Why AI Will Create More Engineering Jobs Not Fewer in 2026

Will artificial intelligence shrink the number of engineering jobs or grow it? In a widely shared opinion piece for Fast Company published on May 24, 2026, veteran technology leader Joe Bertolami argues the answer is growth, not contraction. Bertolami, who has led software engineering at Microsoft, Snap, and Google across two decades, writes that the role was never really about writing code in the first place. It was about solving problems and reducing complexity. If that is true, he argues, then tools that automate the typing do not remove the engineer. They free the engineer to do more of the work that actually matters. At Metaintro, we track these shifts so job seekers can read past the scary headlines and see where the real opportunity is.

What Is Joe Bertolami Actually Arguing?

The core of Bertolami's argument is a redefinition. He contends that the public debate confuses software engineering with the act of writing code, when in practice coding was only ever one tool in a much larger job. The real work, he argues, has always been deciding what to build, understanding why it matters, and navigating the tradeoffs between speed, cost, quality, and risk.

From that starting point, his conclusion follows naturally. AI agents can generate code, scaffold services, write tests, and produce boilerplate faster than any human. But they cannot decide what a company should build, judge whether a feature is worth the maintenance burden, or weigh a security risk against a launch deadline. Those decisions stay with people. As Bertolami puts it, coding is no longer the special sauce. Context is.

It is worth being precise about what this is. This is an opinion and analysis piece from one experienced leader, not a peer-reviewed forecast. Bertolami is making an argument, and reasonable people disagree with parts of it. But it is an informed argument from someone who has run large engineering organizations, and it lines up with a pattern that career data is starting to show.

Does the Job Market Data Back This Up?

The long-range federal numbers lean toward Bertolami's optimism. The Bureau of Labor Statistics projects that employment of software developers, quality assurance analysts, and testers will grow about 15 percent between 2024 and 2034, far faster than the average for all occupations. That category already accounted for roughly 1.7 million jobs in 2024, with a median annual wage of about 133,080 dollars as of May 2024. A field that is supposedly dying does not usually carry a double-digit growth projection.

The shorter-range picture is messier, and honesty requires saying so. Industry trackers and reports compiled by outlets like TechTarget describe a bifurcated market in 2026, where demand for AI and machine learning roles is climbing sharply while postings for general software roles and entry-level positions have softened. So the total may grow over a decade even while the mix of who gets hired this quarter keeps shifting.

There are concrete signs of that growth even inside the turbulence. As we reported in our coverage of how software engineer job listings climbed in 2026, employers have been adding openings tied directly to building and supervising AI systems, even while they trim roles elsewhere. The same dynamic shows up in our look at how AI coding tools are turning developers into design strategists, where the work moves up the value chain toward architecture and decision-making. Read together, the federal projection and the on-the-ground postings tell a consistent story. The pie is getting bigger, but the slices are being cut differently.

The takeaway is not that everything is fine. It is that the headline number and the lived experience are both real. The category is expanding over time, but the composition of that demand is moving toward different skills than the ones that defined the last hiring boom.

Why Would AI Increase Demand for Engineers?

The intuitive fear is simple. If a tool makes each engineer more productive, companies need fewer engineers. Bertolami's counter draws on a well-known economic pattern that has appeared with other automation waves. When the cost of producing software drops, the amount of software the world wants tends to rise, sometimes faster than the productivity gains themselves.

Think of it this way. When building a feature took a quarter, many ideas were never funded because they were not worth the cost. When the same feature takes a week, projects that were previously too expensive suddenly clear the bar. More projects clearing the bar means more systems to design, integrate, secure, and maintain. Each of those systems needs someone accountable for the result.

Bertolami also argues the surface area of software is expanding into industries that barely used it before, from logistics to healthcare to manufacturing. As more of the economy becomes software-driven, the number of places that need engineering judgment grows. Cheaper code, in this view, is an accelerant for demand rather than a replacement for the people who direct it.

There is a second force at work that is easy to overlook. Every AI system a company adopts becomes its own surface that needs engineering attention. Someone has to integrate the model into existing systems, monitor it for drift, secure the data flowing through it, manage its cost, and stand behind the result when it goes wrong. The build-out of AI infrastructure has become a hiring story in its own right, as we covered in our reporting on how AI data centers are creating thousands of new jobs. In other words, the same wave of automation that worries engineers is also generating fresh categories of engineering work that did not exist a few years ago.

Which Engineering Roles Are Actually Being Cut?

A balanced look has to acknowledge the roles that are shrinking right now, because the optimism does not reach everyone evenly. Reporting and market trackers summarized by sources such as TechTarget point to real pressure on entry-level and generalist positions, where the most routine coding tasks overlap most directly with what AI tools now handle.

Several large employers have paired aggressive AI investment with workforce reductions, cutting in legacy product lines while hiring in AI infrastructure. That pattern, layoffs in one area alongside hiring in another, is exactly why a single company can announce cuts and openings in the same season without contradiction. It is a reshuffle, not a simple shrink.

The scale of the current reshuffle is real, and the numbers are not small. Industry trackers like Layoffs.fyi put tech layoffs well past 100,000 roles in the first months of 2026, with a meaningful share explicitly tied to AI and automation as companies redirect budgets toward AI infrastructure. The squeeze on the first rung of the ladder is the sharpest part of this. Our reporting on how entry-level jobs are vanishing for 2026 graduates found entry-level postings down sharply since 2023, with some junior tech and data roles falling far more than the average. That is the uncomfortable counterweight to the optimistic long-range projection. A field can be growing in aggregate while the door to enter it narrows.

