Companies Are Hiring Fewer Developers. That Doesn't Mean What You Think.

Anthropic posted 148 open engineering roles in Q1 2026. At the same time, overall software developer job postings in the U.S. dropped 22% year-over-year according to Indeed's Hiring Lab data.

Both of these facts are true. And both are used to tell completely opposite stories about what AI means for developers.

The common assumption

Developer jobs are either perfectly safe ("AI can't replace real engineering") or about to vanish ("AI writes code now, why would you hire humans?"). Most people hold one of these positions, and the conversation rarely gets past the headline.

What the data actually shows

The drop in developer postings is real, but it's uneven. Indeed's data shows the sharpest declines in roles focused on routine implementation: basic CRUD applications, standard integrations, boilerplate frontend work. These are tasks, not jobs. And they're exactly the tasks where AI coding tools have gotten good enough to reduce headcount needs.

Meanwhile, Anthropic's 148 open roles aren't for the same kind of work. They're hiring systems engineers, infrastructure specialists, and people who can build at the intersection of AI and product. (For a broader look at these emerging positions, see the jobs AI is creating.) Google DeepMind more than doubled its headcount as Google consolidated AI teams under its umbrella. The demand isn't disappearing. It's shifting.

The Bureau of Labor Statistics still projects 17% growth in software developer employment through 2032. But that projection was built before the current generation of coding assistants reached production quality. The gap between the BLS forecast and the Indeed trend is where the real story lives.

The nuance

What's happening isn't replacement. It's compression. A task that used to take a junior developer two days can now be done by a senior developer with AI tools in three hours. That doesn't eliminate the junior developer's job. But it changes what skills that job requires.

Companies like Anthropic are hiring engineers who can work with AI systems, not instead of them. The job posting doesn't say "write code." It says "build systems that work alongside AI." That's a different skill profile, and it's one that most bootcamp curricula and CS programs haven't caught up to yet.

The developers most affected aren't the ones with the most experience. They're the ones whose daily work is most concentrated in high-automation tasks: writing standard CRUD endpoints, converting designs to responsive CSS, writing unit tests for predictable functions. Those tasks aren't gone, but they require fewer humans.

What this means for you

If you're a developer, the question isn't "will AI replace me?" It's "which of my daily tasks has the highest automation exposure, and what am I doing about it?"

Here's a practical check: list the five tasks you spend the most time on each week. For each one, ask yourself whether a current AI coding tool (Copilot, Cursor, Claude) could do 80% of it. If two or more tasks pass that bar, your profile probably looks more like a Transformer than an Innovator, and your next career move should account for that.

The developers who come out ahead are the ones who shift their time toward the tasks AI handles poorly: system design decisions, cross-team coordination, debugging complex distributed systems, and translating messy business requirements into technical architecture. Those tasks show low automation exposure in our scoring model, and they're exactly what companies like Anthropic are hiring for. See how other roles compare in our breakdown by job type.