There’s a particular kind of delusion that grips the software industry every decade or so. In the 2010s, it was the belief that every company was a “tech company.” In the early 2020s, it was the conviction that learning to code was a universal ticket to the middle class. And now, in 2026, the delusion is that software developers can simply keep doing what they’ve been doing while three of the most powerful men in technology publicly announce — with varying degrees of diplomatic courtesy — that the profession is circling the drain.
Let’s listen to them. Not because they’re always right. But because they’re the ones building the drain.
The Three Horsemen of the Coding Apocalypse
Jensen Huang: “Everybody in the World Is Now a Programmer”
Jensen Huang, CEO of Nvidia — the company whose GPUs literally power the AI revolution — didn’t mince words at the 2024 World Government Summit in Dubai. For over a decade, every tech conference stage featured some executive insisting that children must learn to code. Huang walked up to that consensus and set it on fire.
His argument was elegant in its brutality: the entire point of advancing computing is to make programming unnecessary. The new programming language isn’t Python. It isn’t Rust. It’s English. It’s Mandarin. It’s whatever you spoke at the breakfast table this morning.
But here’s the twist that most developers missed while rage-tweeting about Huang’s comments: he didn’t stop at destroying their career narrative. He built a new one. At Davos in January 2026, Huang told BlackRock CEO Larry Fink that the AI boom is creating what he called “the largest infrastructure build-out in human history” — and that it’s generating six-figure opportunities for electricians, plumbers, steelworkers, and construction crews.
“It’s wonderful that the jobs are related to tradecraft,” he said. “We’re going to have plumbers and electricians and construction and steelworkers.”
Read that again. The man who sells the brains of the AI revolution is telling you the money is in the hands. In a separate interview with Channel 4 News, he was even more direct: “If you’re an electrician, you’re a plumber, a carpenter — we’re going to need hundreds of thousands of them to build all of these factories.”
When asked what he’d study if he were 20 again, Huang admitted he’d lean toward the physical sciences over software. The man whose net worth is built on silicon is telling 20-year-olds to go touch grass — and maybe install a breaker panel while they’re out there.
Dario Amodei: “We Might Be 6–12 Months Away”
If Huang was the prophet, Dario Amodei is the coroner. The CEO of Anthropic — the company behind Claude, one of the most capable AI systems on the planet — dropped a prediction at the World Economic Forum in Davos that should have sent every bootcamp graduate into cardiac arrest: AI could perform everything a software engineer does, end-to-end, within six to twelve months.
Not assist. Not augment. Replace.
“In 12 months, we may be in a world where AI is writing essentially all of the code,” Amodei said at a Council on Foreign Relations event. He later told Axios that AI could wipe out half of all entry-level white-collar jobs within five years and push unemployment to 10–20%.
And this isn’t theoretical hand-waving from someone who doesn’t use his own product. Amodei has stated that over 90% of the code produced at Anthropic is now written by AI. Human engineers at his company no longer write code — they review it. Their job has shifted from creation to quality assurance for a machine that learns faster than any junior developer ever could.
Let’s do some quick math on what this means for the average mid-career developer making $150K in a VHCOL city. If AI can do 90% of what you do, your employer doesn’t need ten of you. They need one of you and a subscription. And that one person had better be spectacular — not at writing code, but at orchestrating the machine that writes it. The other nine? Well. Have you considered HVAC?
Sam Altman: “I Have So Much Gratitude”
Sam Altman’s contribution to this funeral procession has been characteristically Altman: a velvet glove concealing a fist made of pure disruption. In March 2025, he stated that AI would “gradually” replace software engineers, while insisting he was more interested in making coders “10x more productive.” It sounded like a compliment until you realized that 10x productivity means you need one-tenth the headcount.
Then, in mid-March 2026, Altman posted on X: “I have so much gratitude to people who wrote extremely complex software character-by-character. It already feels difficult to remember how much effort it really took.”
The internet’s reaction was, to put it mildly, volcanic. “Nothing says ‘you’re being replaced’ quite like a heartfelt thank you from the guy doing the replacing,” one user wrote. Others called the post a eulogy disguised as a LinkedIn endorsement.
In a separate Federal Reserve talk, Altman noted that programmers using AI tools report being “10 times more productive,” and that a complex programming task that would’ve taken a skilled human 20–40 hours was completed by an upcoming AI model in five minutes — for less than a dollar of compute.
A dollar. For the same work that a senior engineer would bill at $10,000.
You can fight that math if you want. But math doesn’t care about your feelings, and neither does a CEO’s budget spreadsheet.
The Case for the Trades: An Exercise in Arithmetic and Humility
So where does a displaced developer go? Let’s look at what the labor market is actually screaming for while tech lays off tens of thousands.
