Software-development job postings in the U.S. have climbed nearly 15% since February 2025, when Anthropic launched Claude Code, even as overall postings fell. On its face, that’s reassuring. Look closer, though: a generation of early-career workers remains on the outside.
Stanford University’s 2026 AI index found that employment in AI-exposed fields among people ages 22 to 25 has fallen 16% since generative AI rose. Among early-career software developers, the drop was 20%. Older developers, meanwhile, saw their roles expand. The ladder’s top rungs are getting sturdier. Its bottom ones are disappearing.
Who Is Benefiting
The rebound isn’t evenly shared. Indeed Hiring Lab data show senior positions accounted for 71% of the increase in software-development postings between May 2025 and May 2026. Job titles mentioning AI made up 37% of that increase. Even after the recovery, postings remain about 27.5% below their pre-pandemic level.
McKinsey Global Institute research says AI wasn’t the sole cause of the earlier slide in developer jobs. A steep fall from a 2022 hiring peak came first, as technology companies pulled back after a pandemic-era hiring surge. Eric Kutcher, North America chair at McKinsey, said on CNBC’s “Squawk on the Street” that “it wasn’t that long ago that we were reading the headline that there will be no more need for software developers.” Developers who work with AI remain in high demand, he said. “We’re seeing much more growth in the places we thought we were gonna see the biggest displacement.”
The Pace of Automation
The tools keep advancing. In 2023, GitHub said AI was close to writing half of all code on projects where it was available. Some AI experts now say AI-written code across GitHub has passed the halfway mark, though GitHub hasn’t confirmed it. In October, the company reported that one in three pull requests involved an AI agent, up from fewer than one in 10 a year earlier. “If that pace holds, within the next two years, most of the code pushed to GitHub could be written by an agent. Much of it may never be fully read by a human,” GitHub said.
That raises a safety question. GitHub warned that humans must keep pace with AI-generated code to monitor its risks. “As the amount of code grows, we must prevent more exposures and reduce the human effort required by those that remain, or else the volume of vulnerabilities introduced will become untenable,” it wrote.
Researchers at Anthropic found that people remain necessary to oversee automation and provide user input. They cautioned, though, that “it is by no means certain that this pattern will persist into the future.”
Protecting the Pipeline
Sneha Puri, an economist at Indeed Hiring Lab, said companies still need developers to oversee AI work, refine products and help clients integrate technology into day-to-day operations. Demand is growing for workers who help businesses apply AI tools, not just build them. For students, she offered practical advice: “Being able to use AI in your work and improve your efficiency is a great place to start. But communication, interpersonal skills and leadership are also really important.”
She also put the onus on employers. “Today’s entry-level talent becomes tomorrow’s senior experts. If employers don’t build those pipelines now, they’re not going to have senior talent they need 10 years down the line.”
The longer-range numbers look steadier. An August report from the U.S. Bureau of Labor Statistics projected software-developer employment would grow 10% from 2025 to 2035, rising from about 1.72 million to 1.89 million jobs and adding roughly 174,700 positions. Economist Steve Rattner’s analysis of BLS data found software developers recorded positive employment growth from May 2023 to May 2025, in line with all occupations and below the pace for traditionally blue-collar jobs tied to building AI data centers.
Why This Matters:
Headline job numbers can hide who’s actually getting hired. Postings are up. Yet 20% fewer early-career developers are working, and senior roles drove most of the gains. If companies cut entry-level hiring while demanding experienced talent, they erode the pipeline that produces that talent in the first place. Puri’s warning about a missing senior bench a decade out points to a collective-action problem no single employer will fix alone. Young workers carry the immediate cost, often with student debt and few cushions. How employers, schools and public workforce programs respond will shape whether AI-driven productivity gains are broadly shared or concentrated among those already established in the field.