In this edition of “Office Optional with Larry English,” Larry discusses the development of programming and how AI will change it but not make it obsolete.
Will AI replace programmers? Headlines over the last few years certainly suggest so.
In 2022, GitHub Copilot debuted, giving programmers an AI tool to accelerate their work. In 2024, Nvidia CEO Jensen Huang said nobody should learn to program, because everybody will be a programmer in the future. A few months ago, Salesforce CEO Marc Benioff announced the company wouldn’t hire any more software engineers in 2025 because of productivity gains from AI.
The reality around AI and programming is more nuanced than headlines suggest. Programming is not dead and programmers do not need to panic — but their roles will evolve as AI advances and disrupts the programming industry.
Where We Are Now: AI and Programming Today
Although software programming is far from dead, the employment of software programmers is on a slight decline. A 2024 report from ADP found employment numbers peaked in 2019 and have been decreasing ever since.
AI is one possible contributor to this trend. Tools like ChatGPT and Copilot can generate snippets of code, not only helping programmers work more efficiently, but also taking care of more basic, routine coding work. When programmers are more productive, companies theoretically now need fewer of them to complete the same amount of work.
But AI also offers opportunities. The technology frees software developers from busy work, allowing them to focus on bigger, meatier problems requiring creativity and collaboration. For smaller companies unable to afford a team of talented developers, AI-powered coding tools open the door to possibilities that would have been out of reach just a few years ago. For larger companies, AI-powered coding offers an opportunity to quickly get new products to market.
Still, AI has a long way to go before it replaces software programmers. AI can’t yet tackle complex business-wide problems that are large in scope and size. The technology can’t stand up entire systems or compile, deploy or provision hardware. And, while it can document and summarize requirements from interviews, it can’t define business requirements from scratch.
AI still makes a lot of mistakes and needs human review and oversight. It can’t offer entirely creative ideas or create something from scratch. True reasoning capabilities also remain out of reach for now. Although some newer AI models are called “reasoning models,” they’re still LLMs with limited reasoning functionality. Software developers are the ones filling in these gaps — and likely will be for some time.
In an interview with Business Insider, Instagram co-founder Mike Krieger summed up the current limitations of AI and programming: “Driving alignment and actually figuring out what to build is still the hard part, right? Like that is actually the only thing that is still best resolved by just getting together in a room and talking through the pros and cons.”
Where We’re Going: Will AI Replace Programmers In The Future?
Experts predict AI will enable coding to be done in natural language. This is what Nvidia CEO Jensen Huang was referencing when he said everyone will be programmers in the future.
But this, too, is nothing software programmers should fear. Natural language is simply another layer of abstraction, like all programming languages that came before it, such as Visual Basic, Python and COBOL. Every layer of abstraction, including natural language, requires the same thing: That programmers define the problem and parameters very precisely. It likely will be some time before AI will be able to replicate the soft skills needed for this precise communication.
As AI advances, programmers will be tasked with solving bigger, more complex and interesting problems because AI will handle the smaller challenges. Instead of writing code, they’ll be designing AI-powered programming systems and overseeing and correcting AI-written code.
Programmers also will play an essential role in the security, stability and resilience of systems. They’ll need an understanding of how systems work and be able to identify existing and potential gaps — an LLM won’t be able to identify an emergent gap or exposure surface. They’ll need to be able to troubleshoot complex systems.
Consider for a moment when no-code and low-code development tools arrived on the market. Suddenly, business processes were easier to create for the non-technical business user. Some software developers probably felt threatened, but programmers still needed to look at the code and troubleshoot issues.
And true, there will likely be fewer generalized software programmers in the future. Some programmers will self-select out of the industry. But the vast majority will be able to evolve their skillset and learn to work alongside AI. Many will lean into specialized verticals and deep coding expertise. Fewer students will study generalized computer science, because everyone will be able to learn coding quickly. Instead, technology-minded students will dive into niche areas of study around AI and specific industries.
Again, AI needs to advance quite a bit for all this to come to fruition. Programmers at my company, Centric Consulting, have been using AI to write code. They say that the tools make mistakes, and those mistakes are often obscure and difficult to find. Skilled programmers are needed to oversee and correct this AI-written code.
Before we reach the point of AI being able to independently write code, leaders need to ask themselves who will be accountable when AI makes a mistake. How do we decide the AI has done enough and the quality is there — especially when AI is also writing its own tests?
What Programmers Should Do Now
To future-proof their careers, programmers should embrace AI-enhanced development by learning how to integrate tools like GitHub Copilot or ChatGPT into their appdev workflow. They should also work on strengthening problem-solving skills, focusing on core programming concepts, architecture and system design and algorithmic thinking. AI-generated code still requires human oversight to ensure correctness, efficiency and security. Finally, programmers should continue to develop domain expertise, gaining deep industry knowledge and expertise at translating business problems into technical solutions.
Software programmers’ roles will look very different in 10 years. Change feels scary, but if you’ve worked for longer than a decade, your job has probably changed quite a bit already. And anyway, wouldn’t it be tiresome to do the same task the same way throughout a long career? AI requires most roles to evolve, adapt, acquire new skills and reimagine the future—this is not a problem for software engineers alone to solve.
This blog was originally published on Forbes.com.
We help companies maintain their competitive edge by partnering with them on their AI journey. Our recent work with a wholesale food delivery company shows how.
Don’t get left behind in the AI revolution. We guide leaders through the disruption AI may cause to help you go from uncertain to excited about the potential of using AI. Ready to get started? Let’s Talk