The numbers are almost comically lopsided:
The U.S. construction industry is short more than 500,000 skilled workers in 2026. Not in ten years. Now. The Associated Builders and Contractors reports that 92% of construction firms actively hiring say they can’t find qualified workers. The Bureau of Labor Statistics projects electrician employment growing 9% through 2034 — nearly three times the average for all occupations — with approximately 81,000 openings per year. McKinsey estimates the data center boom alone will require an additional 130,000 trained electricians and 240,000 construction laborers by 2030.
Meanwhile, in software land: Big tech reduced new graduate hiring by 25% from 2023 to 2024. New grads now account for just 7% of tech hires, down more than 50% from pre-pandemic levels. At startups, the share dropped from 30% in 2019 to under 6% in 2024.
Those aren’t trends. Those are verdicts.
And the money? Let’s talk about the money.
Master plumbers with commercial experience and their own businesses routinely clear $150,000 or more. Industrial electricians specializing in PLCs or high-voltage systems command premium wages well into six figures. HVAC specialists in commercial refrigeration and building automation earn $90,000–$120,000. Construction managers overseeing multi-site builds pull $180,000 or more. Even general electricians hit a median of $60,000–$80,000, with top earners exceeding $104,000.
All of this without a four-year degree, without $80,000 in student debt, and without the existential dread of wondering whether a language model just made your entire skill set redundant overnight.
The Uncomfortable Logic
Here’s the argument distilled to its sharpest edge:
Software development was valuable because it was hard. Writing code required years of training, deep knowledge of systems, and a tolerance for abstraction that most humans don’t possess. That difficulty was the moat. It justified the salaries. It created the scarcity.
AI has filled in the moat.
When Jensen Huang says “everybody in the world is now a programmer,” he’s not being hyperbolic. He’s describing the economic destruction of a scarcity premium. If anyone can prompt an AI to build an app, then the value doesn’t live in building the app. It lives somewhere else.
Trade skills are valuable because they are physical. You cannot outsource a burst pipe to Bangalore. You cannot prompt-engineer a 400-amp service panel into a data center. An LLM cannot crawl under a house to replaster a sewer line. The moat around trade skills isn’t knowledge — it’s embodiment. The fact that you need a human body, in a specific location, with trained hands, executing a task that has real-world consequences if done poorly.
That moat isn’t getting filled anytime soon. Robots can’t navigate the chaos of a residential renovation or the unpredictability of a commercial build site. And the demographic math is brutal: the average skilled tradesperson is in their late 40s, baby boomers are retiring in waves, and a generation was told that trade school was the consolation prize for people who couldn’t hack a computer science degree.
That cultural snobbery is now the single greatest career opportunity in America. The gap between perception and reality is where fortunes are made.
But I Love Coding!
Great. Nobody’s telling you to stop. But love is not a business plan.
The developers who will thrive are the ones who accept that their role is fundamentally changing — from writing software to directing software-writing machines. That’s a smaller, more elite group. It requires taste, judgment, systems thinking, and domain expertise, not the ability to remember the syntax for a SQL join.
For the rest — the mid-level devs cranking out CRUD apps, the bootcamp grads chasing their first React role, the offshore teams building features to spec — the market signal is clear. The demand is contracting. The supply of AI-generated code is infinite. And the price of infinite supply is, historically, zero.
Meanwhile, a 23-year-old electrician in North Carolina named Jacob Palmer skipped college, completed an apprenticeship, launched his own business, grossed $90,000 by age 21, and hit six figures this year. His summary of the situation: “I don’t owe anybody anything.”
No student debt. No layoff anxiety. No existential crisis about whether GPT-7 is about to eat his lunch.
The Punchline
Three of the most influential technologists alive — the man who builds the GPUs, the man who builds the models, and the man who builds the products — are all saying some version of the same thing: the age of the human coder, as a mass profession, is ending.
Jensen Huang wants you to pick up a wrench. Dario Amodei gives you 6–12 months. Sam Altman thanks you for your service.
The world needs 500,000 skilled tradespeople yesterday. It needs 81,000 new electricians every year. Data centers worth trillions of dollars are being built as you read this, and nobody can find enough people to wire them.
The irony is almost too perfect: the machines that are replacing software developers need physical buildings, physical wiring, physical plumbing, and physical humans to construct them.
The AI revolution doesn’t run on code alone. It runs on copper, concrete, and conduit.
Maybe it’s time to learn which end of a wrench is which.
The author accepts no responsibility for any developer who reads this article, enrolls in a four-year electrical apprenticeship, and ends up making more money than their former tech lead. That’s between you and your W-2.