens after you press enter?

Not the marketing version. The actual mechanism - explained in plain English, with no mathematics, no jargon, and nothing you have to take on faith.

This book has no company names in it. No product names. No model names. No dates. Not one. You can check by flipping to any page. That is not a gimmick - it is the whole design. Books about this subject go stale in a year because they are really about this year's tools. This one is about the machine underneath, which is not changing nearly as fast as the logos on top of it.

The explanation runs on a single extended image: a restaurant kitchen you are never allowed to see. You send in an order. A plate comes back. What happened in between is the entire subject of this book - and every time the analogy bends, the book stops and tells you exactly where, because an explanation that hides its own seams is flattering you, not teaching you.

By the last page you will be able to say out loud, without bluffing:

• why a confident answer and a correct answer are produced by exactly the same process


• why these systems invent things, and why that is ordinary operation rather than a malfunction


• what training actually leaves behind, and why nobody can open it up and read the reasons


• why the same question gives different answers, and which dial causes it


• what a benchmark score does and does not tell you about your own work


• where the edge of its competence falls - and why you cannot find that edge by asking it

You will finish with a posture rather than a trick: read the kitchen, and trust each plate exactly as much as it has earned.

AI for Lunch is the foundations volume of the AI at the Table series. Each book stands alone; all three share one kitchen, so nothing you learn in any of them goes to waste.

No mathematics. No hype. No expiry date.