Skip to content
Skip to content

Independent e-magazine

the OUTSPOKEN digest

Anthropic's Own Staff Data: Junior Technical Hires Fell From 27 to 9 Percent as Its AI Agents Outworked Humans

A report from inside the AI lab says its agents now work six times the hours of the people directing them, and the company hires fewer beginners, with caveats it spells out itself.

Outspoken Digest Technology Desk

Sunday, October 18, 2026/2 min read

An outdoor breakout terrace with tables and planting at an office building on Main Street in the South of Market district of San Francisco, photographed on 12 December 2018, a file photograph of a tech-district office and not of Anthropic
Photo: Daniel Prostak via Wikimedia Commons (CC BY-SA 4.0)

Executives have been predicting for a year that AI will squeeze entry-level jobs. This week one AI company published its own numbers, and they are stark. In a report released on 16 October, "How employment is changing at Anthropic", the lab says junior staff made up 27 percent of its technical hires and now make up 9 percent.

The authors, Nathan Wilmers, Regina Catipon, Maxim Massenkoff and Peter McCrory, drew on internal employment, usage and survey data plus 17 employee interviews.

What does the Anthropic employment report show?

The junior share of technical employees fell from 21 percent to 14 percent since early 2025. Technical departments' share of total employment slid from 61 to 56 percent over the past year. About a third of machine-learning managers and a fifth of software-engineering managers said they hire less because of AI.

Meanwhile the agents multiplied. Agent hours, autonomous and interactive combined, outnumbered human hours for the first time in March 2026. By August, interactive agents worked six times the hours of the humans directing them. Code commits per engineer reached five times their level of six months earlier, and coding-agent sessions among non-technical staff rose from two a month in August 2025 to 90 in May 2026.

What changes for the people still working there?

The report describes a job of supervising. Roughly 80 percent of interactive Claude session time is idle, waiting for a human. Meta-work, such as building agent tools and monitoring pipelines, rose from 10 to 22 percent of interactive sessions. One interviewee described "an army of Claudes, with every one asking for my attention." Median meeting time is up 20 percent year on year.

Hiring scorecards changed as well: mentions of a candidate's AI tool use roughly tripled for technical roles and quadrupled for non-technical ones since 2024, and managers say they want people with autonomy and judgement.

How much should you read into it?

Less than the headline suggests, by the authors' own account. The report is descriptive and does not establish that AI caused the hiring trend. Tool adoption is not random, quality proxies may miss errors that are harder to spot in AI code, a survey question changed in April 2026, and Anthropic is an unusually AI-intensive company, so its experience may not generalise. It is also the company selling the technology measuring what the technology does.

Still, it matches what leaders have said publicly. At Davos in January, Anthropic's Dario Amodei and Google DeepMind's Demis Hassabis both reported early signs of falling demand for junior roles, and Amodei repeated his warning that entry-level white-collar jobs could be halved within one to five years.

The test is whether other employers publish their own figures. Until they do, this is one lab's mirror, not the market.

Published in The Outspoken Digest

Editorial desk

Outspoken Digest Technology Desk

Software, hardware, artificial intelligence and what they change for everyone else.

Newsletter

The Digest, in your inbox

One edition, sent when it is ready. No noise, and your address is never passed on.

We send a confirmation first. One click to leave, always.

Share this story

the OUTSPOKEN digest

Beyond boundaries. Independent stories on technology, culture, and the trends shaping how we live.