Anthropic published a report on Friday on how its own workforce has changed over the past year, and the numbers are less flattering than the framing. The company's argument is that because it has moulded its workflows around AI more thoroughly than anyone else, its internal trends are an early signal for the rest of the economy.
What actually changed
- Junior staff in technical departments fell from 21% of the total to 14% since early 2025.
- Junior technical hires fell harder, from 27% to 9%.
- Technical teams went from 61% of the workforce to 56%. Anthropic credits "natural expansion of business functions as the company scales" and "coding tools increasingly allowing non-technical employees to make software themselves".
- Meeting time per employee is up 20% year over year, and 16% for technical contributors.
In an April 2026 internal survey, 27% of employees said Claude had made them more productive, and 20% said their job had shifted "from actually doing work to managing Claudes". One engineer compared it to babysitting: "When you are babysitting, every child brings to you concerns with an equal demand for your attention and the various [Claudes] also do that."
Anthropic's own conclusion is that "human attention and judgement" are becoming the bottleneck, because Claude produces a deluge of text that somebody has to read and organise.
Why it matters
- The junior hiring number is the one to watch. Every company that says AI made its engineers faster is quietly deciding it needs fewer people at the bottom of the ladder, and the bottom of the ladder is where the next senior engineers come from.
- More meetings is not a rounding error. If output per person rises and so does coordination time, the gain gets spent on talking about the work rather than doing it.
- Anthropic admits "more code does not necessarily mean better code or business impact" but says it sees no sign of quality declining. That is a claim from the company selling the tool, measured on its own terms, with no independent check.
- The report is a useful mirror for anyone rolling out agents now. If the most AI-native company in the industry ends up with fewer juniors, more meetings and a review bottleneck, that is probably where your team lands too.