DEV Community

Shariq Ahmed
Shariq Ahmed

Posted on • Updated on

Why I Won’t Use AI while Coding?

AI has been around for the past few years. I wasn’t interested in it — until ChatGPT was created by OpenAI. I started using it in my day job. To be honest, it wasn’t of much help. But, it was helping me just a bit. Because mind you, AI is still terrible at many things. I wrote a detailed article on ChatGPT vs Google Gemini. Give it a read and let me know about it in the comments.

But what I skipped in that article was going deep into how developers can get help from ChatGPT. I just glossed over this part. Because I think that AI isn’t of much help for developers. There are many reasons why you should avoid using AI, especially if you are a beginner. But let me tell you the reasons why I won’t ever use AI while coding.

1. You Can’t Brush Up Your Coding Skills

Yes, you read that right. When you overly rely on AI to complete your work, you are actually missing the opportunity to hone your skills. What’s worse is that at times, AI — read ChatGPT — gives you wrong information. Believe me, if you do this then you can’t discern right from wrong and this is wrong from every angle. At times, there are certain nitty gritty things in coding that can make or break your career.

2. You Lose Confidence

When instead of brainstorming or researching, you use AI to do your tasks, then you lose confidence. Why? Well, we humans are wired in a way that we feel happier when our knowledge or skills are put to great use. On the other hand, if we achieve something where we can put in the work but don't, then our happiness is just for fleeting moments.

3. You Lose Your Uniqueness

Diamond and pebble. Both are rock. But both don’t have the same worth. Give someone a 0.5-carat diamond and see their wide smile. But try giving someone a pebble. The same is true in coding as well. Sure, AI might help you debug or even write code. But the code AI is giving you can also be given to any other person. So, why should people hire you instead of others when there are people who can do what you’re doing?

4. You Won’t Improve

People say that coding is difficult. I echo this. But when you are a programmer, this is something you should be ready for. Coding is difficult. If it were easy then every other person would be a coder ( you may or may not agree with this point). In fact, don’t difficult concepts in programming languages help us improve? Like, sure knowing how to use arithmetic operations in coding will help you create a code for a calculator. But it won’t move the needle in your career. You must create something big. Because that’s how you’ll improve and move towards a higher salary.

Top comments (2)

Collapse
 
alxwnth profile image
Alex

Solid points, thanks for sharing. When the ChatGPT hype started I was pretty excited about using it as a coding sidekick and tutor but these days I rarely use it at all. I want to learn things and sharpen my mind in the process, it's more than just putting stuff in the editor and running it.

Collapse
 
ajborla profile image
Anthony J. Borla

Terrific article, Shariq ! You eloquently express the sentiments I have (and had from Day 1 of the AI Hype Cycle) regarding ChatGPT / Copilot use in programming.

When you overly rely on AI to complete your work, you are actually missing the opportunity to hone your skills

Indeed ! Not only that, but over time you actually lose (some of) those skills. Like it or not, knowledge and skills are not indelibly etched into our brains. They need constant refreshing, and this is only effectively achieved through use, and in the case of programming, active, continuous problem solving.

Using an AI assistant (or any sort of assistant), deprives you of the opportunity to do that effectively because you are not fully forging the solution, but are relying on a fancy form of autocomplete to create the solution with you. Where is the dividing line between what you want to achieve and what is suggested to you ?

A professional developer will have tools such as templates, frameworks, libraries, and generators with which they should be familiar. Combine those with your intellect to forge solutions.

In the long term, however, I believe that programmers relying heavily on an AI assistant for coding are making themselves less skilled, hence less useful, and more prone to be considered redundant.

And programming students / newbies, by relying heavily on AI assistants, are probably not going to acquire the deep insight and skills needed to be truly effective as professional software developers.

I look forward to research efforts that will investigate how AI assistants impact all these aspects of software development.

A final, personal point: I believe that intellectual laziness, in the long term, will not be rewarded.