What Prompt Engineering Is — Problem solving 101 with AI

Using Charly Mangers favorite tools for solving problem and making decisions

Nikita Ponomarenko
ILLUMINATION’S MIRROR
4 min readJun 27, 2024

--

Photo by Yogendra Singh on Unsplash

The ability to solve problems is one of the most crucial skills you can possess, especially in the workplace.

Problems can come in all shapes and forms, and it often feels like no matter what you do, there’s always a new issue arising. It’s like fighting small fires — once you put one out, another pops up.

Throughout my career, I’ve tried various techniques and attended numerous courses at my workplace. But for some reason, these strategies only seem to work during the workshop.

Once you step back into the real world, you quickly realize that what you were taught was just theory, not practical tools.

I remember reading Dale Carnegie’s book on managing stress, and despite my initial enthusiasm, it didn’t provide the solutions I needed. I moved from book to book, each time becoming more frustrated and feeling buried under an ever-growing pile of unread guides.

There’s another way to do it

Basically, I’m here to show you how to solve problems using mental models (its a fancy word for saying how you think) and prompt engineering (and that’s a fancy way to get tailored made from AI and not generic ones)

With these tools, you can solve problems quickly and efficiently.

What is Prompt Engineering?

Prompt engineering is like giving very specific instructions to an AI assistant (system like ChatGPT or DALL-E) to get exactly what you want it to do.

Much like you would guide someone to reach a specific destination or achieve a particular outcome.

To get the parrot (the AI) to say what you want, you give it special instructions called “prompts.” These prompts are like little hints or guidelines that tell the parrot how to respond.

While many are familiar with AI technologies like ChatGPT, they often don’t know how to use them effectively.

Most people resort to naive prompting.

Let me explain what I mean by that. Suppose I am feeling stressed out at work and want to find a solution using ChatGPT. Typical prompts might look like:

  • “Managing stress at work”
  • “How to manage stress at work”
  • “Techniques to manage stress”

The issue with these prompts is that they usually yield very generic responses.

Here is a screenshot from Perplexity.ai illustrating this:

Screenshot by the author

I assume you’re already familiar with most of these points. It’s clear and straightforward.

This is why often, people tell me, “These tools are useless!”

My response is always the same: “It’s not about the tools, it’s about how you use them.”

This might seem obvious and simple, but it’s the truth.

Stick with me, and I’ll demonstrate how to use these tools to achieve the results you want.

Let me show you how to take Charlie Munger’s favorite thinking tool, mental models, and enhance them with AI.

But first…

Mental Models

A mental model is like a mental picture or map that helps you understand how things work.

Just as a real map shows you roads, cities, and landmarks, a mental model simplifies ideas and concepts so you can make sense of them.

For example, imagine you want to understand how a car operates. You might create a mental model that includes an engine burning fuel to generate power, which then turns the wheels to move the car.

While a real car has many complex details, this simple model captures the key components and their relationships.

Mental models help us break down complicated things into manageable chunks. They provide a framework for reasoning, decision-making, and problem-solving by simplifying the world around us.

You can then apply these principles to real-life challenges. One such principle is “The First Principle.”

Let’s use it to tackle the problem of “managing stress at work.” By applying a mental model, this is the result I got:

Screenshot by the author

What you’re seeing here is based on basic assumptions. It’s not the same response we got last time.

Every assumption is essentially a limiting belief, and only by challenging it can we find new ways to solve problems.

What are assumptions?

Assumptions are things we accept as true without questioning or examining them deeply.

You can’t solve a problem without those assumptions, they’re like the building blocks.

We can find new and creative solutions by challenging these assumptions

I’m going to use the WH questions to do that.
Here’s what I discovered:

Screenshot by the author

By simply answering those questions, you’ll start to notice:

  • Your beliefs begin to shift.
  • You have more power to change things than you might realize.
  • You can discover creative solutions.

Conclusions

AI isn’t a new phenomenon, and ChatGPT is just a small part of this technology. For $20, anyone can access an AI tool that can respond to prompts like “I’m experiencing stress, help me with that.”

To receive a truly customized response, you need to master prompt engineering. It’s not about a single perfect prompt; it’s about experimenting with various prompts to see what best addresses your specific issue.

Nikita

--

--