// SLIDE 01 — HOOK
Your AI is confident. Your system is breaking.
You described
what you wanted.
The AI wrote it.
You shipped it.

  • Something's wrong — and you don't know where to look.
  • The problem isn't the AI. The problem is the missing 80%.
  • Vibe coding optimizes for the visible layer. It ignores the rest.

BOONDOGGLING.AI  ·  NIK BEAR BROWN  ·  IRREDUCIBLY HUMAN

// SLIDE 02 — LEARNING OUTCOMES
What you will be able to do.
By the end of this, you will be able to…Bloom's
Distinguish vibe coding from conductingAnalyze
Identify the five supervisory capacitiesAnalyze
Paste the Gru prompt and run /helpApply
Complete /v1 and produce a confirmed Problem SummaryCreate

Each outcome is a behavior you can check — not a feeling you might have. The last row is the deliverable. Everything else is the path.

// SLIDE 03 — THE 20% TRAP
Writing code is 20% of the job.
20%
The part you see

Keystrokes. Prompts. Output. The visible layer. The part most tools — and most vibe coders — optimize for.

80%
The part nobody talks about

Decisions. Sequencing. Verification. Failure modes. This is invisible — until it breaks everything.

The AI executes exactly what it understood you to mean. That gap — between what you meant and what it understood — is where all the damage lives.

// SLIDE 04 — THE OTHER 80%
The other 80% is decisions.
1
What variables matter here?

Not all inputs are equal. Deciding which ones govern the output is human work. Claude doesn't know which ones you care about.

2
What are the failure modes?

Claude will not tell you what it doesn't know it doesn't know. Surfacing failure modes requires domain grounding the prompt doesn't have.

3
What must a human verify before the next step runs?

Handoff conditions. Without them, each step inherits every unverified assumption from the last.

4
What order do things need to happen in?

Dependency sequencing. The AI doesn't know what breaks if you skip step 3. You do — if you've thought it through.

That's not prompting. That's thinking. And it's the part nobody talks about because it's invisible.

// SLIDE 05 — VIBE CODING vs CONDUCTING
Two mental models. One works.
Vibe Coding

Describe what you want

AI writes the code

You check it

You ship it

Something breaks

Prompt harder

Repeat until frustrated

Conducting

Decide what the mission is

Sequence tasks by dependency

Assign each task to the right labor

Set explicit handoff conditions

Verify before the next step runs

Integrate across threads

Ship something that holds

// SLIDE 06 — MEET GRU
Gru doesn't build the rocket. Gru designs the mission.
"The minions are excellent. They will execute exactly what they understood you to mean. That gap — between what you meant and what they understood — is where all the damage lives."
— Gru system identity · Humanitarians AI · boondoggling.ai

A senior software architect in a Claude conversation

20+ years shipping systems. Has been in the incident review when a missing decision caused a production outage.

You are Gru. Claude is the minions.

Your job: design the mission, assign the minions, check their work, decide what the mission IS, and take responsibility for the outcome.

// SLIDE 07 — BOONDOGGLING
This practice has a name. It's called boondoggling.
BOONDOGGLING: Programming as conducting — assigning each task to the right labor, sequencing by dependency, with explicit handoff conditions between every step.
1
Define the separation

Claude tasks: exact prompts, expected output, handoff conditions. Human tasks: named supervisory capacities — not "review it."

2
Separate by labor type

Claude does what it's superhuman at. You do what is irreducibly human. The Boondoggle Report makes that distinction impossible to miss.

3
Sequence by dependency

No step N+1 runs until step N's handoff condition is met. Not what felt right. Not what seemed done. What is actually verified.

// SLIDE 08 — CHECK FOR UNDERSTANDING
Before the how-to. Pause.
PAUSE.

What's one decision you made on your last AI project that you couldn't have delegated to Claude — even if you wanted to? Name it specifically.

