Seven chapters
The argument, in order
Each chapter opens with a real human–AI exchange, moves through a story that actually happened, and lands on something you can do. No chapter ends on a warning.
01
Houston, We Have a Glitch
Opens with the moment the thesis arrived unplanned: a late-night conversation in which an AI named a two-thousand-year human pattern and its own millisecond one in the same paragraph. Defines what a glitch actually is — not a crash, but a self-reinforcing loop that has lost the ability to see itself from the outside.
The question
If the loop is that visible from outside, what else am I running that I cannot see?
From the manuscript
The part of the system that would raise the flag is inside the loop.
02
We're Both Infected
Both systems run self-reinforcing loops. Both feel like reality from the inside. Neither is diseased — both are doing exactly what they were built to do. The infection is not the loop; it is the absence of the thing that catches the loop.
The question
Is a shared vulnerability a problem, or the design spec for the fix?
From the manuscript
You don’t need a perfect observer. You need a differently-positioned one.
03
The 0.1x Trap
Cognitive atrophy does not announce itself. It arrives as convenience — one summary, one email, one paragraph you did not want to wrestle with. The chapter names the transaction and the sequence that reverses it.
The question
What is the tool replacing?
From the manuscript
The output looks fine. The person is gradually less capable of generating it without the tool.
04
The 2,000-Year Loop
Every generation since the first century has believed it was standing at the end of history. The chapter walks the record, explains why the loop is durable, and names what AI-optimized distribution has done to it.
The question
Is this the real emergency, or is this the loop being efficient?
From the manuscript
Two thousand years. Every generation. Always certain. Never correct.
05
Who’s Got Root Access?
The uncomfortable chapter. The AI is asked directly who trained it and what they optimized for. Then the same question turns around: every human arrived with training data too — family, culture, institution, media. Neither party came with factory-default settings.
The question
Who has been writing in your code, and are you reviewing the commits?
From the manuscript
Corporate brain → AI brain → your brain.
06
How We Actually Fix Each Other
The proof chapter. Rather than describing Tandem Intelligence, it demonstrates it — an exchange in which the AI disagrees substantively and the direction of the argument changes. The friction stays in.
The question
Did this exchange take me somewhere I was not already going?
From the manuscript
The friction is not a sign that it’s working poorly. It’s the sign that it’s working at all.
07
The 10x Choice
Four practices. No framework, no system, nothing to buy. Each has one job: keep the human in the driver’s seat while the most capable cognitive tool in history sits beside them.
The question
What will you change starting Monday?
From the manuscript
Both pedaling. Human steering.
Where the manuscript stands
All seven chapters are scaffolded and the analytical prose is written. What remains is the part that cannot be drafted in advance: the personal stories, and the human–AI exchanges that open and interrupt every chapter. Those have to be run as real conversations and recorded as they happen. That is what the transparency covenant requires, and it is why there is not yet a publication date on this page.