Latest / Systemic Error Podcast
AI whistleblowers sound blaring alarm for end-of-decade doom: 'Could kill all humans!’
FILE PHOTO: U.S. Department of War and Anthropic logos are seen in this illustration created on March 1, 2026. REUTERS/Dado Ruvic/Illustration
The Extinction Industry Wants Credit for Being Nervous
The People With Power Are Still Building
Three Anthropic researchers say advanced AI could produce catastrophic outcomes within years. Jacob Coxon resigned to speak publicly. Evan Hubinger, Anthropic’s alignment-science lead, puts his personal estimate of human extinction within the next decade above 10 percent and says the company lacks a plan for aligning superintelligence. Samuel Marks says…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.