How AI Coding Has Made Me Both a Better and Worse Developer

I’ve been writing code and developing websites for 24 years now, AI coding is fairly new. While I had some formal schooling, jumping into the job market as a freelance developer forced me into a state of constant learning. Taking on every project that came my way meant I had to figure things out on the fly. When I got stuck, I turned to online forums like Stack Overflow and Experts Exchange to find solutions.

I haven’t tried too many AI tools. My main go-to is ChatGPT. I did give GitHub Copilot a shot in VS Code, but I wasn’t that impressed with it and ended up turning it off. It couldn’t handle anything beyond the basics and automation.

Table of Contents

• How AI Coding Has Made Me Better
• How AI Coding Has Made Me Worse
• Pros and Cons of AI Coding
• In Conclusion

How AI Coding Has Made Me Better

About a year ago, I decided to give ChatGPT a try for simple problem-solving. To my surprise, it quickly and easily troubleshoots coding errors, saving me from poring over hundreds (or thousands) of lines of code to find mistakes. It is so simple: copy your code, paste it in, describe the error (or don’t, it may find it anyway), and voilà, code is fixed.

I discovered it could:

• Generate simple JavaScript or PHP functions in seconds.
• Automate repetitive tasks.
• Write code in languages I don’t fully understand as long as I knew the concepts.
• Troubleshoot in seconds.

One way I use it frequently is for converting old jQuery scripts into clean Vanilla JavaScript. Over the years, I’ve built a massive library of tried-and-true scripts that I often reuse. However, relying on external libraries like jQuery can put a heavy load on a website.

Now, with AI tools, I rewrite these scripts in cleaner, more efficient JavaScript to improve performance. This also applies to old PHP code I wish to upgrade to modern standards.

Another way I use AI is after I finish developing a function. I feed my code into the AI and ask it to:

• Clean it up.
• Add proper comments.
• Ensure it follows the latest coding standards.

Over the past year, my code has become remarkably clean—textbook clean, even.

How AI Coding Has Made Me Worse

As a self-employed developer, I’m much more efficient and faster than I used to be. But I have to admit that I’ve gotten a bit lazy. I no longer keep all the code in my head because I know I can just ask AI to generate it.

In many situations where the code I need isn’t top of mind, but 5 minutes of reading documentation will get my answer, I don’t bother. I just copy/paste into AI and get my answer.

If the code ChatGPT generates is more complex than my understanding, I don’t always take the time to learn how it works—I just test it to make sure it does. This has made me a little rusty when it comes to coding tests for positions, where I don’t have AI to lean on.

Pros and Cons of AI Coding

In Conclusion

I believe AI coding is where the industry is heading, but it will take a skilled developer to know how to ask the right questions. Asking the right questions of AI has become a skill in itself.

Most recently, with the help of AI, I built a custom CMS with multi-user support and three environments—entirely on my own. The code follows the latest standards, and I completed the project in record time, considering its size and complexity.

I love how efficient AI has made me, but I know I need to keep taking coding tutorials and staying on top of what’s possible. AI coding has made me both better and worse, but overall, it’s a tool I wouldn’t want to work without as I can deliver a better product much faster to my clients.

On some days, I feel like I’m training AI to replace me, but then I see the kind of bugs it gets stuck on and realize I can probably last until retirement. You still need to know what questions to ask.

Do you think AI will eventually replace developers, or will it just make us better?