AI Isn't Replacing Developers

By Jared Lyvers, ldnddev — June 5, 2026

The AI Replacement Myth Doesn't Match the Data

Every week another headline declares that artificial intelligence is coming for software developer jobs. The story is dramatic, easy to share, and almost entirely wrong. The actual data from 2026 tells a different story — one where developers are busier, more valuable, and more in demand than ever.

At ldnddev we have a front-row seat to how AI changes a custom website, Drupal, and WordPress development practice. We use it daily. We have opinions. And after building our own custom AI skills and watching teams across the industry adapt, our position is simple. AI is not replacing developers. It is multiplying them. The teams that figure that out first are the ones winning right now.

This post pulls together two pieces of evidence that should settle the question for any decision-maker still wondering whether to invest in human engineering talent: NVIDIA CEO Jensen Huang's recent comments at Computex 2026 and a series of reports from Oxford Economics analyzing where the "AI layoffs" narrative actually breaks down. Then we will show you how we use AI at ldnddev as an enhancement to our team rather than a substitute for it — and what that means for your project.

Jensen Huang: AI Is Driving Hiring, Not Layoffs

If anyone in technology has standing to talk about what AI will and will not do, it is the person whose company builds the chips powering most of it. Speaking at the GPU Technology Conference at Computex 2026 in Taipei, NVIDIA CEO Jensen Huang called the AI-job-loss narrative "complete nonsense." His argument is worth unpacking because it reframes the entire conversation.

Huang pointed out that the world has roughly 30 to 40 million software developers, and AI is effectively tripling their productivity. That sounds like a setup for layoffs. It is not. The reason is that the demand for software is not fixed. As Huang put it, "We need a trillion lines of code written." When the work to be done is essentially unbounded, making a developer three times more productive does not eliminate the developer. It expands what each one can ship.

His financial framing was even sharper. "If you can hire a software engineer and you could generate $9 trillion worth of productive work, why wouldn't you want to hire more software engineers?" That is not the language of an industry shrinking. That is the language of an industry where every additional engineer compounds output, and the smart move is to hire more — not fewer.

We see this play out at our scale every week. The faster our team can ship a Drupal feature, a WordPress integration, or a custom framework component, the more work clients send our way. Capacity creates demand. Demand creates more capacity. AI did not break that loop. It accelerated it.

Oxford Economics: "AI Layoffs" Are Mostly Corporate Cover

If Huang's view is dismissed as a chip vendor talking his book, the data from Oxford Economics deserves a careful read. Their 2026 analysis of AI-attributed job cuts found something most coverage missed: when companies announce layoffs and blame AI, the numbers usually do not back it up.

From January through November 2025, AI was cited as the reason for nearly 55,000 U.S. job cuts. That sounds like a lot. It is, in fact, about 4.5 percent of total reported job losses in the same window. Layoffs attributed to standard "market and economic conditions" were four times larger — roughly 245,000 jobs. Oxford Economics put it plainly: "Firms don't appear to be replacing workers with AI on a significant scale."

Their working theory for why companies still blame AI is sobering. AI sounds forward-looking. Past over-hiring sounds like a management mistake. Investors prefer the first story. So a meaningful share of "AI layoffs" appears to be a rebrand of routine headcount corrections that have nothing to do with whether AI can or cannot do the work.

One real signal does show up in the Oxford data and it is worth taking seriously. Employment for developers aged 22 to 25 fell nearly 20 percent since 2024, concentrated in boilerplate coding, scripted testing, and routine bug fixes. At the same companies, developers aged 30 and older saw headcount grow. The market is not replacing developers. It is rebalancing toward senior judgment and away from work AI can credibly automate. That is a different problem, and it has a different solution.

What's Really Changing: The Junior-to-Senior Rebalance

Pull the two threads together and the picture sharpens. AI is automating roughly 20 to 40 percent of routine engineering work — boilerplate, scripted tests, documentation, simple fixes. That frees senior engineers to spend their time on architecture, hard debugging, and the cross-system problems where human judgment still wins by a wide margin. Throughput in that range translates to a 20 to 30 percent productivity lift for teams that adopt AI well. Teams that adopt it badly often end up with more technical debt and worse maintenance burdens, which is its own separate warning.

