
The short answer
Use AI assistants to explain, give hints and review your work. Do not use them to write your exercises. AI is a patient tutor that never gets tired, but if it does your thinking, you learn to copy, not to code, and you cannot spot when it is wrong.

AI is like a calculator in a maths class. Wonderful once you understand the maths. A trap if you use it to avoid learning the maths.
Why "just ask AI to build it" fails beginners
AI assistants (such as ChatGPT, Claude or Gemini) predict likely answers from patterns in huge amounts of text. That makes them very good at common tasks, and sometimes confidently wrong:
- They may use an old version of a tool, with commands that no longer work.
- They may invent a function, setting or package that does not exist.
- They may write code that works but is insecure, for example storing passwords in plain text.
A developer who understands the basics notices these problems in seconds. A beginner who copied the code cannot, and ships the bug. That is why understanding comes first: you can only check what you understand.
The three good uses
| Use | What you ask | What you get |
|---|---|---|
| Explain | "Explain this concept / error / line of code simply" | Understanding |
| Hint | "Give me a hint, not the answer" | Progress, with your brain still working |
| Review | "Here is my code. What could be wrong or unsafe?" | Feedback like a senior developer's |
1. Explain
When a lesson uses a word you do not know, or an error message makes no sense, ask for an explanation at your level.
Ask your AI assistant
I am a complete beginner learning web development. Explain what [the concept, for example "an API"] is using an everyday example, then give me a tiny code example and explain each line. Do not assume I know any technical words; define them.
2. Hint
When you are stuck on an exercise, ask for the smallest possible push.
Ask your AI assistant
I am stuck on this exercise: [paste the task]. Here is my attempt: [paste your code]. Do NOT give me the solution. Give me one hint that points me in the right direction, then wait for my next attempt.
3. Review
After you finish something yourself, ask for a review, the way a senior developer would review a pull request.
Ask your AI assistant
Review this code I wrote as a beginner: [paste your code]. It is meant to [describe what it should do]. List problems in order of importance: bugs first, then security, then readability. Explain each simply and show the improved line, but keep my approach.
How to write prompts that work
Vague questions get vague answers. A good prompt tells the AI five things:
- Your level: "complete beginner", "I know basic PHP".
- The tool and version: "Laravel 13", "Python 3.14". Versions matter because commands change.
- What you did, exactly: the command you ran or the code you wrote.
- What happened, exactly: paste the full error message, not "it doesn't work".
- What you want: an explanation, a hint, a review, or a comparison.
| Weak prompt | Strong prompt |
|---|---|
| "My code doesn't work" | "I am learning Laravel 13. I ran php artisan migrate and got this error: [full error]. Explain what it means and how to check the cause" |
| "Teach me Git" | "I am a beginner. Explain what a Git branch is with an everyday example, then give me three commands to practise, explaining each" |
| "Write a login system" | "I want to understand how login works in Laravel. Explain the steps a sign-in goes through, then point me to what Laravel provides so I do not build it from scratch" |
Check every answer
Treat AI output like advice from a smart stranger: probably helpful, never automatically right.
- Run it. Does it actually work on your computer?
- Compare with the official documentation for the tool, at your version.
- Ask "why". If you cannot explain each line, you are not done.
- Watch for security red flags: passwords in plain text, secrets written into code, user input used without validation.

A quick test: close the AI chat and explain the solution out loud, as if to a friend. If you can, you learned it. If you cannot, ask it to explain again.
Never paste these into an AI chat
- Passwords, API keys, tokens or the contents of your
.envfile. - Customers' personal data: names, phone numbers, ID numbers.
- Your employer's or client's private code, unless they have allowed it.
Replace them with placeholders such as YOUR_API_KEY before asking.
A healthy weekly balance
When you are learning, aim for something like this:
- Most of your time: reading lessons, typing code yourself, getting stuck and unstuck.
- Some of your time: AI explanations and hints when you are truly stuck (after 15 to 20 minutes of trying).
- After you finish: AI reviews of your finished work.
As you get more experienced, you will use AI for more of the routine typing, because you can check it quickly. That is the point: experience is what makes AI safe to lean on.
Key takeaways
- AI is a tutor: ask it to explain, hint and review.
- Give it your level, the tool and version, what you did, what happened and what you want.
- Check every answer by running it and comparing with official documentation.
- Never paste secrets or personal data.
- If you cannot explain the code, you have not learned it yet.
Our lessons include a ready-to-copy "Ask your AI assistant" prompt after every hard part. Start with How the web works.
Check your understanding

Every lesson stays free. Tap the thumbs up so we know what works, or say thanks with a soda or a coffee: it pays for hosting and new lessons.



