SAN FRANCISCO — In a sobering window into Silicon Valley’s deepest existential anxieties, a prominent artificial intelligence researcher has revealed that he abandoned his retirement savings nearly a decade ago, convinced that advanced AI poses a coin-flip certainty of wiping out humanity within the next ten years.

Nate Soares, a former software engineer for Google and Microsoft who now leads the Machine Intelligence Research Institute (MIRI) in Berkeley, California, made the stark declaration during a recent appearance on a podcast hosted by The Atlantic CEO Nicholas Thompson.

According to Soares, the accelerating trajectory of machine intelligence leaves precious little margin for error. “Looking at the current situation, it is very likely that we don't even have ten years left,” Soares warned listeners. “I put the probability of not making it past the next decade at over 50 percent.”

That stark calculus has directly impacted his personal financial planning. Soares disclosed that he has not made a single additional contribution to his employer-sponsored 401(k) retirement account in roughly ten years.

When challenged by Thompson—who pointed out that current AI systems still exhibit notable limitations and flaws—Soares drew a historical parallel to the rise of the automobile. He noted that early skeptics among horse owners likely dismissed cars because they lacked certain equine capabilities, missing the inevitable trajectory of exponential improvement.

Crucially, Soares distanced his warnings from the mainstream economic anxiety surrounding job automation. His fears, he noted, transcend mere labor displacement.

“I am not someone who worries about an employment apocalypse; I am someone who worries about AI killing all humans,” Soares said. He added with grim irony: “If that’s the case, we could easily have full employment right up until the moment we all die.”

While Soares compared the looming shadow of artificial general intelligence to the collective dread of past nuclear standoffs, he stopped short of calling extinction a foregone conclusion. He insists that humanity can still alter course—provided it halts the reckless pursuit of systems designed to emulate or exceed human intelligence, while continuing to channel technology safely toward specific medical and scientific breakthroughs, such as cancer research.

For now, however, the veteran researcher prefers to live in the present, betting his energy on mitigating a catastrophe rather than saving for a retirement he believes may never arrive.

[ⓒ Global Economic Times. 무단전재-재배포 금지]