AI in Development: Your Knowledge Determines How Well You Can Guide It

Developers Are Moving Towards AI. Will It Be Effective for Them?

During my journey, I have seen many interns and developers who have just started learning development use trial and error with prompts to build websites. After a few attempts, they do manage to build a website. But the question is: is that website production-ready?

The developer may not even know that they need to tell AI to build for production so that it works towards that requirement. Let us assume the developer does ask AI to build a production-ready website. Now another question comes up: what should be considered when building for production, and what exactly is AI focusing on?

AI Only Works as Far as Your Knowledge Takes It

AI understands that you want a production-ready website. But is it also working on SEO? Does it know your audience? Does it know that you need CSP issues to be resolved? Will it give you a solution with VAPT issues resolved on its own?

In most cases, that does not happen. And if you are a beginner, you may not know about many of these things yourself.

Getting AI to Develop Something Is Not the Challenge

Getting AI to do development work is not the challenge. The challenge is deciding what context you will give it. For that, what matters is how much context you have and how you guide AI while providing that context.

If you leave AI to work on its own, it will work according to what it has learnt. But if you give it clear instructions, it will work according to your requirements.

How Can You Become Better at Giving Instructions?

Before that, let us understand who can give instructions and how.

If you do not have knowledge of something, think about what instructions you can give and how you will give them. You will need to do research even for that. This means you need knowledge of the work before you can give instructions for it.

If You Do Not Know Anything, Can You Still Get AI to Do the Work?

In general, it is wrong to think that you can get AI to produce good work when you do not know anything about the work yourself. AI will either do everything its own way or work only as far as you know how to guide it.

It is important to pay attention to this: AI works its own way, not necessarily the way you want it to.

So, what do you need to do to make AI work according to your requirements? The same point applies again: you need to give it instructions and point out its mistakes. Only then will it work according to your requirements and learn from what you tell it.

What Happens When AI Learns Everything?

AI can never reach a point where it has learnt everything. The world changes every day. Change is the law of nature.

You will need to keep telling AI that something has changed and that it needs to change the way it works or add to its knowledge.

Let us understand this with a simple example. Suppose a new programming language is introduced tomorrow. AI will need to be trained in it.

Let us also assume that you have trained AI in such a way that it can read and learn a language on its own. But what if something completely new comes along, something that has never existed before? You would not have worked on AI for that particular thing, so you would still need to put in the effort to teach it.

Who Can Teach AI?

To teach AI, you first need to understand how it works from the inside.

Think about the people who have built AI or are working to improve it. In the future, you will also need to develop knowledge at their level if you want to properly change or update it from within.