Young Programmers No Longer Know How to Code: AI Is to Coding What Calculators Were to Math Decades Ago

In Isaac Asimov’s short story The Feeling of Power, Myron Aub is a bald-headed little man who rediscovers a skill that humanity had forgotten: He can multiply without the aid of a computer. After showcasing his ability, the members of the New Pentagon are astounded that someone can perform multiplication using only pen and paper.

Originally published in February 1958, the 3,700-word tale is truly remarkable. I highly recommend reading it. It portrays a future where humans no longer know how to perform mathematical operations and depend entirely on computers for everything. When Aub reverse-engineers the calculation process, he sparks an extraordinary situation that leads to a surprising–and sadly predictable–conclusion.

I first read that story nearly 30 years ago and found it fascinating even then. I’m a huge fan of Asimov’s Foundation saga, but I hated the Apple TV+ adaptation for betraying that legacy. The prolific author conveyed a clear message: Perhaps relying too much on machines isn’t such a good idea. Or maybe it is?

The Programmer Who Didn’t Know How to Program

In fact, this is what’s happening in the programming world, which is one of the segments most affected by the rise of artificial intelligence. AI models have proven to be invaluable assistants in programming tasks. Just six months after the launch of ChatGPT, 90% of professionals were using AI for coding.

These models have become increasingly adept at performing programming tasks. The popularity of tools like Cursor demonstrates that the concept of “effortless programming” is becoming more prominent. Even individuals who didn’t know how to program or were unfamiliar with certain coding languages can now create amazing apps. While AI isn’t perfect, it’s clear that it’s continually improving.

However, this advancement poses a risk: People may forget the fundamental programming skills. A developer named Namanyay Goel recently expressed this concern in his blog, saying that “new junior developers can’t actually code.”

While AI enables developers to produce more code than ever before, many young programmers don’t understand why that code works or whether there might be other, better approaches to solving problems.

In his experience, the new generation of programmers frequently uses tools like ChatGPT, Copilot, and Claude. He notes that this reliance enables them to produce more code than ever before. However, he believes that these young developers often lack a deep understanding of why their code works or whether there might be better alternatives. “We’re trading deep understanding for quick fixes, and while it feels great in the moment, we’re going to pay for this later,” Goel points out.

Goel says that not too long ago, sites like Stack Overflow were excellent resources for programmers. When developers asked questions, they typically received answers that helped them understand the underlying principles. This knowledge was freely available. Additionally, experienced developers unknowingly acted as mentors to the younger generation in many cases.

“AI gives you answers, but the knowledge you gain is shallow. With Stack Overflow, you had to read multiple expert discussions to get the full picture. It was slower, but you came out understanding not just what worked, but why it worked.”

More from Xataka On

Anthropic Is One of the Few Companies That Can Keep Up With OpenAI. The Proof: Claude Can Now Write and Run Code

Not everyone shares the same opinion on this matter. One Slashdot user highlighted, “Stack Overflow has been a source of terrible coding advice and overreliance on copypasta by juniors for far longer than [machine learning] has been a problem.”

However, Goel believes that there’s still hope. While AI can certainly assist young programmers, it’s also important to continue learning from it. He says in his blog, “When it gives you an answer, interrogate it. Ask it why.” He also suggests collaborating with your development team to discuss and debate code, which can help generate new ideas. Alternatively, you might engage with developers on platforms like Reddit, Discord, and Mastodon.

The Calculator Shows You the Way

Goel’s reflection is thought-provoking, but for many, it may seem like a meaningless debate. Another Slashdot user explained that they’ve been interviewing junior developers for two decades, adding, “A shocking percentage of them didn’t understand basic concepts… AI has nothing to do with it, it’s the state of education–plenty of schools are turning out graduates that just do not understand their chosen field.”

Other users echoed their sentiments, drawing parallels to past industrial and technological revolutions, when jobs or disciplines were overtaken by machines. While children are still taught to perform mathematical operations by hand, most adults don’t do them by hand. Using a calculator is much quicker.

Calculators may have initially displaced some workers who performed basic arithmetic tasks, but their long-term impact has been overwhelmingly positive.

First and foremost, mathematicians still exist, but they’re now focused on much more complex problems that machines, including AI, can’t solve. Calculators have increased their value by allowing them to concentrate on these intricate issues rather than on calculations that could be prone to human error.

Moreover, calculators have significantly enhanced productivity and efficiency. They’ve relieved professionals of repetitive tasks, enabling them to focus on responsibilities that require uniquely human skills, which machines can’t replace, at least for the time being.

It’s quite possible that a similar scenario will unfold with AI and programming. AI has already demonstrated its value as an aid to programmers, leading to the idea that developers may transition into software architects and engineers. In this new landscape, once the code has been confirmed to work optimally, they would delegate most programming tasks to AI.

In February 2024, Nvidia CEO Jensen Huang said that at this point, no one should learn to program because AI will handle it for us. The industry is moving in that direction. As such, it’s likely that, at some point, knowing how to program may not hold much value if AI can perform coding tasks as well, or even better, than humans. As has happened in many other fields, we’ll have to adapt.

Perhaps, many years from now, a short and bald elderly man named Myron Aub will astonish everyone by revealing that he’s learned to code without the help of machines.

Images | Arnold Francisca | Mohammad Rahmani

Related | The CEO of AWS Has One Thing Clear: Software Engineers Won't Be Coding in the Future, AI Will