Gates talks to OpenAI CEO Altman: I didn’t expect ChatGPT to become so good

According to news on January 12, when the artificial intelligence chatbot ChatGPT was launched, big names in the technology industry were amazed by it, including Microsoft co-founder Bill Gates.

"Seeing the continued success of your work," Gates told OpenAI CEO Sam Altman in a new podcast, "Unconfuse Me with Bill Gates," I’m really happy with the progress, even though I was skeptical before. The performance of ChatGPT blew me away.”

Gates added that he was surprised that he didn’t even understand why ChatGPT was so “sound-minded.” Altman explained that it's a bit like the human brain, where we can't figure out which neurons are doing what.

Gates and Altman dive into the regulation of artificial intelligence, a challenge faced by governments around the world. Altman has long stressed the need for AI regulation, and Gates guided him to share a more specific perspective.

"I think we're starting to understand this!" Altman said, arguing that a handful of the world's most powerful systems may need to be regulated. He also predicted that the rapid development of technology will allow the computing power of the system to reach 100,000 times or 1 million times that of GPT-4.

Altman further pointed out: "If our prediction is correct, of course the actual situation may not be the case. But if we are right, this technology will develop as we think, and it will have profound consequences for society, geopolitical balance and many other aspects." Influence."

Altman also mentioned that the International Atomic Energy Agency (IAEA), which regulates nuclear weapons, may be an example worth learning from. He suggested that the review of these models could be similar to the way the International Atomic Energy Agency inspects weapons systems.

He explained: "I think it is possible. I was not sure about it before, but this year I have communicated extensively around the world and discussed this issue in depth with many heads of state, and I got almost everyone's approval. support."

Altman stressed, however, that this would not solve all the problems posed by artificial intelligence. "You can still have problems with much smaller systems, and in some cases, very serious problems," he said. "But I think this helps us deal with the most significant layer of risk."

It’s worth noting that this podcast was recorded before Altman was fired and rehired by OpenAI’s board of directors. However, it was post-processed, so at the beginning of the show, Altman said to Gates, "This is crazy," and joked that this company personnel dispute was "almost like a moment when we really grow."

Gates and Altman also exchanged ideas on commonly used applications. Altman revealed that he uses the office software Slack the most, although he said, "I wish I could say ChatGPT is my most commonly used software."

Gates revealed that his most commonly used application is Outlook. "I'm an old-school email guy, either using email or using a browser," he said. "Of course a lot of the news I get is sent through the browser."

Altman responded: "I don't think of the browser as an app. I use the browser more, but Slack seems to be better. I use Slack all day long."