Skip to main content
1.1

The Vague vs. Specific

Estimated time: 20 minEstimated cost: ~$0.02Tool: Claude Haiku
After this drill, you can:

After this drill, you can see — with your own eyes — the concrete quality difference between a vague prompt and a specific one.

Why this matters

Every beginner sends vague prompts. The output looks fine — until you compare it to what a specific prompt produces. This drill runs both versions of the same task and puts them side by side. The comparison IS the lesson. You cannot un-see the difference once you've seen it.

How to do it

  1. 1

    Choose a task from your real work or life — something you genuinely need help with

    Examples: summarize a long report, write a proposal, explain a concept to someone, plan an event. Pick something real — it makes the comparison meaningful.

  2. 2

    Send the vague version of the prompt (use template below)

    Copy the vague prompt template. Replace [TASK] with your task. Send it. Copy the response.

  3. 3

    Send the specific version of the same task (use template below)

    Now use the specific prompt template for the same task. The template forces you to add role, context, format, and constraints.

  4. 4

    Compare side by side and write one sentence about what changed

    The sentence should name what specifically was better — not "it was more detailed" but "it included X that the first version missed."

The prompt

PROMPT — Vague Version (replace [TASK] with your actual task)Model: Claude HaikuEst. cost: ~$0.01
Help me with [TASK].
PROMPT — Specific Version (fill in all five fields)Model: Claude HaikuEst. cost: ~$0.01
You are a [ROLE — e.g. senior consultant, experienced teacher, business analyst].

I need help with: [TASK]

Context: [2-3 sentences about your situation, who it's for, what you already know]

Output format: [e.g. bullet list, 3 short paragraphs, a table with columns X and Y]

Constraint: [e.g. keep it under 200 words, avoid jargon, write for a non-technical audience]

Success criteria

  • ✓You sent two versions of the same task
  • ✓The specific version used at least 3 of the 5 prompt components (role, context, format, constraint, task)
  • ✓You wrote one sentence naming what specifically changed
  • ✓You saved this specific prompt to your Prompt Portfolio

Common mistakes

Filling in the specific template with generic content

→ The role should be a real expert (not just "expert"). The context should have actual facts. The constraint should limit something specific (not just "be helpful").

Judging the comparison immediately without reading both responses fully

→ Read both responses from top to bottom before writing your comparison sentence. The key differences are often in the middle or end of the response, not just the opening.

Forgetting to save the prompt

→ The Prompt Portfolio is a module-end deliverable. Save each tested prompt as you go — do not rely on memory.