Appendix
Registry objects
The book rests on a formal registry of theorems, lemmas and propositions. The registry justifies the book; it does not organize it. Listed here are only the objects a published chapter actually invokes, each with the wording that chapter uses. The full registry, and the verbatim statements, are reproduced in the print appendix.
-
LEM-002
measured and recorded activity becomes observable enough to support analysis, attribution, monitoring, or management
-
LEM-006
a buyer who incurs several kinds of cost pays more in total than the access price
-
LEM-016
moving from isolated use into repeated operations can increase the number of usage events and demand points
-
LEM-020
monitoring, limits, attribution, or other management mechanisms can be applied to visible usage or exposure
-
LEM-021
requires such a boundary whenever return on AI is evaluated
-
PROP-032
compute use can be measured
-
PROP-039
measured compute use can inform methods for assigning cost
-
PROP-041
metering a resource or activity records one or more measurements of it
-
PROP-042
measured activity can be distinguished by quantity
-
PROP-043
metering produces records of what was metered
-
PROP-053
monitoring can be used to observe economic exposure
-
PROP-085
management decisions can draw on visible usage
-
THM-002
When a buyer’s cost of adopting AI includes more than the access price, and evaluating the return requires a boundary that includes both cost and value, treating the access price as the whole cost risks understating it.
-
THM-004
Once a deployment reaches scale, economic control requires a governing apparatus. Diligence from the people running the deployment is no longer enough.
-
THM-009
deployed AI is a resource-consuming operating activity, not merely software access
Continuity
Promises between chapters
A chapter that says a later chapter will do something has made a promise. Each is logged when the promising chapter locks and marked paid when the promised chapter locks.
-
Chapter 1 to Chapter 2 paid
This is the first chapter; problem sets begin reaching back to earlier chapters in Chapter 2.
-
Chapter 1 to Chapter 3 paid
Each of these subjects has its own purpose and literature, and Chapter 3 explains where each ends and AI Operations Management begins.
-
Chapter 1 to Chapter 4 open
Chapter 4 examines these instruments in detail.
-
Chapter 1 to Chapter 6 open
It returns in Chapter 6 as the book’s anchor case on realized value.
-
Chapter 1 to Chapter 14 open
Chapter 14 distinguishes the subject of this book from these neighboring disciplines.
-
Chapter 2 to Chapter 1 paid
4 Chapter 1 argued that a flat price relocates the meter to the provider rather than abolishing it.
-
Chapter 2 to Chapter 1 paid
Chapter 1 defined its basic unit: the consumption event, a single use of an AI system that consumes metered computing resources.
-
Chapter 2 to Chapter 1 paid
Chapter 1 described the problem as a category error, and Chapter 3 names the discipline that resolves it.
-
Chapter 2 to Chapter 1 paid
Chapter 1 established that the provider always holds a meter, because the provider must bill and must manage its own capacity.
-
Chapter 2 to Chapter 1 paid
Interleaving: question 4 and problem P2 require Chapter 1’s results.
-
Chapter 2 to Chapter 1 paid
Public reporting does not establish whether Uber made the category error described in Chapter 1 or understood the cost model but failed to carry that understanding into its forecast.
-
Chapter 2 to Chapter 1 paid
Uber’s experience makes the problem from Chapter 1 concrete.
-
Chapter 2 to Chapter 3 paid
As Chapter 3 explains, the theorem formalizes the economic consequence of the third flow’s asymmetry; it does not establish the taxonomy itself.
-
Chapter 2 to Chapter 3 paid
Chapter 1 described the problem as a category error, and Chapter 3 names the discipline that resolves it.
-
Chapter 2 to Chapter 8 open
Chapter 8 builds the record flow, and Chapter 10 turns that record into a budget.
-
Chapter 2 to Chapter 8 open
It returns in Chapter 8, which builds the record flow that organizations are structurally most likely to skip.
-
Chapter 2 to Chapter 10 open
Chapter 8 builds the record flow, and Chapter 10 turns that record into a budget.
-
Chapter 3 to Chapter 1 paid
3.5 Where the discipline ends Chapter 1 set aside three related subjects: how models work, prompt engineering, and finding uses for AI.
-
Chapter 3 to Chapter 1 paid
Chapter 1 described the problem as a category error.
-
Chapter 3 to Chapter 1 paid
Chapter 1 stated the conditions under which deployed AI is a resource the organization consumes.
-
Chapter 3 to Chapter 1 paid
First, apply Chapter 1’s consumption-event inventory.
-
Chapter 3 to Chapter 1 paid
One might treat Chapter 1’s Cursor licenses as a fixed software expense and the other as metered consumption.
-
Chapter 3 to Chapter 2 paid
4 Chapter 2 diagnosed a deployment through its three flows.
-
Chapter 3 to Chapter 2 paid
Chapter 2 ended with that result, which a reader who accepted its argument already believes.
-
Chapter 3 to Chapter 2 paid
Chapter 2 stated the conditions under which a scaled deployment cannot be controlled without a governing apparatus.
-
Chapter 3 to Chapter 2 paid
Interleaving: question 4 requires Chapter 2’s three flows, and the cumulative case below requires both earlier chapters.
-
Chapter 3 to Chapter 2 paid
Second, apply Chapter 2’s three-flow mapping.
-
Chapter 3 to Chapter 2 paid
THM-004 also does not establish Chapter 2’s three flows; that exclusion comes from this book’s framework, rather than the registry.
-
Chapter 3 to Chapter 2 paid
The second function is planning and budgeting, whose absence proved costly in Chapter 2’s opening case.
-
Chapter 3 to Chapter 2 paid
The third function, metering and attribution, builds what Chapter 2 called the record flow.
-
Chapter 3 to Chapter 2 paid
This pairing explains Chapter 2’s statement that THM-004 formalizes the economic consequence of the third flow’s asymmetry.
-
Chapter 3 to Chapter 6 open
The case returns in Chapter 6.
-
Chapter 3 to Chapter 7 open
“Are you paying for capability you do not need, or starving work that needs more?” This is the sourcing question, answered in Chapter 7.
-
Chapter 3 to Chapter 10 open
“What do you expect the AI flow to consume next period, and how will you know when it deviates?” This is the planning and budgeting question, answered in Chapter 10.
-
Chapter 3 to Chapter 11 open
“When capacity is constrained, who decides what runs first, and by what rule?” This is the allocation question, answered in Chapter 11.
-
Chapter 3 to Chapter 12 open
“Where, exactly, does this AI flow pay for itself, and who is on the hook for knowing?” This is the value boundary question, answered in Chapter 12.
-
Chapter 3 to Chapter 14 open
Chapter 14 explains how FinOps and that practice divide the work.
-
Chapter 3 to Chapter 14 open
Chapter 14 explains how the two divide the work.