Prompt Engineering Tutorials — Write Better AI Prompts

1 Beginner 15 min read

OpenAI's Official Prompt Engineering Guide

This is the guide that made prompting "click" for me. It's written by OpenAI (the ChatGPT team) and breaks down six core strategies that work. Before reading this, I was just typing random sentences and hoping for good results. After reading it, I started getting consistently better answers.

The guide is short and practical — no fluff, just techniques you can use immediately.

OpenAI's 6 strategies for getting better results from AI — applicable to any AI tool

The 6 strategies (my summary)

prompt-strategies.md

              1. Be specific

                 Bad:  "Write me some code"

                 Good: "Write a Python function that takes a list

                       of numbers and returns the average"

              2. Provide reference text

                 Give AI something to work from — paste docs,

                 examples, or context so it doesn't have to guess

              3. Break complex tasks into subtasks

                 Instead of "build me a website," ask step by step:

                 "First, create the HTML structure for a nav bar"

              4. Give the model time to "think"

                 Add "Think step by step" to complex questions

              5. Use external tools

                 Let AI use code execution, web search, etc.

              6. Test changes systematically

                 Try different prompts and compare results

The one thing that changed everything

Being specific. That's it. Instead of "help me with my code," I started saying exactly what I wanted: "Fix the bug on line 23 where the array index is out of bounds." Night and day difference in the quality of responses.

Read OpenAI's Full Guide Source: OpenAI Official Docs

2 Beginner 12 min read

Anthropic's Guide to Prompting Claude

If you use Claude (which I do a lot), this is essential reading. Anthropic published their own prompting guide that covers Claude-specific techniques. The biggest thing I learned? Claude responds really well to structured prompts with XML tags. It sounds technical but it's actually simple.

Anthropic prompt engineering documentation

Anthropic's prompt engineering docs — organized from basic to advanced techniques

Claude-specific tips that work

claude-prompt-example

              # Use tags to structure your prompt:

              <context>

              I'm building a personal website with HTML and

              Tailwind CSS. It's a single-page site.

              </context>

              <task>

              Add a responsive navigation bar with a mobile

              hamburger menu. Use dark colors.

              </task>

              <requirements>

              - Must work without JavaScript for basic nav

              - Use Tailwind CSS classes only

              - Mobile-first responsive

              </requirements>

              # Claude responds much better to this structure

              # than a long, unstructured paragraph

Other techniques that work great with Claude: asking it to "think step by step" before answering (especially for complex problems), giving it examples of what you want (few-shot prompting), and being explicit about the output format you need.

Read Anthropic's Prompt Guide Source: Anthropic Official Docs

3 Beginner Free course (9 chapters)

Hands-On Prompt Engineering Tutorial

If you learn by doing (like me), this is the best resource I've found. Anthropic released an interactive, 9-chapter tutorial that you can work through at your own pace. Each chapter has explanations, example prompts, exercises, and a playground where you can experiment.

The best part: there's a Google Sheets version if you don't want to deal with code. Just open the spreadsheet and start practicing prompts right there.

Each chapter has exercises where you practice writing prompts and see how changes affect results

What you'll work through

9

Chapters

30+

Exercises

Free

No cost

Don't skip the exercises

Seriously. The exercises are where the learning actually happens. Reading about prompting is one thing — actually writing prompts and seeing how small changes affect the output is where it clicks.

Start the Interactive Tutorial Source: Anthropic GitHub

4 Intermediate Reference guide

The Prompt Engineering Guide (DAIR.AI)

Once you've got the basics down, this community-driven guide is like a reference manual for prompting across all AI tools. It covers advanced techniques like chain-of-thought prompting, ReAct patterns, and model-specific tips for Claude, ChatGPT, Gemini, and others.

I don't read this front to back — I use it like a reference when I'm trying to figure out a specific prompting technique. It also has a "Prompt Hub" with reusable templates you can copy and modify.

promptingguide.ai — comprehensive reference for prompting techniques across all major AI models

Key topics covered

Zero-shot vs Few-shot

When to just ask vs when to give examples first.

Chain-of-thought

Getting AI to show its reasoning before answering.

ReAct patterns

Combining reasoning with actions (for tool-using AI).

Model-specific tips

What works best for Claude vs ChatGPT vs Gemini.

Explore the Prompting Guide Source: DAIR.AI (Open Source)

Prompting & Workflows