Sam Altman Says OpenAI Paused 1 Model Run as AI Nears AGI
Updated
Updated · Vanity Fair · Oct 5
Sam Altman Says OpenAI Paused 1 Model Run as AI Nears AGI
3 articles · Updated · Vanity Fair · Oct 5
Summary
Altman said OpenAI recently halted one training run to improve alignment and monitorability, arguing the company already slows development when capabilities outpace safeguards.
In the Vanity Fair interview, he said AGI is "close," backed international safety standards and possible U.S.-China verification rules, while insisting technological progress should continue despite serious risks.
OpenAI, he said, would keep its nonprofit control even after a delayed IPO, put mission ahead of shareholders, and accept stock-price volatility rather than compromise on safety decisions.
He also defended OpenAI's handling of mental-health risks and executive departures, while saying AI could help cure diseases within a few years and make society far richer if deployed safely.