A report published on Friday by Anthropic reveals that the proportion of junior-level employees in its technical departments has fallen from 21% to 14% since early 2025.
The company argues these internal trends indicate how artificial intelligence will reshape hiring and labour demand across the wider economy. Anthropic has restructured its workflows around AI, making its own operations a case study for the sector.
Shifting headcount
The drop in junior staff is accompanied by a reduction in technical roles overall. Technical teams now represent 56% of the workforce, down from 61% a year ago. The company attributes this to business functions expanding naturally as the firm scales and coding tools allowing non-technical staff to build software.
Junior technical hires have seen an even sharper decline, dropping from 27% to 9% in the same period.
Managing the output
One employee described an “eternal pressure at Anthropic to move as fast as possible and use Claude for everything.” An April 2026 survey found 27% of staff felt more productive because of the tool. Another 20% said their job had shifted “from actually doing work to managing Claudes”. One engineer compared this to babysitting, noting that every child, like every version of the AI, brings concerns demanding equal attention.
The report suggests human attention and judgement are becoming a bottleneck. Claude generates a “deluge” of text that requires reading and organisation.
Despite the management overhead, the company claims AI adoption has increased productivity. Metrics show more code per person, more sales closed per representative, and more contracts processed. Anthropic acknowledges that higher volume does not guarantee better quality or business impact. It has not, however, found evidence that code quality is declining.
More time in rooms
Meeting time per employee is up 20% compared to last year. Technical contributors spend 16% more time in meetings. The company cites the importance of “human trust and accountability” as the reason. Staff have repeatedly described “face-to-face coordination as indispensable.”
What it means
For people making things at Anthropic, the shift is clear. The need to supervise AI output is consuming time that was once spent writing code. The increase in meetings suggests teams are relying on direct human contact to maintain trust, even as the volume of digital work grows.
