Charles Duhigg, who spent six months inside the worlds of Microsoft and OpenAI, answers questions about the brief ouster of Sam Altman, OpenAI’s C.E.O. and co-founder, which set off a five-day crisis that some people at Microsoft began calling “the Turkey-Shoot Clusterfuck.” You spent a long time reporting one kind of story, and then, at the end, a different, very dramatic story emerged. What does the short Altman saga say about the larger project of developing A.I. for consumers? On the Friday before Thanksgiving, I got a call from someone within Microsoft, who said, “Hey, look at the Internet. They just fired Sam Altman.” So this set off a week-and-a-half sprint to figure out what happened and rewrite the whole story. Most people have thought of OpenAI being in the driver’s seat, and Microsoft being lucky to have allied with them. But I think what this episode shows is that it’s actually the opposite. It’s that OpenAI is lucky that Microsoft allied with them, because although OpenAI is bringing this ferocious and inspiring dedication to developing A.I., it’s really having a partner like Microsoft, which knows how to make stuff actually happen, that is going to bring artificial intelligence into people’s lives in a meaningful way. This brief drama seemed to distill for some people a feeling of chaos or danger that’s been associated with the development of A.I. There was this narrative floating around that the drama at OpenAI was going on because people there thought Altman was moving too fast and taking too many risks. That’s not really right at all. Basically, members of the board just disliked him and felt they were being manipulated by him. And they struck back by firing him. There was no debate over an issue such as safety, in which people were disagreeing with one another. For good and bad, people are still more powerful than the stuff we are creating on machines. There’s a tendency to move to worst-case catastrophes when talking about A.I.—a fear of robots taking over the world and technology escaping our grasp. But the product that they’re now developing with Microsoft is something that helps your word processor. It’s not something that is going to overthrow humanity right now. This was the first crisis for artificial intelligence. The truth of the matter is that OpenAI was thrown into crisis by the most human of problems. And the only way it was solved was through the work of other humans. Support The New Yorker’s award-winning journalism. Subscribe today » |
No comments:
Post a Comment