AI can generate syntax at lightning speed, but only humans bring purpose, ethics, and context to code.
GUEST COLUMN| by Punit Jajodia
HAMDI UJEB
AI has entered the chat, and it’s coding like it has caffeine in its circuits. Think of it like your go-to partner—your new co-worker, your confidante, your reliable co-editor. It doesn’t sleep, doesn’t snack, and it just refactored your code in minutes.
Large Language Models (LLMs) like GPT-4 have come to demonstrate remarkable capabilities, from solving maths problems to writing lines of working code like they’ve been programming with Alan Turing since the 1900s. What seemed like an impossible feat, LLMs can now draft essays and debug entire projects in minutes with simple prompts.
‘…this is not the first time we’ve witnessed a technological revolution that provoked fear of redundancy.’
And, in our awe of the greatness of AI’s capabilities, we have created our own illusion that’s led an entire generation of aspiring developers to go through an existential crisis. Discord servers are full of questions like:
“Why should I learn to code if AI can just do it for me?”
To be fair, it’s a valid question. But again, this is not the first time we’ve witnessed a technological revolution that provoked fear of redundancy.
Why Programming Remains Essential in the Age of AI
In the 80s, the arrival of desktop PCs made graphic design accessible to most individuals and businesses, triggering panic among designers in the printing industry. But even then, machines didn’t replace people. It was people who understood those machines and figured out how best to make them work for humans.
Today’s reality paints a similar picture. AI is trained on documentation, forums, blog posts, bug reports, and open-source code available on the internet. As Apple’s research team described in The Illusion of Thinking paper, AI models produce “thought-like” sequences without possessing anything resembling human thought or understanding. They are simply taking data as input, using algorithms built on layers of mathematical models to generate outputs—without understanding the underlying logic.
‘They are simply taking data as input, using algorithms built on layers of mathematical models to generate outputs—without understanding the underlying logic.’
Code is the blueprint that shapes every neural network, every chatbot, every tool that feels like “magic.” Think of it as raw material that must be processed for the desired outcome. Even as natural language interfaces evolve and the tech appears to understand and speak our language, don’t be fooled: it’s still code, quietly powering everything behind the scenes.
If you want to truly speak AI’s language, you need to speak code. That’s not changing anytime soon.
Why AI Won’t (and Shouldn’t) Replace Human Coders Completely
AI can almost write any functional code. However, what it fails to understand is why that code is important.
From a recent study in the Journal of Information Technology Education, while programs like ChatGPT can generate code with under 2% extraneous content, 70–80% of the output still requires manual editing to be useful. In fact, ChatGPT’s output required complete rewriting in 30% of cases. Why? Because it can write syntax, but has no comprehension of the reason for that syntax.
Further, AI doesn’t comprehend business logic, user needs, and ethics. It just executes prompts without fully understanding their purpose. And that is what differentiates it from humans. We write with purpose, with intent, and with context. We understand ethical boundaries and make informed decisions. We can even be held accountable for our actions—something AI cannot replicate.
‘…AI doesn’t comprehend business logic, user needs, and ethics. It just executes prompts without fully understanding their purpose. And that is what differentiates it from humans.’
So yes, AI can write functional code when commanded in English, but because it cannot reason, human supervision is integral. The role of a developer isn’t becoming obsolete; it’s evolving—from manually writing code to directing AI and refining outputs. Developers must ensure the software being generated is ethical, aligned with human values, and free of critical flaws.
Without supervision, AI systems can be counterproductive, spreading false information or even leaking sensitive data like passwords. As stated in Ethical Challenges in AI-Driven Software Engineering, “human oversight is essential” to correct biases, ensure responsible use, and safeguard the community impact of software systems.
In short: AI cannot replace coders. It needs to work with them.
Learning to Code Is Embracing a New Skill Set and a New Mindset
Thanks to the evolution of modern programming languages, coding is more accessible than ever—even to those outside traditional computer science or IT backgrounds. With AI co-pilots and smart coding assistants, you don’t need to memorize every function or command.
But coding is more than writing instructions for a machine. It reshapes how you approach problems, decisions, and complexity. Coding is fundamentally about learning to think logically: breaking big problems into smaller ones, spotting patterns, and designing scalable solutions. These are life skills—useful whether you’re building a mobile app, running a business, or tackling everyday challenges.
Coding is preparing for a future where logic, creativity, and clarity of thought matter more than ever. It’s a form of literacy for the modern world, where structuring ideas is just as significant as the ideas themselves.
‘Coding is preparing for a future where logic, creativity, and clarity of thought matter more than ever.’
The Future Is a Skill-athon and Coding Is the New Excel
In the 90s, anyone who knew Excel was considered a superhuman. They were indispensable, and Excel skills could catapult careers to new levels.
Today, the equivalent of Excel is code—not necessarily as a full-time developer, but in the ability to generate, automate, and interact with technology in unique ways.
- A marketer who writes Python to clean up campaign data works at 10x efficiency.
- A product manager who prototypes without engineers saves their team huge amounts of time.
- An analyst who automates dashboards doesn’t just report on strategy—they shape it.
Coding isn’t a niche advantage anymore. It’s a differentiating skill for anyone aiming to be a top future professional.
So let’s stop asking if you should learn to code, and start asking how. In an AI-generates-everything world, relevance will come from how well you understand the code behind every new advancement in technology.
Because in this AI revolution, those who code won’t be replaced—they’ll be the ones leading it.
—
Punit Jajodia is Co-founder and CEO of Programiz, which he has grown from a humble blog into a global learning platform reaching millions. A tech entrepreneur and educator with expertise in web app development and technical writing, he has also scaled other ventures by shaping strong company culture. His contributions to technology and education earned him a place on Nepal’s prestigious 40 Under 40 list of influential youth. Connect with Punit on LinkedIn.
0 Comments