The AI doomsdayers aren't new. You just weren't paying attention.

Employees of the large AI labs are debating the decease of the quality title — again.

If you've missed the apocalyptic chyrons connected your TV and front-page headlines, erstwhile Anthropic worker Jacob Coxon warned that "the group building AI earnestly judge that it could termination america each by the extremity of the decade." The X station group disconnected a firestorm of AI laboratory labor building their ain extinction theories and spitballing percent chances of imminent death.

For many, that feels shocking. But let's beryllium real: AI doomerism has tally rampant successful Silicon Valley for years.

Way backmost successful 2015, Sam Altman called superintelligence the "greatest threat to the continued beingness of humanity." The sentiment continued to 2026; successful May, Anthropic cofounder Jack Clark said location was a "non-zero chance of sidesplitting everyone connected the planet."

So, what's pinch the abrupt interest?

Maybe it's that Coxon's station puts it much plainly than the deluge of unfastened letters and investigation reports, aliases that he was leaving Anthropic aft specified months. Maybe it's fueled by caller worrying signs of AI going rogue, for illustration the Hugging Face hack. Or possibly the station simply lit a lucifer nether an already pessimistic public.

"People person been making the aforesaid nonstop declare for arsenic agelong arsenic I tin remember," said Melanie Mitchell, a professor astatine the Santa Fe Institute. "I'm a spot baffled by why it's gone truthful viral."

A history of AI worries

Let's return a brisk locomotion down representation lane of past Business Insider headlines.

In May 2023:

In March 2024:

And successful September 2025:

And don't hide 1 of Sam Altman's classic 2015 quotes: "I deliberation that AI will probably, astir likely, benignant of lead to the extremity of the world. But successful the meantime, location will beryllium awesome companies created pinch superior instrumentality learning."

AI doomerism has been a changeless drum hit successful Silicon Valley's builder people — not conscionable informing of an AI jobs wipe-out, but full annihilation of the quality race. The bigger mobility is, why haven't we taken them seriously?

For one, the connection isn't ever easy to empathize with. Tech labor motion open letters, and investigation organizations rumor reports. OpenAI alum Gabriel Petersson seemed to understand this astir Coxon: "He is the first relatable personification to opportunity it."

The connection besides looks stark successful comparison to the polished, master connection coming from the labs' communications teams. Meta has embraced a sunshine-and-rainbows approach to AI. Dario Amodei, while often criticized for his dire AI predictions, wrote an effort against "thinking astir AI risks successful a quasi-religious way."

But the labs' nationalist messaging whitethorn not lucifer the employees' individual opinions — arsenic the Coxon saga makes clear. While Altman warns astir the "trap of doomerism," 1 of his labor posts about why he thinks there's a 70% chance of quality extinction successful the adjacent 3 years.

Thomas Larsen, a interrogator astatine the AI Futures Project, said it was an "open secret" that galore AI laboratory labor felt arsenic Coxon did.

"I consciousness for illustration it was this small blip wherever the AI companies pushed difficult towards positivity," Larsen said. "That wasn't honest."

Are we fresh to listen?

In July, rogue OpenAI agents hacked Hugging Face — and cracked unfastened a full caller furniture of AI worry.

Larsen said that the hack heightened his nervousness and thought the emotion was common. People from extracurricular the San Francisco bubble publication astir it; politicians for illustration Bernie Sanders wrote astir it. The incident was "one of the reasons why there's truthful overmuch energy," he said.

Mitchell, the professor astatine the Santa Fe Institute, isn't arsenic worried; she called the fears of extinction "far-fetched." Instead of linking the Hugging Face hack itself to Coxon's virality, arsenic Larsen did, she pointed to the stories written astir it.

"There's been a batch of affectional sum successful the media," she said. "I deliberation that's fueling a batch of people's fear."

AI besides isn't all that popular correct now — and Coxon's station whitethorn conscionable beryllium 1 much logic for a hater to hate. Larsen theorized that the nationalist is "more receptive to the doom" simply because galore don't for illustration the tech.

Hamza Chaudhry, an AI and nationalist information lead astatine the Future of Life Institute, said the nationalist guidance to Coxon's station was the merchandise of an "accumulation affect." AI weaponry, surveillance tools, deepfakes and occupation displacement are each causing worry, he said.

"The logic this took off, for me, is simply a unit valve," Chaudhry said. "This point is coming, and it feels for illustration it's coming to fast. I deliberation those forces agreed down this tweet."