The company building some of the world's most-used AI has published a report on what AI did to its own workplace — and the headline is not 'everyone went home early.' Anthropic says its workforce has become significantly less junior in the past year, and its people are spending more time in meetings, TechCrunch reported on October 17.
The numbers are stark. Junior-level employees in technical departments fell from 21 per cent to 14 per cent of that workforce since early 2025, and the junior share of new technical hires fell even more sharply, from 27 per cent to 9 per cent. Technical teams overall shrank from 61 per cent to 56 per cent of the company, which Anthropic attributes both to business functions expanding as it scales and to coding tools letting non-technical employees build software themselves.
Inside the work, the report describes a shift 'from actually doing work to managing Claudes': in an April 2026 survey, 27 per cent of employees said Claude had made them more productive, and 20 per cent said their job had shifted toward supervising agents. One engineer compared it to babysitting — every Claude brings concerns with an equal demand for attention. Agents logged six times their handlers' hours by August, by one account.
The result, Anthropic says, is that 'human attention and judgement' are becoming the bottleneck, as Claude produces a 'deluge' of text that still needs reading and organizing — one reason meeting time per employee rose about 20 per cent year over year.
Anthropic argues its trends could preview how AI shifts hiring and labour demand across the economy, with the obvious caveat: no other employer has moulded its workflows around its own AI quite this thoroughly.
