Placement and Judgment Over Raw Execution

status: active · created 2026-06-26 · 6 sources · #synthesis #cross-domain #ai-and-work #data-engineering #principle #craft

The wiki now spans two domains that look unrelated — AI and work (how agents change knowledge work) and engineering craft (how to build data products). Yet they keep landing on the same three structural moves. The reason they rhyme: both are responses to execution getting cheap. When doing the work — compute, AI output, raw labor — is abundant, the scarce, value-determining inputs are where you place the work and where you spend human judgment. Execution is the commodity; placement and judgment are the craft.

This is the placement-side companion to What's Scarce When Execution Is Cheap, which argued value migrates to the scarce, high-information human contribution. That synthesis named judgment as the scarce input. This one adds the second scarce input — placement — and shows that both express themselves through three repeating moves, each visible in both domains.

Move 1 — Do it once, at the right place

When redoing work is the waste, the craft is finding the single correct point and doing it there. Filed as its own hub: Do It Once, at the Right Place.

Data craft: derive once at the contract; compute where the data has gravity; persist a human decision so it's paid once.
AI & work: Personal vs Team Agents is the same move — maintain one shared "unit" agent that many people inherit, rather than each person re-building and re-mending a personal replica. Do the upkeep once, at the team layer.

Move 2 — Reserve human judgment for the exception

Automate the collapsible bulk; spend the scarce human on the cases that actually need judgment, and capture that judgment so it compounds.

Data craft: Human-in-the-Loop by Exception — automate the structured bulk, route only true exceptions to a person, and make each confirmation teach the system.
AI & work: The Human Sandwich — the human frames and judges; the agent collapses the middle. Both modes "require a human to work well," and The Bottleneck Shift explains why: when execution is cheap, deciding what to do is the bottleneck. Same split, different vocabulary.

Move 3 — Put the thing where the mass already is

Don't relocate the heavy object; bring the light thing to where the mass — of data, of attention, of existing work — already sits.

Data craft: Data Gravity moves compute to the data because the data is too heavy to move. Build It as an Extension puts a feature where the user's mental model already is and reuses the nearest proven mechanism.
Org & AI: Build With, Not Against makes the durable layer the other team's extension rather than a rival build — the same "fit into the existing mass, don't erect a competing one" logic at the org scale, and the lesson behind Personal vs Team Agents.

So what

The three moves are corollaries of one shift: as raw execution commoditizes, advantage moves to placement and judgment. That reframes a lot of decisions — platform choice (Choose the Platform by the Hardest Recurring Part), where to spend humans, where to compute, where to put a feature — as the same question: given that doing it is cheap, where does doing it correctly actually live? It also suggests a test for new heuristics entering the wiki: does it tell you where to do something, or who/what should judge? If so, it probably belongs to one of these three.

Open questions

Is "placement" really separable from "judgment," or is choosing the right place itself the highest-leverage act of judgment? (If so, this synthesis collapses back into What's Scarce When Execution Is Cheap — judgment all the way down.)

Do the three moves have counterexamples in either domain — places where re-doing, centralizing judgment, or relocating the mass is actually right?

Connections

What's Scarce When Execution Is Cheap — the judgment-side synthesis this extends; read them as a pair.
Do It Once, at the Right Place — Move 1, broken out as a reusable hub.
The Bottleneck Shift — the engine under the whole thing: cheap execution shifts scarcity to deciding what to do (and where).
Working Heuristics — Data, Matching, Product, Org, Tooling — the source whose ingest surfaced the cross-domain rhymes.

Sources

Synthesizes Working Heuristics — Data, Matching, Product, Org, Tooling with the AI-and-work corpus (The Human Sandwich, Personal vs Team Agents, The Bottleneck Shift, Agents as Employees vs Human-Agent Collaboration) and the prior synthesis What's Scarce When Execution Is Cheap.