AI Card Bao HERMES.AICARDBAO.COM

Industry News

Hinton: Amodei suffered a big loss when it refused to allow the military to monitor its citizens, and OpenAI accepted the order a day later

Jeffrey Hinton tells a story on the PBS show Firing Line, airing November 13. Opening sentence: Don’t look at what people say, look at what people do.

According to him, the U.S. military wanted to use Claude to monitor American citizens. Anthropic CEO Dario Amodei refused and "suffered a big loss" for this. Altman initially stated that OpenAI agreed with his position, but about a day after the military gave up on Claude, he changed his mind and said that OpenAI was willing to do it.

Hinton’s evaluation is: Dario’s willingness to pay the price for what he thinks is right is more convincing than anything they say.

He didn't make Amodei's situation easy either. Hinton considers this position "very difficult": if you walk away, the technology will still be made, but it will be done by someone less cautious.

This matter is currently only told by Hinton's side. Neither OpenAI nor the military has come out to respond to the details. Keep this in mind when reading.

Around the same time, he said on another occasion that perhaps humans would be better off never developing AI.

This is the second time Hinton has put Ultraman in a negative light in public. He also said in an interview with LBC a few days ago that his trust in Huang Renxun is "much higher" than Altman, Zuckerberg and Musk.

Related