The people most exposed are those whose value was concentrated in producing standard code quickly, especially early-career engineers who have not yet built the judgment and context that Bertolami says now carries the premium. That is a genuine challenge, and it is the part of the story the optimistic headline tends to skip.

What Skills Matter Most Now?

If the job is shifting from typing to directing, the skills that matter shift with it. Bertolami frames the change as a move from the specialist coder to what he calls the generalist orchestrator, an engineer who coordinates AI tools rather than hand-writing every line.

Based on his argument, the abilities that gain value are problem framing and deciding what to build, understanding the business and user context behind a project, mapping constraints and tradeoffs, and owning system resilience and security. Supervising and coordinating AI agents becomes a core competency rather than a novelty. Communication also rises in importance, because directing tools and aligning teams is fundamentally about expressing intent clearly.

It helps to make this concrete. The skills moving up in value are the ones that are hard to hand to a tool, such as deciding which of three plausible architectures fits the company's real constraints, recognizing when an AI-generated solution is subtly wrong, and being able to explain a tradeoff to a non-technical stakeholder so a sensible decision gets made. These overlap heavily with what we describe in our guide to the human skills AI cannot replace, where judgment, communication, and ownership consistently outrank raw technical throughput. The practical version of this for a job seeker is simple to state and hard to fake. You want to be the person who can be handed an ambiguous problem and an AI toolkit and come back with a defensible answer.

None of this means coding fluency stops mattering. You cannot judge what an AI tool produces if you cannot read and reason about code. The shift is additive. You still need the foundation, and now you need the judgment layered on top of it. Aspiring engineers who treat AI tools as something to master rather than something to fear are building exactly the profile the growing part of the market rewards. You can follow how these expectations are evolving through the career coverage at Metaintro.

What Does the Wider Industry Say?

Bertolami is not alone in this view, which is worth noting because a single opinion is easy to dismiss. Analyst Josh Bersin has argued that AI is fundamentally a job-creation technology despite the prevailing fear, pointing to historical patterns where new tools expanded employment over time. The Pragmatic Engineer newsletter has documented a 2026 market that is recovering unevenly, with strong demand for experienced and AI-fluent engineers alongside a tougher road for newcomers.

Even consulting groups studying the talent pipeline, such as Boston Consulting Group, have long warned of engineering talent shortages rather than surpluses, a framing that sits awkwardly with the idea that engineers are about to be obsolete.

The skeptics deserve a hearing too, because the optimistic case is not unanimous. Some economists and engineering leaders warn that if AI tools keep improving, the productivity gains could eventually outrun demand, leaving fewer total seats even in a growing market. Others point out that a decade-long projection anchored in 2024 may not fully price in how fast generative coding tools have moved since. The fair reading is that the long-range data and the working leaders currently lean optimistic, while acknowledging real short-term pain, and that no one can be certain how the next few years play out. What is not in serious dispute is the direction of the skill shift, away from raw code production and toward judgment, even among those who disagree about the headcount math.

The honest synthesis across these voices is that the field is restructuring rather than disappearing. That is also how Metaintro reads the moment, and it is the difference between a story about endings and a story about repositioning.

What This Means for Your Career?

If you are an engineer or want to become one, the practical message is to lean into the skills that are getting more valuable rather than retreating from the field. Get genuinely good at directing AI tools instead of treating them as a threat. Practice articulating what to build and why, not just how. Build the judgment, security awareness, and systems thinking that no agent can supply on its own.

For early-career professionals facing the tightest part of the market, the path is harder but not closed. Differentiate on context and ownership early. Take on projects where you have to decide tradeoffs, not just complete tickets. Show that you can use AI to ship more, then explain the choices behind what you shipped. That is the profile that survives a reshuffle.

How you present that profile matters as much as building it. Vague claims about being AI-savvy get filtered out fast by both recruiters and the screening tools they use, so specificity wins. Our walkthrough on how to show AI fluency on your resume makes the same point Bertolami implies for the work itself, which is to name the tools you actually used, the workflow you actually built, and the outcome you actually shipped. The engineers who frame their experience around problems solved and decisions owned, rather than lines of code produced, read as exactly the kind of judgment-first hire that the growing part of this market is competing for. That framing is also a hedge. Whatever the next tool does, the ability to direct it toward a real outcome is the part of the job that keeps its value.

Lacey Kaelani, CEO of Metaintro, told People Managing People that "AI is not completely eliminating roles, but instead restructuring roles and therefore slowing hiring for some jobs." That restructuring is the whole story in one sentence. The engineers who thrive will be the ones who shift toward the roles AI is creating rather than clinging to the ones it is automating.

People Also Asked

Will AI replace software engineers in 2026?

A: Most evidence points to restructuring rather than replacement. The Bureau of Labor Statistics still projects about 15 percent growth for software developers through 2034, and leaders like Joe Bertolami argue AI automates the typing while leaving the judgment, design, and accountability to people. Routine coding roles are under the most pressure, but the broader field is expanding.

What engineering skills are most valuable as AI grows?

A: The skills gaining value are problem framing and deciding what to build, business and user context, mapping constraints and tradeoffs, system security and resilience, and the ability to direct and supervise AI tools. Coding fluency still matters as a foundation, but judgment layered on top of it is what increasingly commands a premium.

Is now a bad time to become a software engineer?

A: The market is harder for entry-level candidates than it was a few years ago, with softer demand for routine coding roles. But the long-term outlook remains strong, and newcomers who build context, ownership, and AI fluency early can still position themselves for the part of the market that is growing rather than shrinking.

Future-proof your career with Metaintro, AI-powered tools that help you adapt and thrive. Sign up to stay ahead.