A former Google engineer who now leads one of the country’s best-known artificial intelligence safety groups says he has stopped saving for retirement. His reason is that he doesn’t expect humanity to last long enough for him to need the money.
Nate Soares, president of the Machine Intelligence Research Institute (MIRI), made the admission on “The Most Interesting Thing in AI,” a show hosted by journalist Nicholas Thompson. Soares said he still has a 401(k) from his time at Google but hasn’t put money into it in about a decade.
Soares, who also worked as a software engineer at Microsoft, puts the odds that advanced AI causes human extinction within the next 10 years at more than 50%. That number is his own estimate, not an established scientific forecast.
The researcher has spent years sounding the alarm about superintelligent machines. In 2025, he co-wrote the book “If Anyone Builds It, Everyone Dies” with AI researcher Eliezer Yudkowsky. The book argues that building AI smarter than humans could end in catastrophe for the human race.
Soares doesn’t consider that outcome locked in, however. He has argued that the danger can still be avoided if world governments halt the training of general-purpose AI systems and end the global race to build ever more powerful models.
His prediction is far darker than those of many others in the field. Geoffrey Hinton, the computer scientist widely known as the “Godfather of AI,” has estimated the risk of extinction at 10% to 20% over a few decades. Anthropic researcher Evan Hubinger has put the risk above 10% within the next decade.
A survey by the research group AI Impacts found that about half of the AI researchers it polled gave at least a 10% probability to human extinction or severe disempowerment caused by advanced AI.
Not everyone buys the doomsday case. Skeptics point out that many extinction scenarios rest on two unproven assumptions. The first is that AI systems will eventually outwit humans. The second is that their goals will clash with ours. Some critics have also warned that fixating on far-off apocalyptic scenarios makes it easier to overlook the harms the technology is causing right now.
Soares isn’t the only big name in tech to question the point of saving for retirement in the AI era, though others reach that conclusion for a very different reason. Tesla and SpaceX CEO Elon Musk has also cast doubt on retirement savings. Musk’s argument is that AI, robotics and energy could create such abundance that traditional nest eggs become unnecessary. Soares, by contrast, sees a vastly capable AI as an existential threat.
Soares’ remarks have been picked up across the tech and business press, including by Fortune. They come at a time of mounting public debate over how fast AI companies should be allowed to move, and over who should decide.
For now, the MIRI president says he is putting his money where his mouth is.
