The Zero-Attention Economy, Second Edition

01What the first edition missed

In July I published a free book arguing that a machine becomes valuable at the moment it earns the right to be ignored. Almost all of it still seems right to me. Attention is the remainder of what we failed to predict. Absorption is prediction success. Trust is amortized attention, retired one uneventful week at a time. The hours a good system returns are real, and they vanish unless someone decides in advance where they go.

What the book never did was look closely at the hours that are left. It followed the returned hours to their fates and never asked what the remaining hours are made of, once the machines have taken everything they can predict. That turns out to matter as much as the dividend. So there is a second edition.

The machines take the easy work. What reaches you is only the hard part, and the easy work was where you rested.

02A new Part Four: The Exception Queue

Three new chapters follow the person who receives what absorption leaves behind.

1. Chapter 19

   Fresh Horses

   Whitehead wrote that civilization advances by the operations we no longer think about, and in the next sentence that thought is a cavalry charge: limited, in need of fresh horses, and best saved for decisive moments. Absorption removes the light stretches of a working day and leaves the charges, so the day gets shorter and heavier at once. Rest stops being a luxury and becomes maintenance.

2. Chapter 20

   Two Clocks

   Every decision a system forwards runs on the situation’s clock and the person’s. Holding a hard decision until someone is fresh is a kindness only if it will keep that long. A good system sorts each one, the way battlefield surgeons learned to: absorb it, report it, hold it, or deliver it now and small. It never asks about what time has already decided.

3. Chapter 21

   The Charge Is Yours

   Surprise belongs to a person, so a hard decision costs a particular person and not a job. The model that sorts someone’s day is the most valuable thing an agent holds, and it has to belong to the person it describes. This is the book’s full case for personal AI, and for letting people bring their own.

03What else changed

• Rest is booked as maintenance. The first edition said an hour of rest changes nothing about next quarter. It does: it keeps the faculty that makes good decisions from wearing out.
• The interrupt contract has two clocks. An interruption should arrive while the situation can still be acted on and when the person is fit to act.
• Total attention includes the walk. The practical shape is now four things: absorb the predictable, fund the reserve, pace the charges, spend the dividend where you chose.
• Fifteen rules instead of twelve, a new section of open questions, an expanded glossary, and a handful of first-edition errors corrected.

04One correction worth naming

The first edition said that asking a question is authorship of the denominator, meaning that a good question sets your own Expectation on purpose. The framework the book rests on forbids that. Nobody sets their own Expectation; it is inherited, never authored. What a question actually does is narrower, and still the whole point. It states an expectation you already hold, precisely enough to be broken, and goes to stand where the Actual that could break it will arrive. The chapter has been rewritten to say exactly that. The argument about the question premium survives intact, and is more honest for the change.

05Get it

The second edition is free, with nothing held back for a paid version. It is set for US Letter and prints cleanly on a home printer. Download the PDF, mark it up, and hand it to someone else. The first edition stays available. The thirty-minute narration on the book page was recorded for the first edition, and the argument it tells has not changed.