Part 1 — The New Engineering Problem  ::  Chapter 4

1.4 Two Problems the Method Must Solve

The preceding substrate properties leave two engineering problems that MAGE addresses directly. A finite reasoner needs representations that keep relevant state tractable as systems and tasks grow. Probabilistic autonomous work needs authority outside the producing run so that important obligations do not depend on memory, confidence, or repeated human inspection.

1.4.1 The Representation Problem

A broad reasoner can operate on abstractions, but its active reasoning state is finite and large systems contain far more detail than any one engineering question needs. How should a large, evolving system be represented so that a finite reasoner can work coherently over the properties that matter without continually reconstructing them from implementation?

Part II develops MAGE's answer: Modeling.

1.4.2 The Authority Problem

Autonomous work can be locally constrained and checked without a comprehensive model. But probabilistic implementation cannot be its own final authority, and richer system-level obligations require the environment to obtain their meaning from somewhere. How should engineering obligations be given authority over autonomous work so that satisfying them does not depend on the producing agent remembering, agreeing, or self-certifying—and so that deterministically checkable questions do not require a human to reconstruct the answer each time?

Part III develops MAGE's answer: Alignment.

1.4.3 How Representation and Authority Relate

The two problems are related but neither is reducible to the other. Representation shapes the scale and semantics of what humans, agents, and tools can reason over; authority determines what the resulting work must satisfy before it is accepted. Local controls can already be authoritative without rich models. Better abstractions, however, enlarge the set of consequential properties that can be stated, reasoned about, and checked without recurring human reconstruction. Part II develops the abstractions; Part III shows how obligations acquire authority; Part IV combines the two into the method. Figure 1.4-1 traces that derivation. Commodity intelligence reduces the implementation work required to realize many engineering decisions; it does not remove the decisions themselves.

From substrate to method Two derivations run left to right. In the first, a finite working state raises the representation problem, which the method of modeling answers. In the second, probabilistic productive work over observable, interposable seams raises the authority problem, which the method of alignment answers. These are the two core activities Part I derives; Part IV composes them into the MAGE method. From substrate to method SUBSTRATE TRAIT PROBLEM METHOD finite working state Representation problem Modeling represent agents’ reasoning probabilistic productive work observable, interposable seams Authority problem Alignment give obligations authority Part I derives the two core activities; Part IV composes them into the MAGE method.
Figure 1.4-1. From substrate to method. Finite reasoning state makes representation consequential; probabilistic autonomous work makes external authority consequential. Parts II and III develop MAGE's corresponding answers: Modeling and Alignment.
© James C. Davis, 2026–present