Anthropic, the company behind Claude, has published a study of its own workforce. Its finding: as its staff hand more work to AI, the company is hiring far fewer junior engineers and researchers.
What the numbers say
Junior employees made up 21% of Anthropic's technical departments in early 2025. That share is now 14%. The share of new technical hires who are junior fell faster, from 27% to 9%. Job postings for junior roles dropped most sharply after December 2025, the report says.
This is a change in the mix, not a hiring freeze. Anthropic's technical hiring still grew in absolute numbers. By April 2026, its engineering and research teams had moved away from hiring for lower-band technical roles. In an April survey, about a third of machine-learning managers and a fifth of software engineering managers said they were hiring less because of AI.
Agents now outwork the staff
In March 2026, the hours worked by AI agents at Anthropic passed the hours worked by its people for the first time, the report says. By August, agents that employees ran in interactive sessions were logging six times as many hours as the humans directing them. Code commits per engineer reached about five times their level of six months earlier.
The authors add caveats. Agent hours are not the same as six extra workers, since about 80% of interactive session time is spent waiting for a person to respond. Staff did not adopt the tools at random, so the productivity estimates may be biased. And some of the shift reflects ordinary business growth.
Hiring managers now put more weight on autonomy and judgement than on narrow expertise, the report says. It also notes that "a workforce that is anxiously worried is not a workforce doing its best work."
