What If AI Doesn’t Kill Developer Jobs?

There’s been no shortage of opinion pieces about AI replacing software engineers. Most rehash the same logic: if the machine writes code, the human becomes optional. But last week, Matt Asay published a short, sharp column in InfoWorld that questions that narrative. He suggests that AI might not shrink development teams, but expand them...by increasing demand for software, not reducing the people who write it.

It’s a take worth serious attention. And while I’m not ready to fully endorse it, I wouldn’t rule it out either...especially in the near to mid-term. The numbers seem to back him up.

The mechanism Asay points to is the Jevons paradox: when a resource becomes more efficient to use, we tend to use more of it, not less. Here, that resource is developer time.

And the efficiency gains are real:

• An internal GitHub study showed developers using Copilot completed coding tasks 55% faster, with a higher success rate (78% vs. 70%) compared to those without AI assistance.

• A multi-company field study covering over 1,900 engineers found that access to Copilot led to 13–22% more pull requests per week.

• Microsoft, ZoomInfo, and others report time savings of 40–50% on common dev tasks.

AI-Enabled Nearshore Engineers: The Ultimate Competitive Edge

The future of software engineering isn’t just AI... it’s AI-powered teams. By combining AI-driven productivity with top-tier remote nearshore engineers, companies unlock exponential efficiency at a 40-60% lower cost, all while collaborating in the same time zone.

✅ AI supercharges senior engineers—faster development, fewer hires needed
✅ Nearshore talent = same time zones—real-time collaboration, no delays
✅ Elite engineering at significant savings—scale smarter, faster, better

But here’s the key observation...as Asay writes, "we don’t ever respond to such productivity improvements by dialing back software development." We use those savings to build more.

And more code means more complexity: more integrations, more testing, more security risk, more documentation, more maintenance. The time saved by AI gets funneled straight back into a growing backlog.

This isn’t theoretical. Gartner forecasts that by 2027, 80% of developers will need at least baseline AI proficiency. And between 2023 and 2024, demand for AI/ML engineers rose 148%, even while general “developer” listings declined slightly. The market isn’t contracting...it’s shifting.

What I find interesting in Asay’s argument is his framing of developers as a kind of creative, interpretive force. They’re not just coding...they’re shaping the rules for how AI writes, responds, and adapts. They’re deciding which outputs to trust. They’re tuning prompts, debugging hallucinations, and building guardrails. That’s not replacement. That’s orchestration.

So yes, AI might shrink the effort required to write a line of code...but it’s also expanding the perimeter of what we can build. And with that expansion comes the need for judgment, context, and coordination...things the machine doesn’t do well (yet).

Long-term? Who knows. Maybe five, ten years from now, one engineer will do the work of a dozen. Maybe we’ll see micro-teams scaling entire product stacks. But that’s not what’s happening today.

Today, we’re building more. Faster. Broader. And there’s a strong case to be made that...at least for now...we’ll need more developers, not fewer, to make it all work.

Recommended Reads

✔️AI will require more software developers, not fewer (InfoWorld)

✔️Productivity and Competition Effects of AI (American Economic Association)

✔️The rise of the "AI engineer" and what it means for the future of tech jobs (Business Insider)

✔️Where developers feel AI coding tools are working—and where they’re missing the mark (Stack Overflow)