Entry-Level Dev Hiring Down 20%, but Experienced Developers With AI Skills Are Getting Raises

Two major data releases this month paint a surprisingly specific picture of what AI is doing to the job market — and the answer isn’t “replacing everyone.” It’s more like a sorting machine: punishing one group of workers while richly rewarding another.

At the 2026 Stanford SIEPR Economic Summit, researchers presented data showing entry-level software developer hiring for workers ages 22–26 is down 20% from its late-2022 peak. Call center hiring is down 15%. But here’s the part that doesn’t make headlines: employment for older, more experienced workers in those same AI-exposed occupations has remained stable or grown.

The numbers tell two different stories

A new Dallas Federal Reserve study crystallizes the divide. Since ChatGPT launched in fall 2022, total U.S. employment has grown about 2.5%. But in computer systems design — one of the most AI-exposed industries — employment has dropped 5%. Across the top 10% of AI-exposed industries overall, employment is down 1%.

Now look at wages. Nationwide average weekly pay has climbed 7.5% since fall 2022. In the most AI-exposed industries? Wages grew 8.5%. In computer systems design specifically, pay has risen 16.7%.

Fewer jobs, higher wages. That’s not a contradiction — it’s a skill filter.

The Dallas Fed researchers explain it this way: AI substitutes for entry-level workers who perform “codifiable, book-learned tasks” while complementing experienced workers whose value comes from “tacit knowledge — the judgment and intuition built through years of hands-on experience.”

The pyramid is becoming a diamond

The Stanford panel identified something structural happening in professional fields like law, banking, and medicine. These industries used to look like a pyramid: lots of entry-level roles at the base, fewer positions higher up. AI is reshaping that into a diamond — hollowing out the entry level while expanding mid-career and senior roles that require judgment, client relationships, and domain expertise.

This isn’t theoretical. Entry-level hiring at the 15 biggest tech firms fell 25% from 2023 to 2024, and some data shows a 55% decline since 2019. Junior tech job postings in the U.S. have dropped by as much as 67% in certain categories. The “learning curve” that companies used to invest in is being automated.

The wage premium is massive — if you have the right skills

The World Economic Forum reports that AI-related roles offer a 56% wage premium on average, up from 25% last year. PwC’s Global AI Jobs Barometer found that human-AI collaboration has driven a fourfold increase in productivity growth. And critically, AI skills on a résumé helped offset conventional disadvantages in hiring: older applicants and candidates without advanced degrees saw their prospects improve substantially when they demonstrated AI competence.

The WEF data points to an important nuance: wages are growing twice as fast in AI-exposed industries versus less-exposed ones, but the growth is concentrated among workers who use AI to augment their skills, not among those whose tasks are simply being automated away.

What actually works

Stanford researchers found a clear dividing line: employment is falling among workers who use AI to automate tasks, but growing for those who adopt AI to learn new skills. The difference isn’t between “uses AI” and “doesn’t use AI.” It’s between “uses AI as a shortcut” and “uses AI as a force multiplier.”

For career changers and people early in their professional journey, the message from the data is counterintuitive but clear: the path isn’t to compete with AI on speed and cost. It’s to build domain expertise faster — using AI as a learning accelerator — so you can move past the entry-level roles that are disappearing and into the mid-career roles where AI makes you more valuable, not less.

Jobs requiring physical presence and human judgment — skilled trades, healthcare, and hands-on work — are holding up best. But in knowledge work, the winning strategy is the same across every field: develop the kind of expertise that AI amplifies rather than replaces.