Exclusive: Even Bill Gates was surprised by ChatGPT
Add Axios as your preferred source to
see more of our stories on Google.
Image courtesy of Gates Notes
When ChatGPT came out, even many people in tech were surprised at just how good it was. Bill Gates confessed to OpenAI CEO Sam Altman that he was part of that group.
Driving the news: "I was privileged to see your work as it evolved, and I was very skeptical," Gates said in an interview with Altman that airs today as part of Gates' podcast "Unconfuse Me with Bill Gates."
- "I didn't expect ChatGPT to get so good," Gates says during the half-hour podcast, shared first with Axios. "It blows my mind."
Gates added that he is particularly surprised given that we don't even really understand why it chooses what to "say." Altman noted that's somewhat similar to the human brain, where we don't know exactly which neurons are doing what.
Zoom in: Gates and Altman spoke in detail about how AI should be regulated, a question that governments around the world are now wrestling with. Altman has long talked about the need for AI regulation, but Gates pressed him on just what he wants.
- "I think we're starting to figure that out," Altman said, suggesting that what's likely to be needed is regulation of a handful of the world's most powerful systems. Altman said that he expects the tech to advance rapidly toward systems with 100,000 or a million times the compute power of GPT-4.
- "If we are right, and we may turn out not to be, but if we are right, and this technology goes as far as we think it's going to go, it will impact society, geopolitical balance of power, so many things," Altman said.
The OpenAI CEO pointed to the International Atomic Energy Agency, which regulates nuclear weapons, as a potential model and suggested that inspections of the models might resemble the way the IAEA inspects weapons systems.
- "That feels possible to me," he said. "I wasn't that sure before, but I did a big trip around the world this year, and talked to heads of state in many of the countries that would need to participate in this, and there
was almost universal support for it."
Yes, but: He said that won't fix all of AI's problems. "There are still going to be things that are going to go wrong with much smaller-scale systems, in some cases, probably pretty badly wrong. But I think that can help us with the biggest tier of risks."
Of note: The podcast was taped before Altman's ouster and rehiring, though the episode begins with an update from Altman acknowledging to Gates that "It's been so crazy" while noting the corporate drama "was like a real moment of growing up for us."
Between the lines: Gates and Altman also compared notes on music and what apps they use the most.
- Altman said that he uses Slack the most. "I wish I could say ChatGPT," he added.
- Gates said his most-used app is Outlook. "I'm this old-style e-mail guy, either that or the browser, because, of course, a lot of my news is coming through the browser."
- "I didn't quite count the browser as an app," Altman said. "It's possible I use it more, but I still would bet Slack. I'm on Slack all day."
Go deeper: Listen to the full interview.
