ChatGPT is the most powerful tool to make money online. Bu…

What you'll learn in this thread:

• How to create sophisticated and tailored AI prompts
• How to understand what audiences want and need
• How to optimize prompts for maximum output
• How to use AI to create new products
• How to monetize AI prompts

1. Defining the problem

To create a good prompt you need to understand the problem.

• What is the expected output of ChatGPT?
• Can you do the task manually?
• What input is required?

The first step should be doing the task manually.

2. Validate the problem

Pretend you are ChatGPT, ask someone you know to prompt you.

Write down:
• What was your thought process?
• Did you need more information?
• Is this task valuable for the person you are helping?

Define the:
• Input
• Steps
• Output
• Benchmark

3. Mess around with ChatGPT

• Try asking it to just do the task in one go.
• Try asking how it would break down the task.
• Try formulating the task differently.

Evaluate?

• Did it understand the steps?
• Was the quality good enough?
• Was it consistent?

4. Debug

The words you use matter.

Try:

• Use the same words you would use.
• Use the right terminology.
• Don't have typos.

See how using words impacts the output.

5. Try adding more information.

Just like you, ChatGPT needs to do research.

Try giving background information on the task you want it to do:
• Related books or people.
• Explanations and definitions.
• Example of how it should do the steps.

Give it a chance to reflect.

6. Make it think

ChatGPT jumps to conclusions. You need to make it think.

Instead of answering a question directly, make it independently asks itself follow-up questions whose answers lead to the final answer.

7. Select the plattform

The prompt needs to be usable for your customers.

You have three options:
1. Sell the prompt.
2. Make an app around the prompt.
3. Use the prompt in an existing app.

How?
1. Sell on gumroad
2. Use no-code + Openai api
3. Your app and + Openai api

8. Iterate the prompt

Use customer feedback to optimize the prompt.

• Did they understand the output?
• Did it meet their expectation?
• Was the step-by-step process intuitive?

In almost all cases, you will face these problems.

9. Aligning customer expectations.

In generative AI the customers have expectations.

You need to surpass expectations.

How?
• Show examples to expect.
• Explain output.
• Output content that is difficult to make.

10. Pivot

At Tribescaler customers didn't like our generated twitter thread content, except for one sentence, the hook.

In user testing:
"Can I save that hook? I want to use it".

The customer shouldn't be able to imagine life without your prompt.

11. Monetize

Using OpenAI costs money.

Important principle:
Charge a percentage of the value created, not to cover costs.

A good prompt should create enormous value.

11. What is a good prompt?

Prompts do different things:
- Answer: Ask about anything.
- Creating: Article generation.
- Upskill you: Write code.

The highest value is all at once.

Example:
- How to survive in the arctic?
- Get ideas.
- Make mindmap.
https://t.co/SqDMUaFrC2

What does chatGPT think about a subject?

You can ask it to make a mindmap.

Here is the prompt you can generate a VISUAL mindmap on any topic in seconds:
— Alexander Leirv\xe5g (@Claybight) January 18, 2023

12. Advanced topics

You need more advanced techniques to survive as an AI startup in 2023.

Here are tools to get you started:

• Prompt engineering: https://t.co/hHodEcEs9i
• GPT Memory: https://t.co/o5H7Y2D2Z8
• Integrate Google: https://t.co/RO3YghYhGH

And that's it!

If you enjoyed this and you're into AI and prompt engineering:

1. Follow me @Claybight for more badass threads like this

2. Retweet the 1st tweet of this thread to help inspire others

TL;DR (Part I)

13-step process I use to create prompts to make $13,000 every month.

• Understand the problem and validate it
• Mess around with ChatGPT
• Debug your AI prompts
• Integrate more information
• Make it think
• Select the platform
• Iterate your prompts

TL;DR (Part II)

• Align customer expectations
• Pivot to a better solution
• Monetize your skills
• What is a good prompt?
• GPT-3 stack technologies