1 of 1
Story summary
- Jared Kaplan warned that self-training AI could trigger an intelligence explosion.
- He said the decision point could arrive between 2027 and 2030.
- Former Anthropic engineer Jeffrey Ladish cited Kaplan’s warning to U.S. lawmakers.
- Anthropic CEO Dario Amodei urged labs to pace development as AI accelerates.
- A Politico poll found two-thirds of Americans now believe AI could destroy humanity.