Hands-on: Your First ChatGPT Conversation

Welcome to Your First Real Practice!

You've learned about LLMs, training, tokens, and prompt engineering. Now it's time to put it all together with hands-on practice!

Exercise 1: The Content Creator

Scenario

You're starting a blog about healthy eating and need help creating content.

❌ First Attempt (Beginner)

Write a blog post about healthy eating

What happens:

• Generic, unfocused content
• No specific audience
• Random structure
• Could be for anyone, about anything

✅ Improved Version (Apply What You Learned)

You are a nutritionist writing for busy professionals who want
to eat healthier but have limited time.

Write a 300-word blog post titled "5-Minute Healthy Breakfasts
for Busy Mornings" that includes:

1. A brief intro explaining why breakfast matters (2 sentences)
2. Three quick breakfast ideas with ingredients
3. One time-saving tip
4. A motivating conclusion

Tone: Encouraging and practical, not preachy
Avoid: Complicated recipes, expensive ingredients, nutritionist jargon

Why it's better:

• ✅ Clear role (nutritionist)
• ✅ Specific audience (busy professionals)
• ✅ Exact structure (numbered list)
• ✅ Tone specified (encouraging)
• ✅ Exclusions listed (what NOT to do)

Exercise 2: The Code Helper

Scenario

You're learning Python and stuck on a concept.

❌ Common Mistake

explain loops

Problems:

• Which programming language?
• What's your current skill level?
• What specifically is confusing?
• How do you learn best?

✅ Better Approach

I'm learning Python and confused about for loops vs while loops.

I understand basic concepts like variables and if statements,
but I don't know when to use each type of loop.

Can you:
1. Explain the difference in simple terms
2. Give me a real-world analogy for each
3. Show two short code examples (one for each type)
4. Tell me when to use which

Explain like I'm a beginner, but don't oversimplify - I want
to truly understand the concept.

Why it works:

• Specifies language and topic
• States current knowledge level
• Clear, numbered requests
• Examples requested
• Learning style specified

Let's simulate this with code:

Exercise 3: The Problem Solver

Scenario

Your laptop is running slow and you need troubleshooting help.

❌ Too Vague

my computer is slow

⚠️ Better But Still Not Great

my computer is slow, how do I fix it?

✅ Excellent Problem Description

My laptop is running significantly slower than usual.

Details:
- Windows 11, Dell XPS 15, 16GB RAM
- Started 3 days ago after Windows update
- Slowest when opening Chrome or switching between apps
- Hard drive shows 80% full
- No obvious malware (ran Windows Defender)

I'm comfortable with basic troubleshooting (restarting, updating
drivers) but not advanced technical fixes.

Can you suggest 3-5 most likely causes and step-by-step fixes,
ordered from easiest to hardest?

For each fix, let me know:
- How long it takes
- Risk level (can I break anything?)
- Whether it requires technical expertise

Why this is powerful:

• Specific symptoms
• Context (when it started, what changed)
• What you've already tried
• Your comfort level with solutions
• Exactly what format you want

Exercise 4: The Creative Writer

Scenario

You need help writing a birthday message for your boss.

❌ Lazy Attempt

write a birthday message for my boss

✅ Context-Rich Request

Help me write a professional birthday message for my boss's
card that the whole team is signing.

Context:
- She's been with the company 10 years
- Known for being supportive and mentoring team members
- Recently helped our team through a challenging project
- The team culture is professional but warm (not stiff/formal)

Tone: Appreciative and warm, but professional
Length: 2-3 sentences (it's going in a card)
Avoid: Generic platitudes, overly casual language, anything
that could be seen as brown-nosing

I want it to feel genuine and specific to her impact on the team.

Better output because:

• Relationship context provided
• Specific things to mention
• Tone carefully calibrated
• Length constraint realistic
• Clear "avoid" list prevents awkward AI suggestions

Practical Patterns That Always Work

Pattern 1: The "I need X for Y" Structure

I need [TYPE OF CONTENT] for [PURPOSE/AUDIENCE].

Context: [RELEVANT DETAILS]

Format: [HOW YOU WANT IT]

Constraints: [LENGTH/TONE/STYLE]

Pattern 2: The "Help Me Understand" Structure

I'm trying to understand [CONCEPT].

What I know: [YOUR CURRENT KNOWLEDGE]

What confuses me: [SPECIFIC CONFUSION]

Can you explain [WHAT YOU WANT] using [HOW TO EXPLAIN]?

Pattern 3: The "Review My Work" Structure

I've [WHAT YOU DID] for [PURPOSE].

Here it is:
[YOUR WORK]

Please review it and provide:
1. [WHAT TO CHECK]
2. [WHAT TO IMPROVE]
3. [WHAT WORKS WELL]

Target audience: [WHO IT'S FOR]

Common Mistakes to Avoid

Your Practice Exercises

Now it's your turn! Try these exercises with a real LLM:

🎯 Exercise A: The Explainer

Task: Get the AI to explain "how WiFi works" to three different audiences:

1. A 10-year-old
2. A high school student
3. A college engineering student

Challenge: Same topic, three different prompts. Notice how you adjust language and depth!

🎯 Exercise B: The Translator

Task: You need to write an email declining a meeting invitation politely.

• The meeting conflicts with another commitment
• You want to suggest an alternative time
• Keep it professional but warm

Practice: Write a detailed prompt that gets you a perfect email!

🎯 Exercise C: The Debugger

Task: You have this code that isn't working:

def calculate_average(numbers):
    total = 0
    for num in numbers:
        total + num
    return total / len(numbers)

print(calculate_average([1, 2, 3, 4, 5]))

Challenge: Write a prompt that:

1. Explains what's wrong
2. Shows the fix
3. Explains WHY it wasn't working
4. Gives a tip to avoid this in the future

Real vs Simulated AI Responses

Since we can't call a real LLM API in this lesson, let's simulate how different prompt qualities affect responses:

Test Your Learning

Key Takeaways

🎯 Context is King: Always provide background

• Who's it for?
• What's the purpose?
• What's your knowledge level?

🎯 Structure Your Requests: Be specific

• Use numbered lists
• Specify format and length
• Give examples when possible

🎯 Iterate and Improve: First try rarely perfect

• Refine based on output
• Add constraints if needed
• Learn what works for you

🎯 Save Your Best Prompts: Build a library

• Save prompts that work well
• Modify for similar future tasks
• Share great prompts with colleagues

Congratulations! 🎉

You've completed Module 1: AI Fundamentals!

You now understand:

• ✅ What LLMs are and how they work
• ✅ Training vs Inference
• ✅ Tokens and Context Windows
• ✅ Prompt Engineering basics
• ✅ How to have effective conversations with AI

What's Next?

In Module 2: Prompt Mastery, you'll learn:

• Advanced prompting techniques
• Few-shot learning
• Chain of Thought prompting
• Prompt patterns that solve specific problems
• Building your own prompt library