What this means for your business is not "fire your developers." It is "make sure the developers you work with know how to use AI as a multiplier rather than a crutch." A senior engineer with a sharp AI workflow is not three engineers — they are a senior engineer who can ship the volume of three engineers without sacrificing the judgment that made them senior in the first place. Strip out the senior judgment and you get a stack of plausible code with no one accountable for whether it works.

If you are evaluating an agency, the question is not "do you use AI?" Everyone says yes. The real questions are: how do you use it, where do you not use it, and who is on the hook when something breaks? Those answers will tell you whether you are buying augmented engineering or AI-generated risk dressed up as a deliverable.

How ldnddev Uses AI as a Team Multiplier

Our position is straightforward. AI is part of every workflow at ldnddev. It is part of no decision that we will not personally defend. That distinction matters.

Concretely, here is what AI does on our team. It accelerates first-pass implementation of Drupal modules, WordPress blocks, and custom framework components. It pulls together draft content, SEO audits, and accessibility reports faster than we could by hand. It helps us scaffold tests, generate translations, and audit existing codebases for patterns we want to retire. Every line of that work passes through a human engineer before it touches a client environment. That review is not a formality. It is where our value lives.

What AI does not do at ldnddev: it does not pick your hosting strategy, design your information architecture, decide what to deprecate, or approve a deployment. Those decisions sit with engineers who can read your business context and explain the tradeoffs to a stakeholder who is not in the code with us. If we ever automated those decisions, we would be selling you the worst version of our service — code with no accountable owner.

A concrete example helps. When we run an accessibility audit on a client Drupal site, AI handles the initial sweep — flagging missing alt text, low-contrast color pairs, and unlabeled form controls across hundreds of templates in minutes. A senior engineer then walks the flagged list, separates the real findings from the false positives, prioritizes the fixes by user impact, and explains the tradeoffs to the client. The AI did the volume work that used to take days. The engineer did the judgment work that determines whether the audit is actually useful. Neither could have shipped the same outcome alone.

This is why our broader perspective on AI-powered development in 2026 centers on tooling that augments engineers rather than replaces them. The framework, the agent skills, the deterministic helpers — they all exist so a human engineer can move faster without losing the context that makes their work trustworthy. If you want a deeper look at how we are building those AI tools internally, our piece on building custom AI skills to enhance your workflow walks through the architecture.

What This Means for Your Project

If you are a decision-maker or marketing lead trying to plan a 2026 web or marketing initiative, the practical implication of all of this is simple. You should be skeptical of any partner who promises that AI alone will deliver your project. You should be equally skeptical of any partner who refuses to use AI at all and quietly bills you for work that should take a fraction of the time.

The agencies and in-house teams winning right now are the ones operating in the middle. They use AI aggressively for the parts of the work where it has earned trust — drafts, scaffolds, audits, refactoring suggestions, content variations — and they keep humans firmly in charge of architecture, accessibility, security, and the final say on what ships. That is the model we run. It is also the model the data supports. Productivity goes up. Quality stays up. And the team gets stronger over time because senior engineers spend their hours on the work that grows them as engineers.

For our clients, the visible result is faster delivery on the boring parts and more thoughtful attention on the parts that actually move the needle. You get the speed without paying for it in trust, security, or accessibility debt. For a closer look at how we apply the same principle to AI-powered web development in 2026, that companion piece covers our day-to-day toolchain in more detail.

The Future Belongs to AI-Augmented Teams

Here is the part the doom headlines miss. The teams adopting AI well are not shrinking. They are expanding what they can take on, sharpening their senior bench, and earning more trust from clients who care about outcomes. The teams refusing AI entirely are quietly falling behind on cycle time, and the teams trying to replace their engineers with AI are walking into the technical-debt trap that recent productivity research has been warning about all year.

The future of development is not human or AI. It is human with AI. Senior judgment on top, AI doing the heavy mechanical lifting underneath, and a clear line of accountability connecting the work to a person who can explain it. That is the playbook Jensen Huang is describing when he says AI creates jobs. It is the playbook Oxford Economics is describing when they say the layoff narrative does not match the data. And it is the playbook we run every day at ldnddev — because it is the only one that produces work we are willing to put our name on.

If you are weighing what to do next — whether that is modernizing a Drupal site, rebuilding a WordPress stack, or designing an AI-augmented workflow for your internal team — we would be glad to walk through it with you. The goal is not to replace anyone. The goal is to give your team the leverage to do the work that actually matters.

Until next time, Jared Lyvers