Anthropic's own study: ai already eyes 80% of jobs, starting with the youngest
The company that sells Claude just published the chart every intern fears: a heat-map showing which desks turn blue next. Anthropic’s Labor Market Impacts of AI slices 1.2 million real-world tasks and finds that today’s models can already swallow at least part of four out of five occupations. The red zone—what bots actually do inside offices—still looks small, but the gap is closing at roughly one vertical spreadsheet a week.

The 22-to-25 trap
Hiring rates for that age bracket have plunged 15 % since 2022, precisely where corporate AI budgets spiked. Stanford economists who co-authored the paper call it “eroding the bottom rung”; without entry-level grunt work, the escalator to seniority jams. Meanwhile CFOs at Google, Amazon and Meta quietly file 10 % productivity gains—translation: six employees now ship what ten did last year, and the other four never got their badge.
Blue-collar bodies and preschool hugs still survive; algorithms remain clumsy with torque and toddlers. Yet any job that lives inside a browser tab—paralegal, media buyer, junior analyst—has a neon target on its back. The report coins a sterile phrase, “the implementation gap,” but the English is simple: software is only waiting for legal clearance and a cheaper API key.
Sam Altman admitted as much in Davos: “a painful labor-market reset is coming.” He didn’t add that his own products write the redundancy memos. Anthropic tries to soften the blow, claiming new roles will bloom for those who “master” AI. Translation: learn prompt engineering faster than your landlord cashes the rent check, or the next Claude will write your farewell email—signatures included.
