A measurable change within a single company
The share of new junior technical hires at Anthropic fell from 27% to 9% in just over a year. The figure, reported by TechCrunch based on an internal report published by the company, concerns new hires in technical departments: it does not mean that nine out of every hundred employees are junior, nor that the absolute number of hires necessarily declined in the same proportion.
The same analysis indicates that the junior component of the technical workforce fell from 21% to 14% since the start of 2025. Meanwhile, technical departments went from representing 61% to 56% of the total workforce. Anthropic attributes this latter change also to the growth of non-technical business functions, while assisted programming tools are said to allow some employees to perform tasks previously reserved for specialists.
Work does not disappear: tasks and coordination change
The report describes a broader transformation in the distribution of work. In an internal survey conducted in April, 27% of employees said they had become more productive with Claude; 20% reported that their work had shifted from direct execution to managing multiple agents. These are self-reported responses, useful for understanding how workers perceive the change but not equivalent to an independent measurement of productivity.
The company also reports a 20% increase in time spent in meetings per employee compared with the previous year, with a 16% increase for technical staff. The apparent paradox—more automation and more human coordination—is one of the most interesting elements of the report. When different tools produce code, analyses, and documents in parallel, someone must set priorities, verify the results, and resolve conflicts among proposals.
A useful case, not a forecast for the entire market
Anthropic is an unusual company: it develops AI models and has organized many processes around its own products. Its choices may anticipate certain dynamics, but they do not automatically represent traditional companies, small businesses, or sectors with stricter security obligations and procedures. The report does not prove that AI alone caused the decline in junior hires: company growth, team composition, economic conditions, and recruitment policies may all contribute together.
For young professionals, the question mainly concerns the first rung of the career ladder. If basic tasks are automated, companies will have to explain how the craft is learned: through supervision, progressive projects, work review, or new forms of apprenticeship. Eliminating repetitive tasks can free up time, but those same activities have often allowed new hires to gain hands-on experience.
The decisive test will come from data comparable over time: number of hires by level, roles that have actually changed, quality of the work produced, and internal mobility. For now, the report offers a meaningful snapshot of Anthropic and an open question for the sector: who will train the next generation of specialists if entry doors narrow?
