That’s a compelling—even comforting—thought for many people. “We’re in an era where other paths to material improvement of human lives and our societies seem to have been exhausted,” Vallor says.
Technology once promised a route to a better future: Progress was a ladder that we would climb toward human and social flourishing. “We’ve passed the peak of that,” says Vallor. “I think the one thing that gives many people hope and a return to that kind of optimism about the future is AGI.”
Push this idea to its conclusion and, again, AGI becomes a kind of god—one that can offer relief from earthly suffering, says Vallor.
Kelly Joyce, a sociologist at the University of North Carolina who studies how cultural, political, and economic beliefs shape the way we think about and use technology, sees all these wild predictions about AGI as something more banal: part of a long-term pattern of overpromising from the tech industry. “What’s interesting to me is that we get sucked in every time,” she says. “There is a deep belief that technology is better than human beings.”
Joyce thinks that’s why, when the hype kicks in, people are predisposed to believe it. “It’s a religion,” she says. “We believe in technology. Technology is God. It’s really hard to push back against it. People don’t want to hear it.”
How AGI hijacked an industry
The fantasy of computers that can do almost anything a person can is seductive. But like many pervasive conspiracy theories, it has very real consequences. It has distorted the way we think about the stakes behind the current technology boom (and potential bust). It may have even derailed the industry, sucking resources away from more immediate, more practical application of the technology. More than anything else, it gives us a free pass to be lazy. It fools us into thinking we might be able to avoid the actual hard work needed to solve intractable, world-spanning problems—problems that will require international cooperation and compromise and expensive aid. Why bother with that when we’ll soon have machines to figure it all out for us?
Consider the resources being sunk into this grand project. Just last month, OpenAI and Nvidia announced an up-to-$100 billion partnership that would see the chip giant supply at least 10 gigawatts of ChatGPT’s insatiable demand. That’s higher than nuclear power plant numbers. A bolt of lightning might release that much energy. The flux capacitor inside Dr. Emmett Brown’s DeLorean time machine only required 1.2 gigawatts to send Marty back to the future. And then, only two weeks later, OpenAI announced a second partnership with chipmaker AMD for another six gigawatts of power.
Promoting the Nvidia deal on CNBC, Altman, straight-faced, claimed that without this kind of data center buildout, people would have to choose between a cure for cancer and free education. “No one wants to make that choice,” he said. (Just a few weeks later, he announced that erotic chats would be coming to ChatGPT.)
Add to those costs the loss of investment in more immediate technology that could change lives today and tomorrow and the next day. “To me it’s a huge missed opportunity,” says Lirio’s Symons, “to put all these resources into solving something nebulous when we already know there’s real problems that we could solve.”
But that’s not how the likes of OpenAI needs to operate. “With people throwing so much money at these companies, they don’t have to do that,” Symons says. “If you’ve got hundreds of billions of dollars, you don’t have to focus on a practical, solvable project.”
Despite his steadfast belief that AGI is coming, Krueger also thinks the industry’s single-minded pursuit of it means that potential solutions to real problems, such as better health care, are being ignored. “This AGI stuff—it’s nonsense, it’s a distraction, it’s hype,” he tells me.