Anthropic published internal research this week that is hard to dismiss. Their own data shows entry-level technical hires dropping from 27% of all technical hires to just 9% since early 2025. That is not noise. That is a real structural shift happening at the company building some of the most capable AI on the planet.
The cause is visible in the numbers. AI agents now work six hours for every one hour worked by a human employee at Anthropic. An agent, in this context, is software that operates autonomously, taking on tasks like writing and reviewing code, fixing bugs, and monitoring systems without constant human direction. The humans supervise, decide, and correct.
The productivity gains extend across every department. Salespeople closed more deals after adopting AI tools. Lawyers processed more contracts. One account executive reported that AI automations replaced roughly 20% of her previous daily tasks and freed her to reach 30% more customers than before. Code commits per engineer are running at five times the pace of six months ago.
What changed is the nature of the bottleneck.
A year ago, execution capacity was the limiting factor. There were only so many hours a developer could code, only so many contracts a lawyer could review. Now the constraint is human attention and judgment. One Anthropic engineer described managing AI agents as facing “an army of Claudes, with every one asking for my attention.” Just reading the output from one employee’s AI sessions would consume over half the work week.
The hiring shift follows directly from this. Anthropic managers now prioritize autonomy and judgment over specific technical skills. A junior candidate who needs steady guidance and focuses on execution? Less relevant when AI handles most of the execution. A mid-level person who can define problems clearly, set direction, and catch when automation produces bad results? Extremely relevant.
This pattern will not stay inside AI companies. It is already spreading to any team that integrates AI into daily workflows. The employees who know how to direct AI tools effectively and review AI output critically are outpacing those who treat AI as a novelty.
For business owners, the practical takeaway is this. The question is no longer “can we afford AI tools?” It is “are we building the workflows and team skills to actually run on them?” The companies doing that work now will have a real structural advantage over those waiting to see how it plays out.
This is not the future. This is already how work operates at the companies paying close attention.
Want to explore how AI automation could benefit your business? Let’s talk.