The Race — Building Democracy in the Age of AI. People race toward a shared finish line.
The Race
← Trends

Supporting case

Anthropic’s employment study finds a smaller junior share alongside continuing headcount growth

What happened

Anthropic published this internal employment study on October 16. Researchers combined staffing and hiring records, AI-use and output measures, surveys and seventeen interviews. From early 2025, junior employees’ share of technical staff fell from 21% to 14%, and their share of technical hires fell from 27% to 9%. Technical headcount nevertheless increased in absolute terms; other functions grew faster.

The study also describes reassignment of work and movement between departments. Its agent-runtime measure was about six times estimated human working hours in August. Runtime is not an equivalent amount of useful human labor. A comparison of 192 engineers adopting agents with 74 nonadopters was observational, and code-output measures do not establish product quality or the overall business value of the work. This is evidence from one unusually AI-intensive employer, published by that employer. Study and methods ↗.

The entry-level hiring pattern is relevant to who acquires implementation knowledge and gains a foothold in technical careers. Internal redeployment also shows why automation need not translate directly into an equivalent reduction in jobs. Neither observation establishes economy-wide displacement.

Dates and implementation stage

Development: 2026-10-16; publication 2026-10-16

Supporting sources

Related trends