Why Coding Is More Relevant Than Ever in the Age of AI

Rising fears that artificial intelligence will automate all fields has led some to claim coding is fast becoming redundant.

The emergence of no-code tools like GitHub Copilot has convinced many—particularly non-technical people—that the very coders who built these tools have worked themselves out of a job.

But leading tech voices, including Bill Gates and Andrew Ng, an AI and machine learning pioneer, have refuted this claim, explaining why it’s more important than ever to learn to code.

In a recent essay, Andrew Ng wrote:

“As these tools continue to make coding easier, this is the best time yet to learn to code, to learn the language of software, and learn to make computers do exactly what you want them to do.”

They are not alone. The Future of Jobs Report 2025 states:

“Technology-related roles are the fastest growing jobs in percentage terms, including Big Data Specialists, Fintech Engineers, AI and Machine Learning Specialists, and Software and Application Developers.”

And Worldmetrics, in their 2024 Market Data Report, projected that AI would create 540,000 software engineering jobs in 2025.

Just as literacy went from an esoteric skill among elites to the global standard for education, coding is becoming essential for everyone—if we are to understand, maintain, and wield the AI that will define the future.

Bill Gates, speaking to Sam Altman on his podcast Unconfuse Me with Bill Gates, stated that AI will change all industries beyond recognition—but said three jobs will remain the same, with coding first on his list.

AI Can’t Replace Coders—It Can’t Survive Without Them

Despite dramatic advances in artificial intelligence, Gates argues that developers will always be needed to debug, refine, and improve AI models.

More than just surviving, Gates believes coders who maintain and guide AI systems will increase in value—while humans won’t be needed “for most things” within 10 years.

Will Kencel, Senior Software Engineer on Microsoft’s Xbox team and a Codesmith alum with an MA in AI, agrees.

“It’s gonna open up a lot of new jobs where we need people to maintain these systems. You can replace some things with AI—bartenders and toll booth attendants—but the most over-hyped trend in 2025 is replacing engineers with AI. You can get AI to code, but it's not maintainable or reusable.”

In an interview with Codesmith, Kencel and other senior engineers, including Travis Huff—Engineering Manager at PayPay and former Honey lead—pointed out that while interest in AI is high, it’s nowhere near ready to replace human engineers.

Huff, who is currently looking to build his own company, said he plans to use AI tools to their fullest potential “to see how far one can go with these tools,” especially given how useful they are to developers who know how to use them properly.

This view—that AI can make engineers better, not obsolete—is echoed by Andrew Ng, who emphasized this point again in a March 2025 essay.

You Must Speak the Language of AI to Use It Effectively

Ng argues that while it’s never been easier to learn to code thanks to AI, it’s also more important than ever—especially if we want to effectively harness the tools shaping our future.

He recalls how, in 1960, Nobel laureate Herbert Simon predicted the “programming occupation will become extinct” because “computers will program themselves.”

Six decades later, it’s clear that this premonition—even from one of the era’s most brilliant minds—was wildly inaccurate.

The popular phrase, “AI won’t replace developers, but a developer using AI will,” often attributed to Jeff Atwood (co-founder of Stack Overflow), reflects today’s reality. Ironically, AI chatbots like ChatGPT have surpassed Stack Overflow as the go-to resource for programmers.

Yet, many experienced developers still believe AI is best used for low-level, time-consuming tasks, while the critical thinking and decision-making in real software development must remain human-driven.

Marselena Sequoia, Codesmith Engineering Mentor, voiced these concerns in an interview with CEO Will Sentance. She described herself as an “AI skeptic,” explaining:

“The downsides are so obvious." 

Whilst happy to use LLMs in laborious tasks like parsing through documentation, she is wary of the flaws in using AI to code, like “introducing hallucinations into the code base” and actively encourages students to build under the hood mental models rather than lean on AI.

Stephen Wolfram went even further, saying anyone who wants to stay ahead of AI must “learn computational thinking”, in an interview with Codesmith CEO Will Sentance and AI ML instructor Cyrus Yari.

Learn Computational Thinking to Wield AI Tools

In the same spirit, Stephen Wolfram, a renowned computer scientist, said anyone who wants to stay ahead of AI must develop computational thinking.

“Learn computational thinking. For every field X there's going to be a computational X.
If you figure out how to think about things computationally and you know the best tools, then you're in good shape.”

Wolfram, known for his deep understanding of software and large language models, makes it clear: the edge always belongs to the human who understands the system, not the system itself.

In a discussion with Will Sentance and AI/ML instructor Cyrus Yari, Wolfram reflected on decades of AI development. Just like Andrew Ng’s observations on failed predictions in the 1960s, Wolfram noted that:

Despite decades of AI hype, developers with deep computational knowledge have only grown in importance.

Yann LeCun, Chief AI Scientist at Meta, responded to Ng’s essay with agreement:

“We might have super intelligent AI assistants in a while, but we’ll be their boss. We will need to understand the foundations in order to lead them.”

LeCun, like Wolfram, has seen many AI trends come and go—and remains healthily skeptical about the idea that AI will replace engineers anytime soon.

At Davos, he addressed the topic head-on, stating:

“Human-level artificial intelligence is going to take a long time. Contrary to what you might hear from some people, we do not have a design for an intelligent system that would reach human intelligence.”

Now Is the Time to Learn to Code

From seasoned engineers to AI pioneers, there’s one message being repeated: AI won’t replace coders. It will rely on them more than ever.

If you want to understand, control, and shape the future of AI, you must learn to code—because this is just the beginning.

TLDR – Why AI Won’t Replace Coders

The Myth

• Many believe AI and no-code tools like GitHub Copilot are making coding obsolete.
• Non-technical voices suggest developers are automating themselves out of jobs.

Expert Voices Say Otherwise

• Bill Gates and Andrew Ng argue now is the best time to learn to code.
• AI will augment coders, not replace them.
• Developers are essential for debugging, refining, and advancing AI systems.

The Data

• The Future of Jobs Report 2025 highlights software and AI-related roles as the fastest-growing.
• Worldmetrics predicts 540,000 new software engineering jobs in 2025.

AI Makes Good Coders Better

• AI tools are helpful for repetitive or low-level tasks but lack reliability and long-term maintainability.
• Developers who learn to integrate AI into their workflow will outperform others.
• Senior engineers like Will Kencel and Travis Huff emphasize that AI is a tool for leverage, not replacement.

Computational Thinking Is Key

• Stephen Wolfram advises engineers to develop computational thinking to stay relevant.
• For every field, there will be a "computational X"—understanding this will be essential.
• Knowing how to think and work with AI tools gives humans a lasting edge.

You Still Need to Code

• Predictions of programming becoming obsolete have repeatedly proven wrong.
• Tools like ChatGPT can assist, but experienced engineers are still irreplaceable.
• AI leaders like Yann LeCun confirm we are far from human-level AI, and foundational knowledge is still critical.