Room two · Case two
2025: the men running the two most valuable AI labs in the world warn, loudly and specifically, that AI will eliminate enormous numbers of jobs. 2026: both companies file for trillion-dollar public offerings. Both warnings soften, almost in step.
Dario Amodei, Anthropic, 2025, including a 60 Minutes interview
AI could eliminate up to 50% of entry-level white-collar jobs within one to five years, potentially raising unemployment to 10 to 20%.
Sam Altman, OpenAI, 2025
A significant number of jobs "will go away" because of AI — framed at the time as a serious societal risk requiring policy attention, not a side effect to manage quietly.
Both men, 2026, walking it back
Altman, on his own earlier warning, reportedly called it "pretty wrong." Amodei's revised position: automation may "expand the work people do" rather than eliminate it. Both shifts land in the months leading up to OpenAI's and Anthropic's own trillion-dollar IPO filings.
Mustafa Suleyman, Microsoft AI — a third voice, contesting, not echoing
AI will not take over white-collar jobs within the next 18 months, though it will automate mundane tasks — a flat contradiction of Amodei's original claim, from inside the same industry.
Does a forecast that changes shape exactly when the financial stakes change reflect authentic, owned revision — or a bet quietly adjusting itself to the market it claims only to describe?