Why demand for code is infinite: How AI creates more developer jobs

Not only is there a future for software development, but we’re on the cusp of enormous demand for code developed by humans.

February 9, 2026

Why demand for code is infinite: How AI creates more developer jobs

Not only is there a future for software development, but we’re on the cusp of enormous demand for code developed by humans.

• Credit: Alexandra Francis*

Much has been said about AI decimating the job market for developers. In an industry changing this quickly, we certainly can’t blame people—especially junior and aspiring engineers—for worrying that the AI automation wave might sweep their jobs out from under them.

More existentially, some are wondering whether the age of AI, and particularly the rise of vibe coding, signals the demise of software development. But reports of its death are, to paraphrase Mark Twain, greatly exaggerated.

Not only is there a future for software development, but we’d like to suggest that we’re on the cusp of enormous demand for code developed by humans. From our perspective, AI represents a platform shift that’s changing what it looks like to build software and ushering in a period of explosive demand for ambitious, innovative, and highly specialized code.

A recent conversation between Stack Overflow CEO Prashanth Chandrasekar and OpenAI Head of Developer Experience Romain Huet got us thinking about how developers will build everything that’s suddenly becoming possible. Let’s explore how AI will drive new jobs (and new ways of approaching those jobs) for developers.

The platform shift perspective

Anytime you want to understand where you’re headed, look at where you’ve been. AI isn’t the first major platform shift, and each of those shifts has fundamentally changed how we work.

In the mid-90s, the internet emerged as a mainstream technology. Handwritten college applications gave way to online forms. Physical libraries became digital repositories. Entire business models that couldn't have existed before—ecommerce, search engines, social networks—became ubiquitous.

Then came mobile computing and the cloud. Arguably, they’re part of the same shift: The client-server model for the early internet was browser-data center; it evolved into mobile device-cloud.

Smartphones changed where and how we interact with technology. Apps went from something you ordered with drinks to the world’s interface. Mobile-first companies proliferated. Again, fears of job displacement gave way to whole new careers: mobile developers, UX designers.

Cloud computing abstracted away the complexity of managing physical infrastructure. DevOps emerged as a discipline. Companies that once needed massive IT departments could spin up global-scale applications overnight. More abstraction, more possibility, more jobs.

Like these seismic shifts, AI is redefining how we learn, create, and solve problems. Consider the evolution of abstractions in learning to code. Once, you learned from textbooks, painstakingly working through examples and asking classmates or instructors if you got stuck. In 2008, Stack Overflow democratized that knowledge. Suddenly, you could tap into the collective wisdom of millions of developers worldwide, finding answers to problems that would have taken hours to solve. That was a major abstraction layer: from personal networks to global knowledge sharing.

Now AI coding assistants have introduced another abstraction layer. We’ve gone from searching for solutions to conversing with an intelligent system that can generate, explain, and iterate code in real time.

None of these abstraction layers eliminated the need for developers. Instead, they changed what skills and experiences organizations were looking for. They unlocked new possibilities and drove demand for people who could build them.

Imagination drives inevitable innovation

Prashanth Chandrasekar, Stack Overflow’s CEO, is a lifelong Trekkie. When asked about how AI will drive demand for code, he points to the technology of the Starship Enterprise: Replicators that materialize objects from thin air. Holographic environments indistinguishable from reality. Voice-activated AI that anticipates crew needs. Warp drives that fold space-time.

"Once you imagine something," Chandrasekar observes, "it's inevitable that we're gonna go build it at some point."

The human mind is an imagination engine, constantly coming up with better ways of doing things. Each of those imagined futures requires software to become reality.

For every solved problem, we discover new ones to fix. In curing one disease, you might discover biomarkers that point to five others. In optimizing one supply chain, you might recognize inefficiencies in related systems and understand how to fix them. In building one AI capability, you might imagine a dozen other applications (If it can do x, what about y?). Progress doesn't satiate our ambitions; it whets them.

Consider one domain AI is reshaping: Drug discovery is becoming, at least in part, a computational problem. Scientists are moving from trial-and-error chemistry to AI-guided molecular design. Simulations that once took months now take days. Coming from a family of doctors, Chandrasekar reflects, "It'd be amazing if we could use AI to actually solve or cure some of the world's biggest ailments that debilitate a lot of people." Every disease we target, every biological pathway we map, every personalized treatment we develop—all of it requires sophisticated software, maintained and improved by developers.

The Cambrian explosion of AI companies

Look at any AI market map and you'll see thousands of companies, each attacking a different layer of the stack or a different vertical application. Venture capitalists are funding this explosion because they see the writing on the wall: AI is fracturing into myriad specialized niches.

This Cambrian explosion is driving demand for developers across every layer:

• The hardware layer is in full reinvention mode. General-purpose CPUs are giving way to specialized AI chips: GPUs, TPUs, neuromorphic processors, quantum computing experiments. Each architecture requires firmware, drivers, optimization libraries, and toolchains. Semiconductor companies are hiring engineers to build the physical infrastructure required.
• The model layer is diversifying rapidly. There's a proliferation of specialized models fine-tuned for specific domains, from medical diagnosis and legal document analysis to code generation, image synthesis, and protein folding. Each model needs training pipelines, evaluation frameworks, deployment infrastructure, and continuous improvement cycles, driving demand for data scientists and ML engineers.
• The infrastructure layer is being rebuilt for AI workloads. Serving LLMs efficiently requires new approaches to compute allocation, caching strategies, load balancing, and cost optimization. People are building entire businesses around making AI inference faster and cheaper. Every one of these businesses needs engineers who understand distributed systems, performance optimization, and the unique characteristics of AI workloads.
• The application layer may be where the most explosive growth is happening. Every industry, every workflow, and every use case is being reimagined with AI as a central component. Legal tech companies are building AI contract analyzers. Financial services companies are designing fraud detection systems. Manufacturing companies are working on predictive maintenance platforms. Educational companies are creating personalized learning systems. You get the idea.

[...]

Original source