For years, the advice to “learn to code” has been repeated as a gateway to secure jobs and opportunities in technology. It became almost a cultural mantra, often suggested as a way into stable work or even as a fallback career path. But as artificial intelligence steadily automates programming tasks once seen as exclusively human, the usefulness of that advice has come under scrutiny. What was once considered a foundation of digital literacy now appears more complex, and opinions differ widely on whether it still carries the same weight.
Automation Changes the Landscape
Artificial intelligence has already reshaped software development. At Google, a substantial share of the company’s code is now machine-generated, and Microsoft has said that certain projects rely on AI to write between a fifth and a third of their code. These companies are also deploying automated systems to review the results. Tools such as GitHub Copilot and Codex have accelerated the pace of development, producing working blocks of code that once took engineers hours to draft. The shift has prompted debate about whether programming remains an essential skill or whether new abilities are becoming more valuable.
A Case for Understanding the Basics
Several executives[1] believe coding still matters, even in a world where machines write much of it. Matthew Prince of Cloudflare, who leads a company built around online security and infrastructure, explained that his effectiveness as a chief executive depends on understanding how his engineers approach their work. He acknowledged that he no longer writes much code himself, but said that grasping the fundamentals makes him better at leading technical teams. For him, AI may lighten the workload, but security-critical industries will continue to rely on engineers with hands-on skills.
Liz Centoni of Cisco echoed this view, describing coding as a core discipline that shapes the way people think through problems. Having worked as a software engineer earlier in her career, she emphasised the value of practical knowledge when deciding how and when to apply technologies such as machine learning or generative AI. To her, those who know the tools in detail are better equipped to solve real business problems and adapt as technologies shift.
Jay Graber, who runs the social platform Bluesky, took a similar stance. She pointed out that while AI can generate and structure code, it cannot replace the ability to judge quality or design entire systems. In her view, those who lack a solid foundation risk outsourcing their thinking to machines and may struggle to evaluate whether the output is usable.
A Shift Toward Broader Skills
Other leaders placed more weight on abilities beyond pure coding. Andrew Anagnost of Autodesk argued that while programming once served as a reliable entry point into technical careers, the differentiator today lies in system-level thinking. He described future workers less as coders and more as creative orchestrators, guiding AI tools and combining disciplines to address complex challenges. For him, success depends on seeing how processes, technologies, and people fit together, not just on producing lines of code.
Salesforce executive Jayesh Govindarajan went further, suggesting that agency — the drive and capacity to decide what problems to solve — is now more valuable than technical coding skills. He described AI systems capable of carrying out a wide range of tasks, but stressed that humans must still provide purpose and direction. From this perspective, coding itself becomes secondary to initiative and problem selection.
Okta’s chief executive, Todd McKinnon, was even more sceptical about the universal push to code. He noted that learning programming suited his own way of thinking and proved useful to his career, but he dismissed the notion that everyone should be expected to acquire the skill. To him, insisting on coding as a universal requirement ignores the variety of strengths and perspectives people bring to the workforce.
Coding as a Shared Language
Others argued that coding should remain a common literacy. Google’s head of research, Yossi Matias, compared it to mathematics, saying that knowing the basics helps people understand how digital systems work. He suggested that the ability to code opens up opportunities rather than closes them, particularly in an era where AI expands what can be built on top of those foundations.
Dropbox executive Morgan Brown presented a more pragmatic view. He explained that deep coding may not be necessary for every role, but understanding the structure of the technology stack remains essential. In his experience, familiarity with tools such as SQL helps product leaders understand data and evaluate how systems function. Without that awareness, he argued, it is easy to miss possibilities or overlook problems.
The Community Perspective
Outside executive circles, the developer community also continues to wrestle with the same questions. On forums[2] such as Reddit[3], many seasoned programmers have said that AI should be treated as an accelerator rather than a substitute. They argue that those who do not understand how code works will struggle to use AI effectively, since generating fragments of software without comprehension often leads to errors and wasted effort. Experienced coders also highlight that while AI can speed up repetitive tasks, it still falls short on architectural judgement and complex debugging, areas where human expertise remains critical.
At the same time, other contributors see opportunity in the changes. Some point out that AI can make learning easier than in the past, acting as a tutor that explains concepts and accelerates progress. Beginners can ask for clarification and receive detailed feedback, shortening what once required years of trial and error. These participants describe the present moment as both a good time to learn and perhaps one of the last windows to do so before AI advances further.
A More Nuanced Outlook
Taken together, the views suggest that learning to code is no longer a simple recommendation. For some, it remains a vital foundation, useful not just for building software but for shaping the way problems are approached. For others, the real advantage lies in combining technical knowledge with creativity, initiative, and the ability to guide AI tools toward meaningful outcomes. What is clear is that programming itself is not disappearing; rather, it is shifting into a broader landscape where human judgement, design skills, and cross-disciplinary thinking hold increasing value.
For those entering the field today, the decision to learn code may depend less on whether machines can write it, and more on whether understanding it equips them to steer technology rather than be steered by it.
Notes: This post was edited/created using GenAI tools. Image: Carlos Perez/Unsplash
Read next: AI Search Traffic Rises Fast, But Organic Still Converts[4]
References
- ^ Several executives (www.businessinsider.com)
- ^ forums (www.reddit.com)
- ^ such as Reddit (www.reddit.com)
- ^ AI Search Traffic Rises Fast, But Organic Still Converts (www.digitalinformationworld.com)