# Prompt Iteration Workflow

## The Idea

A simple, distraction-free workflow where:

1. **You write prompts in a text file** (`prompt.txt`)
2. **The AI monitors the file** for changes
3. **When you save**, the AI detects the change and processes your prompt
4. **The AI responds** in a separate file (`respons.txt`)

### How It Works

```
┌─────────────────┐
│  prompt.txt     │ ← You edit this in any text editor
│  (Your Input)   │
└────────┬────────┘
         │ (save triggers detection)
         ▼
┌─────────────────┐
│   AI Monitor    │ ← CLI-AI watches for file changes
│   (File Watcher)│
└────────┬────────┘
         │ (processes prompt)
         ▼
┌─────────────────┐
│  respons.txt    │ ← AI writes response here
│  (AI Output)    │
└─────────────────┘
```

### Benefits

- **Distraction-free**: Use any editor (vim, nano, LibreOffice, etc.)
- **Async workflow**: Save when ready, AI picks it up automatically
- **Persistent**: Your prompts and responses are saved as files
- **Simple**: No complex interfaces or tool switching
- **Iterative**: Refine your prompt, save again, get new response

---

## Prompt Iteration vs Prompt Engineering

### Prompt Engineering

**Definition:** The practice of designing, testing, and optimizing prompts to get the best output from an AI model.

**Characteristics:**
- One-shot or few-shot optimization
- Focus on finding the "perfect" prompt formulation
- Often involves A/B testing different phrasings
- Typically done in AI chat interfaces or playgrounds
- Emphasis on technique and methodology

**Example:** "Write a function that..." vs "Create a robust, production-ready function with error handling that..."

### Prompt Iteration

**Definition:** A workflow where you continuously refine prompts in an external file, with the AI responding to each iteration.

**Characteristics:**
- Continuous refinement loop
- Focus on the **process** of improvement over time
- Uses file-based workflow for separation of concerns
- Allows deep thinking and editing in your preferred editor
- Emphasis on **iteration** and **evolution** of ideas

**Example:** 
1. Save: "Write a simple function"
2. AI responds
3. Edit & save: "Add error handling and type hints"
4. AI responds
5. Edit & save: "Now make it production-ready with logging"
6. And so on...

### Key Differences

| Aspect | Prompt Engineering | Prompt Iteration |
|--------|-------------------|------------------|
| **Focus** | Finding the right prompt | Evolving the prompt over time |
| **Workflow** | Test → Compare → Select | Write → Save → Refine → Repeat |
| **Tool** | AI chat interface | External text editor + file monitoring |
| **Mindset** | Optimization | Iteration & refinement |
| **Output** | Best single response | Series of evolving responses |
| **Distraction** | Context switching in AI interface | Focused writing in your editor |

### Summary

**Prompt Engineering** is about *crafting* the optimal prompt.  
**Prompt Iteration** is about *evolving* your ideas through continuous refinement in a distraction-free environment.

The iteration workflow complements engineering - you're still engineering better prompts, but through a file-based, iterative process rather than direct chat manipulation.