Nerves are already rattled over recent artificial intelligence advances in writing code. Tensions continue to rise alongside unending releases of new programming features. For example, ChatGPT recently went multimodal. ChatGPT Plus and Enterprise versions now have voice and image capabilities, which leads to “napkin developers,” as NativeScripting founder and lead trainer Alex Ziskind called them in this video demo. This upgrade enables ChatGPT to write full code from nothing more than a basic idea scribbled on a napkin.
AI model upgrades like this one raise the bar considerably in the AI versus software developers job competition. But it still doesn’t mean developers will lose to the machine.
“What’s interesting about the impact AI has over engineering is that it’s only replacing one skill — coding,” says On Freund, an ex-WeWork VP of engineering and now CEO of Wilco, an immersive upskilling platform for software developers. “The job of a software engineer, however, requires so many other skills, whether hard skills, such as systems thinking, architecture and debugging, or soft skills, such as verbal and textual communication. These skills are now becoming even more important than ever.”
But it’s not just the skill set requirements that are changing in software development — it’s the nature of the work, too.
“Generative AI’s unique ability to understand language opens up an opportunity to finally migrate the enormous tech legacy that has been piling up for 40 years. Many of these old applications are not properly documented and hard to properly gauge. GenAI can help provide a report on the logic of these applications and accelerate the modernization and migration of them into more modern systems,” says Thierry Bonfante, chief product officer at Unqork.
But shifts of that magnitude will also create some job casualties if developers aren’t fast to change their game.
“This unique opportunity to finally migrate away from legacy 40-year-old code will impact people who have created a highly sought after niche market for themselves; for example, COBOL [Common Business Oriented Language] developers,” Bonfante adds.
The Current Score
In a nutshell, AI will consume software developer jobs that focus on the lower end of programming tasks. Anything repetitive and predictable is a candidate for automation, which also lends itself to AI tasking. That’s why AI models are heavily focused on writing code. But that’s not the threat to software developers that many would assume it to be.
“Software development is less about writing software and more about understanding the problem you are trying to solve,” says Louis Lang, CTO and co-founder of Phylum. “While the likes of ChatGPT and Copilot might make the writing process quicker, it has a long way to go before it can reason through a novel problem domain. Making development faster with AI only applies to scaffolding new projects and writing well-trodden code and even this seems problematic from time to time. If you try to produce something that requires deep expertise, AI will not help you.”
But what jobs it does destroy, it replaces with new roles. For example, AI is itself software and as such requires developers.
“With the rise of generative AI, software developers play a pivotal role in designing, building, and maintaining the underlying infrastructure that powers AI applications,” says Adam Prout, CTO of SingleStore, a cloud-native database. “Their position is vital to implementing algorithms, creating data pipelines, and optimizing models in close collaboration with data scientists and machine learning engineers. The expertise of a software developer is integral to bringing AI projects from conceptualization to real-world deployment.”
But it’s not just AI models that require more programmers and engineers, it’s also the changes AI brings about.
“We’ll see the supply of programmers increase dramatically, along with demand for engineers,” says Stephen O’Brien, head of product at Tines. “Large language models are built and operated by software engineers, built upon by software engineers, and their immediate economic impact demonstrates the need for more upstream and downstream software. This technology will transform, not end, software engineering.”
AI Reshapes the Software Development Job Market
However, it’s naïve to think that some programming jobs won’t soon be permanently gone. Even so, some software developers will enjoy renewed job security and more career opportunities.
“Some skills, like rapid recall of arcane API interfaces, are suddenly irrelevant, while others, like understanding what real-world problem needs to be fixed and why, are proving durable,” O’Brien says.
Other developer groups and teams won’t be affected at all by AI, at least in the short-term.
“All programmers are not the same and they work in many different ways,” says Richard Luna, CEO of Protected Harbor, an IT firm. “For example, hardware programmers are very different then GUI programmers, who are very different than network programmers.”
“Script programmers, programmers who write servers scripts should be concerned because AI at this moment can write the script and do a good job. GUI programmers, programmers who do application design don’t have to worry too much yet because so much about GUI programming is subjective,” Luna adds.
By all accounts, the software development job market is being reshaped by AI. Roles will change, certainly. But change is nothing new to the professionals in this field and winners always rise from the challenges.
“Software developers bring logical reasoning, problem-solving acumen, and algorithmic prowess to the table. Give a progressive developer AI tools, and they won’t just adapt, they’ll dominate,” says Fabian Kochem, Head of Global Product Strategy at 1NCE, a provider of IoT software and connectivity.