OpenAI’s New CEO Said Something Terrifying a Few Months Ago

"That should cause you to s*** your pants."

Doom and Gloom

Emmett Shear, the guy currently tasked with running OpenAI amid its leadership tumult, has some pretty bizarre ideas about AI for someone who's now at the helm of the most famous AI company in the world.

In a resurfaced clip from June, the new OpenAI interim CEO's appearance on "The Logan Bartlett Show" admitted that he ascribes to the "AI doomer" argument put forth by computer scientist Eliezer Yudkowsky, who is perhaps the world's foremost pessimist regarding the technology.

"I know Eliezer thinks that we're all doomed for sure," Shear, who also cofounded Twitch, told the podcast host. "I buy his doom argument, I buy the chain [of thought] and the logic."

Yudkowsky's become one of the most famous — or infamous, depending who you ask — AI alarm-ringers in the world over the past decade, and in the year since ChatGPT was unleashed upon the world, he's even called for bombing AI data centers to avoid the doom and death that artificial general intelligence (AGI), or human-level AI, will wreak upon our planet and species.

Shear doesn't quite swallow Yudkowsky's doomerism wholesale, and told Bartlett as much — but he does think there's between a five and 50 percent "probability of doom" for the technology he's now in charge of building.

The venture capitalist host went on to cite the beliefs of AI researcher Paul Christiano, an ex-OpenAI-er who has also made something of a name for himself in the tech doomer cottage industry for his estimated probabilities of AI annihilating humanity and who, as Bartlett pointed out, put that same risk possibility at about 25 to 50 percent.

"It seems like if you talk to most AI researchers," the host continued, "there's some preponderance of people that give some percentage."

"That should cause you to shit your pants," Shear responded — and honestly, he's not wrong.

So Edgy

These newly-resurfaced podcast comments weren't the only time OpenAI's temporary head honcho has expressed his intense concerns about the future of AI — he has straight up called himself a "doomer," after all — and they aren't the only strange things we've heard about him saying since he burst onto everyone's radar this week, either.

Yesterday, Futurism and other outlets reported that in some not-so-old tweets, Shear made some super weird observations about rape fantasies and said that he'd prefer Nazis taking over the world to particularly risky AI — which does, to be honest, seem very much in line with his doomer streak.

So there you have it: the guy who's now in charge of OpenAI during its chaotic post-Sam Altman firing era is an AI doomer who made edgelord tweets within the past year. Do with that information what you will.

More on OpenAI: OpenAI Employees Say Firm's Chief Scientist Has Been Making Strange Spiritual Claims