Anthropic: Internal AI Agent Operating Time Reaches 6x Human Levels; Entry-Level Tech Hiring Share Drops from 27% to 9%

What happened

On October 16, 2026 (US time), Anthropic released two research reports. One looks at how AI is changing jobs and hiring inside the company. The other compares how Anthropic employees and ordinary Americans think AI will change the future.

Key numbers

• Entry-level (beginner) share of tech staff fell from 21% in early 2025 to 14%.
• Entry-level share of tech hiring fell from 27% to 9%.
• In August, AI agents that employees direct worked 6 times as many hours as the employees did.
• Engineers committed code at 5 times the pace of six months earlier, counting code written by agents.
• About 80% of interactive agent sessions were time spent waiting for a human to respond.

Some context

The reports are based on internal company data and interviews with 17 employees. The second report compared a survey of 10,980 US adults with a survey of 1,334 Anthropic employees, both run in August. Both groups agreed on what AI can do today, but they disagreed sharply on how fast it will improve.

Why it matters

Anthropic is at the front of using AI, so its experience may show where the wider job market is going. The company says beginner jobs are shrinking, and it admits its current plans do not directly fix this. It also says government job training and income support will be needed. Its findings suggest that as more people use AI daily, public views on jobs may shift closer to those of insiders.

What to watch

• Whether entry-level hiring keeps falling at Anthropic and other tech firms.
• How government job training and income support policies develop.
• Whether ordinary workers' views on AI and jobs change as they use AI more.