Anthropic report: junior technical hires fell from 27% to 9%

The share of junior-level employees in technical departments at Anthropic slipped from 21% to 14% since early 2025, and the proportion of junior technical hires fell faster still, from 27% to 9%. The company leads with those figures in a report published by the Anthropic Institute on October 16 and authored by Nathan Wilmers, Regina Catipon, Maxim Massenkoff and Peter McCrory.

By August, the report says, interactive AI agent hours outnumbered employees 6 to 1. In March, the combined autonomous and interactive agent work hours "leapt past human work hours for the first time." "Code commits per engineer have risen the fastest, coming in at 5x their velocity as 6 months ago," the authors write.

The gains spread beyond engineering. Lawyers processed more contracts, accounts payable processed more invoices, salespeople closed more deals. The composition of the work shifted with it. "Meta-work," which the report defines as building visibility tools and monitoring agent pipelines, grew from 10% of the average employee's interactive Claude sessions in fall 2025 to 22% in spring 2026. The median employee spent 20% more time in meetings than a year ago.

One unnamed source inside the company put the shift bluntly: "I think we've already moved past the point of Claude augmenting away the really entry-level admin type role." Another: "that 20% that was strategic becomes 90% of the job."

Hiring managers tripled their discussion of AI tool usage in technical candidate assessments between 2024 and 2026, and quadrupled it for non-technical roles, prioritizing "autonomy and judgment" over "technical domain expertise and implementation." Technical employment at the company fell from 61% to 56% of total staff over one year. Employees with at least a year on the job were 50% more likely to move across departments each month in 2026 than in 2025, after Anthropic stood up what it calls "an early warning system to track jobs at risk of disruption."

The report does not publish absolute headcounts behind the share figures, and the dataset is a single company reporting on itself. Two of the people we track had passed the link around before most feeds caught it.