Taylor Lorenz’s Power User
Taylor Lorenz’s Power User

Are We All Going to Die!?!?! The Dark Money Behind AI Doomer Panic

September 16, 2026

AI Summary

5 min read

The AI Doomsday Machine: Who Benefits When Everyone Thinks We're About to Die

Last Tuesday, 27-year-old AI researcher Jacob Coxton resigned from Anthropic and posted on X that AI had a 10% chance of killing all humanity. The post has been viewed nearly 200 million times. Lawmakers and major news outlets immediately began calling him a whistleblower—despite the fact that, as Washington Post technology journalist Natasha Tiku explains to Taylor Lorenz, "he didn't whistleblow anything." Coxton had no internal documents, no specific evidence of negligence, and no policy agenda. When Tiku asked him what he hoped to achieve now that he had the world's attention, he said he didn't have a specific policy goal—just that raising concerns might help people think about coordination between countries, or maybe a pause on development.

The Whistleblower Who Didn't Whistleblow

Continue reading the full summary in the app — free to try.

Read Full Summary →

Free • No credit card required

Listen to Audio Summary Open in App

What you'll learn

  • 1 (03:00) **The Jacob Coxton Whistleblower Moment** - How a 27-year-old's 10%-extinction-risk tweet went viral and why it lacks substance
  • 2 (04:24) **What Did Jacob Actually Whistleblow?** - The absence of evidence behind the extinction claim
  • 3 (08:12) **The "Recursive Self-Improvement" Gap** - Why the doomer claim rests on faith, not fact
  • 4 (11:35) **Real vs. Existential Harms** - The concrete dangers being ignored for sci-fi scenarios
  • 5 (16:26) **The Contradiction of the Doomer Whistleblower** - Why they simultaneously want to build the technology
  • 6 (19:39) **How the Narrative Shifted** - The social dynamics that made doomerism the smart, high-status position
  • 7 (21:03) **The Infrastructure of a Cultural Victory** - How hundreds of millions of dollars created a self-reinforcing belief system

+ Full timestamped outline available in the app

Guests on this episode

Taylor Lorenz

Show Notes

Why does everyone suddenly think AI could kill humanity?SUPPORT MY WORK: 

Buy a paid subscription to my newsletter at https://www.usermag.co         

Support my work on Patreon: http://patreon.com/taylorlorenz

In this episode of Power User, I'm joined by tech journalist Natasha Tiku to investigate the viral claim that AI has a 10% chance of causing human extinction. From the halls of Anthropic and OpenAI to the corridors of Washington, D.C., we uncover how the insular, billionaire-funded Effective Altruism (EA) movement transformed sci-fi doomsday theories into mainstream media panic and undemocratic policy proposals.

  

We break down the math behind the so-called "10% extinction risk," expose how dark money and ideological fellowships shaped mainstream AI coverage, and discuss why big tech labs actively benefit when lawmakers regulate frontier models out of fear rather than tackling immediate, real-world harms like wage discrimination, data privacy violations, and corporate consolidation.

 


Taylor Lorenz’s Power User

More from this podcast

Taylor Lorenz’s Power User →