AWS is hiring 11,000 juniors. Anthropic says they won't be needed. Here's who's right.

Matt Garman, AWS CEO, said it plainly in an interview: replacing juniors with AI is one of the dumbest ideas he's heard. Same day he announced AWS is hiring 11,000 interns in 2026. Same as last year. After Amazon cut thirty thousand people in twelve months.

Meanwhile Boris Cherny (built Claude Code at Anthropic) and Martin Casado at a16z have been running the same line for months. Programming is dying. AI writes code. No programmers needed. Pack it up.

So someone is lying.

Nobody is lying. Casado and Cherny are looking through one hole, Garman through another. Each sees something real.

What's actually dying

Casado and Cherny track one number: lines of code per day written by a human. That number is collapsing. AI writes code faster and cheaper than a senior. Any junior with Cursor ships in a day what took a senior a week in 2022. Disputing this is pointless.

But they take that one number and extrapolate it to the entire profession. That's the sleight of hand.

Programming was never "writing code." Writing code was one task inside a programmer's job, and yes, it ate most of the workday for a long time. Because the tools were terrible. We had vim, Stack Overflow, and Notepad++. "Understand what the client actually needs," "design the architecture," "make sure it doesn't die in prod" got the scraps.

Now it's reversed. Code is an hour of work. Everything else is everything.

What's actually needed

Garman tracks a different number. AWS doesn't have enough people who can understand a customer's problem, build an app around it, and keep it running under load. Demand for those people grows faster than hiring. Hence 11,000 interns. Not generosity, just trying to keep up.

Look at exactly what he said. Not "we need more Java developers." This:

"Narrow technical skills (like writing Java) are losing value. The focus is shifting to building full applications and solving customer problems."

That's a different profession. Not the one juniors got hired into in 2018. Back then: here's a five-page spec, implement it. Now: here's a live customer with a real problem, figure out what they need, build it, make sure it holds.

AI handles the first. The second is not even close.

Three apps

That same week someone brought me three projects. All three built through Cursor, minimal human involvement. All three work. None of them sell.

A B2B landing page. Eight fields in the capture form. All eight defensible: company, role, team size. The user closes the tab on field three. Just gets tired.

An e-commerce checkout. Logic is clean, tests are green. The button says "Confirm Order" instead of "Pay." 28% of people at that step hit back. They assume another confirmation screen is coming. Of course they do.

A four-step onboarding. Technically flawless. Step two has an empty state that reads "Nothing found." No instruction on what to do next. Users close it.

Each app technically works. Each one needed someone to walk through it as a customer and say: this doesn't sell. Good thing they showed me before launch. These founders genuinely thought they were ready.

People with that eye, taste, product instinct, pattern recognition, are in catastrophic short supply. That's exactly who Garman is hiring at AWS. He just calls them "junior software engineer."

Where the bottleneck moved

Used to be: idea, build, test, ship. The long part was building. Maybe 80% of the time. Seniors got paid well because they could build fast.

Now building takes an hour. Not "elegant distributed architecture," just a normal feature. "Add a feedback form, post to Slack" takes one hour, with tests.

The bottleneck moved to where there basically wasn't one before. The "let's check whether this is actually ready for users" stage.

There's a real difference between two words.

"Works": unit tests pass, page loads, form submits. AI handles this on its own.

"Ready" is something else. Does the user understand what to click. Do they see that something happened. Are we losing 40% of the funnel on step three. Does the email template break in Outlook (it always breaks in Outlook). What happens if the user has a VPN, slow 3G, and iOS 17.4. Does AI know that in Brazil checkout has to accept PIX, not Stripe. It doesn't.

About numbers. In April, Wiz found a vulnerability in Lovable, a $6.6 billion platform, that left every user's source code and credentials publicly accessible for 48 days. Their own audit of 1,645 Lovable projects: 10.3% had user data exposed before the project had its first customer. One in ten. Moltbook, a social network whose founder bragged on every stage that he hadn't written a single line of code, leaked 1.5 million auth tokens three days after launch.

CodeRabbit analyzed 470 pull requests: AI code contains 2.74× more vulnerabilities than human code. In Q1 2026, 91.5% of vibe-coded apps had at least one hole that AI simply invented.

My point isn't "AI writes bad code." It writes fine. Nobody looks at what it wrote through the user's eyes.

Why Anthropic doesn't see this

Casado and Cherny aren't stupid. They aren't villains. They look from their own angle.

They sell a tool. Their business model has a direct financial interest: the more expensive seniors are and the cheaper AI gets, the more revenue they make. That doesn't make their words a lie. It makes their words point hard in one direction.

They also look at the industry from a 20-person startup. And in a 20-person startup, one senior with Cursor really does close out the work of five 2018-era juniors. Not because juniors are useless, but because in a 20-person startup, product decisions get made by the founder. The founder is the one walking through the app as a customer, cutting form fields, vetoing features.

AWS can't clone a founder across 11,000 interns. To ship product, AWS doesn't need "programmers." They need thousands of small founders. They call them "junior software engineer" because HR systems don't have a better term.

What to do about it

If you're a junior now or studying, don't panic. The profession isn't dying, it's changing. If you've been learning "how to write Java," stop. It's a tool, and AI handles it better than you. You can't build a career on that.

What builds a career: taking a customer problem, breaking it down, figuring out what to build, shipping it, and verifying it actually solves what was promised. A completely different job. Universities don't teach it. Learn through AI, but the goal isn't "write code." It's "understand what it wrote, and why."

If you're a founder, two pieces of news. Good: the person who can walk through your product as a customer and say "not ready" is probably already on your team. Bad: you probably don't realize how valuable they are. Don't let them leave. Give them veto on releases. Pay them like a senior engineer.

If you read Casado's tweet and thought "great, no more juniors, I'll use AI instead," you have six months to reconsider. In six months you'll have two seniors and Cursor. The seniors will burn out because all the unglamorous work lands on them. Cursor doesn't burn out. But it also doesn't grow into someone who can run the product on their own in a year.

Programmers aren't going anywhere. What's going away is one narrow type: the ones who wrote code to a spec. The ones who solve a customer's problem with an app are needed more than ever.

Garman gets it and is hiring. Anthropic doesn't see it and is selling seniors on the idea that juniors are obsolete.

I'm betting on Garman.