3 GPT Prompts That Make Me a 10x Engineer
Unlocking AI superpowers: prompts that changed how I work
Hey friend,
Last year, I was playing around with AI tools. Mostly just asking them basic stuff like “Summarize this article” or “Generate a tech documentation about …” Cool, but nothing life-changing.
I realized I wasn’t using AI wrong, but I also wasn’t using it right. It was like having a Ferrari and only driving it in a parking lot. So I started digging into how to talk to AI better. Like really talk to it.
That’s when I found something that changed my workflow, my side projects, even how I explain things to friends.
Here are three powerful prompting techniques that turned AI into my most trusted coding partner, thought organizer, and brainstorming buddy.
Let’s dive in.
1. Chain-of-Thought Prompting
Sometimes, when you ask AI a question, it just spits out the answer. No explanation. No thinking out loud. Just… boom. Final answer.
But here’s a better way to ask:
Design a basic system architecture for a ride-hailing app. Explain your decisions step by step. Starting with the core services and why you chose them.
Now instead of a one-size-fits-all diagram, AI walks you through its logic (step by step). What services it prioritized, how they talk to each other, and trade-offs it considered.
Why this is powerful:
You learn how the AI makes decisions
You catch flawed assumptions early
You can tweak the parts that don’t feel right
I use this when I’m exploring designs, reviewing code suggestions, or preparing engineering strategy docs. It’s like having an architect who always explains their thinking—and never pushes back on feedback.
2. Role-Based Prompting
One of the best ways to get more useful answers from AI? Give it a role and a specific audience.
For example, if you ask:
Explain the difference between SQL and NoSQL databases.
You’ll probably get a technical breakdown—useful, but maybe not what you actually need.
Now try this:
You’re an experienced database expert explaining to a product manager with limited technical background. Help them understand the key differences between SQL and NoSQL, and when to choose one over the other—using simple, real-world examples.
Suddenly, the answer shifts. Less jargon, more clarity. Focused on what the listener needs to know, not just what the AI knows.
I use this approach often—especially when aligning with cross-functional teams. Whether I’m drafting strategy docs, onboarding new hires, or prepping for stakeholder briefings, this trick helps AI tailor the tone and depth so I don’t have to do all the translation work later.
Think of it as turning AI from a know-it-all into a great communicator.
3. Modular Prompting
Big tasks are where things fall apart. Unless you break them down.
One time, I asked AI:
Write a complete guide on mentoring new grad engineer.
The answer? Too long, all over the place, and overwhelming.
So I broke it down:
Step 1: Explain the importance of code reviews.
Step 2: Discuss how to give constructive feedback.
Step 3: Share tips for setting learning goals.
Step 4: Suggest ways to encourage independent problem-solving.
Now the AI was thinking like a team lead—not just dumping ideas. The output became a clean, usable draft.
This method is gold for:
Creating technical design docs
Writing sprint plans or retrospectives
Mapping out hiring roadmaps
Or even prepping talking points for a tech leadership sync
The key? Guide AI through your thinking. One logical step at a time. You’ll reduce noise, avoid rework, and get results that fit how you operate.
So, Why Does This Matter?
Because this is where things are going.
Tech companies, and honestly, smart solo builders too are looking for people who don’t just use AI, but know how to guide it. Prompting is the new literacy.
These three techniques are how I’ve gone from AI “user” to AI collaborator. And once you start using them, it’s hard to go back.
Try This
Before you leave, challenge yourself:
Next time you use ChatGPT (or any AI tool), don’t just ask a question.
Use one of these techniques. See what happens. You might be surprised at how much better your results are.
If this post made something click for you—share it with someone. Let’s help more folks stop driving Ferraris in parking lots 😉
Until next time,
Adlet
🎉 We're approaching 2,000 newsletter subscribers in just a couple of days! Thank you for your support!
You’re the real MVPs!
Loved this post? 💙 Just hit the like button below. It's a small favor from you, but it helps me grow a lot.
Stay curious!
Recently I did: Based on commit hash xyz, tell me the changes made by it with respect to abc feature.
Probably even better:
You are a product owner whose app is for <insert customer group>.
In release notes, write up to two bullet points based on commit hash xyz, telling changes made by it with respect to abc feature.