Why Coding Isn’t Going Away
Coding? What is it? I’ve been doing this since 1999 and I still don’t know what the hell I’m doing. You can say that coding is the process of writing instructions in a programming language, which are then translated by a compiler or interpreter into machine code (0s and 1s) that a computer can execute. In a way, it is like typing English words into Google Translate and then letting it translate them into Spanish, for example. Coders, programmers, software engineers or whatever you want to call us, are people who know one or more programming languages that can be translated into something the computer can understand. At least that’s how modern programming languages work and how the code is used.
But what is it really? Is it just typing words? Well, no. If you type the following into Google Translate and then translate it into Spanish, it will still lack enough details for someone to understand it and act on it, right?
Go look at all parking spaces and find how many are free
This is an instruction you can give to someone, and depending on the context, they might or might not know what to do. For example, if there are many parking lots, you’ll need to specify which one. You might also need to say whether you’re looking for empty spaces of any size or a specific size.
When you are coding, you need to give very specific instructions to the computer to execute. You cannot be vague, or else your program will result in unexpected behavior—or worse. Coding is more than just knowing a programming language and being able to type it. That’s one part. The most important part is knowing how to give specific and clear instructions, and also attempting to account for scenarios to avoid mistakes and errors. This is what we call the logic. Your code needs to be written in a way that’s safe and efficient. This takes experience and a lot of practice.
What’s coding then? It is the ability to understand a problem or task well enough to create specific instructions for a system via code and data so it can execute these instructions successfully, resulting in the expected output.
It is common to find errors or unexpected behavior in your code, no matter your experience. It happens. We call these “bugs.” There are many ways to avoid them. There are also processes to find them early, before you release your set of instructions (application code) to a system in production. You can have automated tests that review the inputs and outputs, people who test it manually, and tests that check the functionality and logic efficiency of a small part of the code, we call these unit tests, etc.
So what’s coding now? Now that we have LLMs capable of writing code, and some even able to predict what code is needed to accomplish a task, how does coding change, really? In my opinion, not much. Although some people claim that today’s models are capable of creating complex applications with little effort, the truth is very different. These LLMs still fall short when you try to create an application that needs to connect to other systems such as APIs or databases. Having it orchestrate complex tasks all on its own unfortunately doesn’t work. You need someone to review it and fix the misunderstandings and errors it makes.
What’s changing is that now we have another layer that can write the code for us. The code it writes uses the same programming languages we humans do, so it’s an abstraction, another layer between the human and the machine code. The coder is still needed, since these AI tools are only good at generating code when the instructions are clear. Even then, someone with experience needs to check that the logic written is correct and that the code not only works, but is also safe and efficient.
I think coding will change, and this is obviously a step forward in that direction. Once we have a system that is capable of giving instructions to a computer without writing in a programming language made for humans—once such a system can talk to a computer’s CPU directly in a language that both understand without compiling, transcribing, or translating, then we can say that coding as we know it today has changed. Right now, we’re in an important moment. More people, even without much coding experience, can create simple apps and web pages. And believe me, all the input and output is being recorded to help make these systems faster, more accurate, more efficient, and eventually self-sufficient. We are all helping them with all of our inputs and feedback.
What happens when we have a system that can translate anything someone says into machine code so a computer can execute the instructions reliably and efficiently? That’s what Artificial General Intelligence (AGI) would do, not only for coding, but across different fields.
What about when such a system doesn’t even need a person’s input, it just does the task? That’s Artificial Superintelligence (ASI), and we’re not there yet. Theoretically, an ASI won’t need human input for any task. It will act on its own, more intelligently than a human. I imagine it will still need people to maintain its hardware, energy supply, and network connections.
For now, and for many years ahead, I don’t see coding disappearing from human hands. The tools may change, and AI may handle more of the typing, but the core skill, understanding a problem, shaping it into precise instructions, and judging whether the solution is sound, remains human. These same skills are what make us effective at guiding and supervising AI output.
The old advice applies: trust but verify. Answers are easy; correct answers are hard. Keep learning so you can recognize when something’s wrong. We’re still in charge, at least for now.
As for me, I’m learning to use these tools and trying to make the best use of them.
This post and many of my previous ones, was written 100% by me, but I’ve been using a custom GPT to help with grammar and flow. I’m using it as my writing editor. If you want to try it out, I just made it public:
It’s called Keep My Voice, and it’s available on ChatGPT:
https://chatgpt.com/g/g-iBvzMbkni-keep-my-voice
Have a great weekend!
