Lesson 2 of 42

Why prompts matter

Vague vs Specific: Same Model, Different Result

4 min read

Two prompts. Same model (Claude Sonnet 4.5). Same task — write an email. The only thing that changes is how clearly the prompt tells the model what we want. Watch what happens.

The same task, written two ways — toggle to see the difference

Vague
Write me an email.

Prompt 1 — vague

Write me an email.

Output:

# Email Template

**To:** [Recipient's email]

**From:** [Your email]

**Subject:** [Your subject line]

---

Dear [Recipient's name],

I hope this email finds you well.

[Main body of your message goes here - explain your purpose, provide necessary details, and include any requests or action items.]

[Add a second paragraph if needed for additional information or context.]

Please let me know if you have any questions or need any further information.

Thank you for your time and consideration.

Best regards,

[Your name]

[Your title/position]

[Your contact information]

---

**Note:** I've created a basic template for you. To write a more specific email, please let me know:
- Who you're writing to
- What the email is about
- The tone you want (formal, casual, professional, etc.)
- Any specific details you want to include

Captured from Claude Sonnet 4.5 (claude-sonnet-4-5) on 2026-04-27. Re-runs may differ slightly.

A blank template. The model genuinely tried — but you didn't tell it anything, so it gave you the shape of an email and asked you to fill in the rest.

Prompt 2 — specific

Write a 90-word email to my landlord asking to extend my lease by 6 months.
Tone: warm and professional. I have lived here 2 years, paid rent on time,
and have a stable job. Sign as "Sara". Output only the email body, no subject line.

Output:

Dear [Landlord's Name],

I hope this message finds you well. I've thoroughly enjoyed living here for the past two years and would love to extend my lease for an additional six months.

As you know, I've consistently paid rent on time and maintained the property with care. My job remains stable, ensuring continued reliable tenancy.

This extension would be ideal for my current situation, and I'm happy to discuss any terms or paperwork needed.

Thank you for considering my request. I appreciate your understanding and look forward to hearing from you.

Warm regards,
Sara

Captured from Claude Sonnet 4.5 (claude-sonnet-4-5) on 2026-04-27. Re-runs may differ slightly.

Useful, signed, on-tone, the right length, ready to copy-paste.

What changed

Nothing about the model changed. What changed was the information density of the prompt. The vague prompt forced the model to guess everything: who, what, why, length, tone, format. The specific prompt removed every guess.

FieldVagueSpecific
Recipientunknownlandlord
Goalunknownextend lease 6 months
Toneunknownwarm + professional
Lengthunknown90 words
SignatureunknownSara
Formatunknownbody only

That's the whole game. Most "ChatGPT gave me garbage" stories are really "I gave it nothing, so it gave me an average answer."

Vague vs specific prompt — same model, different result

Generic

Vague prompt

RecipientUnknown
GoalUnknown
ToneAverage
LengthUnspecified
Cons
  • Returns blank template with placeholders
  • Asks you for the missing details
  • Wastes a turn — you have to retry
Useful

Specific prompt

RecipientLandlord
GoalExtend lease 6 months
ToneWarm + professional
Length90 words
Pros
  • Returns ready-to-send body
  • Signed and on-tone in one shot
  • Copy-paste, you're done

Next: a related trap — treating the model like a search engine instead of a writer. :::

Quiz

Module 1: Why Prompts Matter

Take Quiz
Was this lesson helpful?

Sign in to rate

FREE WEEKLY NEWSLETTER

Stay on the Nerd Track

One email per week — courses, deep dives, tools, and AI experiments.

No spam. Unsubscribe anytime.