Photo by Christina Morillo
Picture this: a world where coding is as effortless as conversing in English, where programming becomes akin to a casual chat with computers. According to Jensen Huang, CEO of NVIDIA, this isn’t a mere fantasy, it will soon be a reality for us all — thanks to AI.
In his recent speech at the World Government Summit at Dubai, Huang suggested that we stop emphasising coding and computer science education.
“Over the course of the last 10, 15 years, almost everybody who sits on a stage like this would tell you it is vital that your children learn computer science. Everybody should learn how to program and in fact it’s almost exactly the opposite,” he quipped.
As passionate educators in the world of coding and computer science, we find Jensen Huang’s claim to be a bitter pill to swallow. Jensen Huang is one of our heroes, but we simply don’t agree with his stance. Instead, we strongly believe that in the age of AI-generated code, coding and computer science knowledge will become even more vital in ensuring our youths’ success. Here’s why:
1. We will still need to think like a Programmer.
Just like how a person holding a supercomputer doesn’t automatically become a mathematician, a person who can generate code with AI through natural language doesn’t automatically become a programmer either.
In the age of AI-generated code, knowing how to prompt AI effectively will become more important than ever. This will involve an ability to abstract away details that are unnecessary for AI to know and systematically provide instructions to AI so that it can write out the code in a logical manner.
Photo by Christina Morillo
In other words, individuals will need to know how to think like a programmer — tapping on computational thinking skills such as algorithmic thinking, pattern recognition, decomposition and abstraction. The job of a programmer then becomes more of a conversation between the human and an AI, where the AI will excel at churning out boilerplate code and snippets of code, but the human programmer will still need to steer the development of the entire codebase, especially more so as the codebase gets bigger and more complex.
Without these, whatever is asked of AI will always end up falling short, resulting in half-baked code that cannot be deployed large-scale. Hence, for youths to effortlessly tap on AI-generated code, they need to be properly educated about how computers and computer programs work.
2. Coding Knowledge will remain Critical.
At present, AI is slightly misleading. It gives you the illusion that it is “thinking”, but if you look under the hood, it’s merely a prediction engine. It has been fed with so much information that it has become splendid at recognising patterns in data and spitting out useful predictions. However, this doesn’t change the fact that AI is incapable of logical thought and reasoning.
Let’s use ChatGPT as an example. If you asked ChatGPT to write a summary of the movie, “The Matrix”, it correctly identifies the movie and regurgitates all the important facts about the movie. That’s roughly 2 hours saved in watching the movie! But now, let’s say you ask it to give you a summary of a made-up movie, let’s call it, “The AI”. It ends up returning you with a polished paragraph of words sequenced in such a perfect manner that it almost appears as though this movie exists. However, it’s complete nonsense and falsehoods. There is no such movie!
The fact of the matter is that AI models thrive when there is a plethora of data. On the flip side, in the absence of data, AI flounders and ends up predicting a string of words that are completely irrelevant, foolish and outright wrong!
This underscores the importance of being knowledgeable in coding. As a project becomes more complex, chances are AI would not have sufficient data to predict accurate lines of code. It would begin hallucinating package functions and modules that don’t exist. This can be frustrating for a non-coder as they would not know why their AI-generated code isn’t working as it should. On a micro-level, AI-generated code might introduce subtle bugs, especially those that result from interactions between far-flung lines of code. Those kinds of bugs are the hardest to track down, requiring sophisticated critical thinking skills that a language prediction engine simply lacks. Trying to correct the mistakes would again require knowledge of coding, which non-coders would not have, thus leading to a catch-22 situation.
Look at recent troubles at Boeing. For years, they prioritised financiers over engineers within their management, which has inevitably run the aviation giant to the ground. Over the recent months, their 737s have either plunged mid-flight, lost metal door plugs mid-air or jammed up upon landing. The aviation titan is in crisis as they have forgotten the importance of engineers. Similarly, by de-emphasising coding education, we risk producing programmers over-reliant on AI, and lacking in sophisticated coding skills. The result of which would be programs that end up failing, and potentially resulting in disasters, rather than helping the world run smoothly.
3. Coding will become a key competitive advantage in more fields.
In a world where coding becomes as simple as talking to an AI machine, being able to develop games, mobile applications or even websites, would no longer be seen as a task reserved solely for programmers. Instead, we expect that with the help of AI, more people will begin dabbling in programming. Coding skills would then become a differentiating factor for professions such as researchers, lawyers, engineers and perhaps even teachers.
Imagine an architect that writes code to generate intricate building blueprints, that would have otherwise been too labour-intensive to create by hand. Or a hawker who programs an automated queueing system as soon as they notice an untenably long queue during the peak lunch hour. Administrators might turn to code to automate their routine paperwork and filing duties, saving them from hours of tedium. These citizen developers can improve their working environment, and make the world a more pleasant place, because they are on the ground experiencing these issues. Thanks to AI, they would be empowered to develop their own solutions through coding. But of course, they won’t be able to do so without first having some coding and computer science skills.
Just as knowing English is no longer a luxury but is now a necessity to thrive in our world, we believe coding and computer science will likewise become one of the foundational pillars forming the backbone of one’s employability in the future.
Final remarks…
Although we agree with Huang that AI will bring about a drastic change in how coding will be conducted in future, we disagree with him that we should stop emphasising coding and computer science education.
We expect instead that code will become ubiquitous in an AI-driven world and therefore believe teaching computational thinking, debugging skills, and most importantly high-level creative coding will be quintessential for youths to succeed in the unpredictable world that lies ahead of them.