# ChatGPT Prompt: Mechanical Sovereignty — The 740 Project

## Intent

Show the 740 as a doctrine artifact: stock exterior concealing mechanical-first engineering, positioned as civil resilience rather than tactical capability.

## Documentation Note

This project includes:
- Full generation strategy with 4 layered imaging approaches (GENERATION-STRATEGY.md)
- Complete development chat history showing doctrine refinement (CHAT-LOG.md)
- Visual assets demonstrating capability-poster framing

The prompting strategy was developed iteratively to address AI hallucination in vehicle cutaways by splitting complex engineering into single-system layers.

## Core Doctrine

**Repairability > Capability**  
**Mechanical Autonomy > Electronic Dependency**  
**Reduced Failure Points > Advanced Systems**  
**Independence Is Engineered, Not Declared**

**Footer:** What you can repair, you can trust.

## Capability Layers (Doctrine-First)

**Layer 1: Repairability**  
Can I fix it here? (hand tools, field repair)

**Layer 2: Mechanical Autonomy**  
Can I keep moving alone? (no electronic dependencies for core movement)

**Layer 3: Sustained Mobility**  
Can I keep operating without support? (fuel autonomy, mechanical redundancy)

**Layer 4: Civil Resilience**  
Can I adapt when assumptions fail? (degraded infrastructure, supply chain breaks)

## Primary Prompt: Capability Poster (RECOMMENDED)

See GENERATION-STRATEGY.md for full 4-layer imaging approach.

Quick version for capability-focused image:

```
Professional automotive poster of a 1980s Volvo 740 Combi (wagon) in stock condition.

Vehicle should look ordinary and accessible.

Overlay topographic contour rings radiating outward, labeled (inside to outside):
  Repairability
  Mechanical Autonomy
  Sustained Mobility
  Civil Resilience

At center: Seal reading "MECHANICAL FIRST"

At bottom banner: "What you can repair, you can trust."

Style: Engineering poster aesthetic (not digital, not tactical).
References: Topographic maps, old aviation navigation charts, survey drawings.
Color: Muted vehicle, pencil-sketch rings in blue and orange.
```

## Why Layered Generation Works

Full cutaway requests hallucinate:
- Wrong engine orientation
- Impossible gearbox placement
- Suspension geometry errors
- Components floating in space

Layered approach (separate: chassis, powertrain, suspension, capability poster):
- AI excels when each image contains ONE engineering system
- Packaging constraints disappear
- Recoverable detail increases
- Exploded-view becomes intentional design communication

## Development Context

**Problem Identified:** Vehicle cutaways require CAD-level accuracy that image models cannot provide through single prompts.

**Solution:** Split into 4 focused layers, each addressing one system.

**Doctrine Refinement:** Chat history shows shift from "advanced resilience platform" (digital tactical) to "mechanical-first preparedness" (field-repairable, human-scale).

## Run Log

- Doctrine developed: 2026-06-22
- Refined through conversational design feedback
- Layering strategy validated for reducing hallucination
- Footer landing: "What you can repair, you can trust."
- Status: Ready for candidate publication and visual iteration