The article argues that an AI-driven collapse in employment is unlikely in the near term, despite prominent industry warnings that automation could rapidly replace large numbers of workers. It matters because evidence so far suggests that companies are deploying AI more slowly and less extensively than its capabilities might imply, while productivity and unemployment figures have not shown the dramatic changes predicted.

Anthropic found no systematic rise in unemployment among workers most exposed to AI since late 2022, and said use of the technology remains far below what is theoretically possible. Claude was reported to cover 33% of computer and mathematics tasks, despite being capable in principle of far more; meanwhile, labour-productivity growth has been slower than during the 1990s IT boom. The article also notes that even Sam Altman has questioned apocalypse predictions, while critics argue that tasks AI cannot reliably perform may remain economically important.