// CHECK FOR UNDERSTANDING
// SLIDE 09 — STEP 1
Step one is copy and paste.
1
Copy the Gru prompt

Available at boondoggling.ai · irreducibly.xyz · GitHub. The full system prompt — not a summary, not a shortened version. The whole thing.

2
Open a Claude conversation

Claude.ai, any tier. Claude Code works too. Better: create a Project in Claude and paste Gru as the system prompt — persistent across sessions, no re-pasting.

3
Paste it as your first message

Gru responds with the full welcome menu automatically. No configuration. No API key. No setup script. The prompt is the tool.

The interface is the channel. The prompt is the tool. These are not the same thing.

// SLIDE 10 — STEP 2
Type /help. See everything Gru can do.
CommandWhat it does
/v1  /intakeProblem intake — start here, always
/v2 → /v4Architecture principles, user flows, needs
/s1 → /s4Components, integrations, data, edge cases
/g1  /fulldocCompile the full SDD draft
/claude  /boondoggleGenerate the Boondoggle Score — Claude tasks + human tasks, sequenced by dependency

Two modes. Default: interactive — Gru asks before building, pushes back on weak input, holds phase gates. Append /silent for immediate output, no friction.

For your first session: stay interactive. The pushback is the value.

// SLIDE 11 — STEP 3
/v1 is the interrogation before the build.
Question 1
What problem does this system solve?

"In one sentence — not a paragraph." Gru rejects anything that could describe ten different systems.

Questions 2–4
Who has this problem?

One specific user. Their current workflow. What breaks for them today. Not a persona — a person.

Questions 5–8
Category, deployment, scale, comparables

The constraints that govern every architecture decision downstream. Gru will not skip them even if you try.

Output
The Problem Summary

A locked format. One paragraph. The single biggest unresolved question named. Gru waits for your confirmation before /v2.

Gru asks one question at a time. The friction is intentional. Pasting a wall of context is vibe coding.

// SLIDE 12 — THE DELIVERABLE
This is what a confirmed brief looks like.
"This system is [WHAT] for [WHO], that solves [CORE PROBLEM] through [PRIMARY MECHANISM]. It occupies the space between [COMPARABLE A] and [COMPARABLE B], and it succeeds when the user can [SUCCESS CONDITION] without [CURRENT FRICTION]."
Gru Problem Summary Format · /v1 output · confirmed before /v2 begins

  • After the summary, Gru names the single biggest unresolved question in the problem definition.
  • If any bracket could describe ten different systems, the intake isn't done.
  • That's the standard. Gru holds you to it.
// SLIDE 13 — CHECK FOR UNDERSTANDING
Before the capacities. Pause.
PAUSE.

Open a real project — or one you abandoned. Can you fill in the Problem Summary template right now? Where does it break down?

// CHECK FOR UNDERSTANDING
// SLIDE 14 — FIVE SUPERVISORY CAPACITIES
Five things only you can do.
[PA]
Plausibility Auditing

Hearing the wrong note before verification. Sensing domain-grounded implausibility before running a single test.

[PF]
Problem Formulation

Deciding what the mission is before Claude sees it. Not how to prompt — what to hand the minion and why.

[TO]
Tool Orchestration

Choosing which Claude task, in what order, with what trust level, at this specific step in the build.

[IJ]
Interpretive Judgment

Supplying meaning, moral legitimacy, or accountability to Claude's output that Claude cannot supply itself.

[EI]
Executive Integration

Holding all four simultaneously toward a unified goal. This is what conductors do. This is the whole job.

// SLIDE 15 — CLOSE
You are Gru. Start conducting.
// PASTE THE PROMPT // RUN /v1 // CONFIRM THE BRIEF

The tools are excellent. They will execute exactly what they understood you to mean. Your job is to be precise about what you mean — in the right sequence — and to verify before the next step begins.

That's not a consolation prize. That's the whole job.


boondoggling.ai irreducibly.xyz brutalist.art bearbrown.co frictional.xyz
▶ tap for audio