Dear AI Companies: Stop the “Doom Trolling” | AI Reality Check

June 25, 2026

AI Summary

5 min read

In March 2023, just a few months after ChatGPT launched, Yuval Harari, Tristan Harris, and Aza Raskin published a New York Times op-ed warning that AI had “seized the master key to civilization.” They predicted that by 2028, the U.S. presidential race might no longer be run by humans. Their conclusion: “We have summoned an alien intelligence.” Cal Newport, a computer scientist and technology commentator, remembers reading that and thinking it was wildly out of proportion with how large language models actually work. He wrote a New Yorker piece pushing back, but it was too late. The AI companies themselves had already embraced the strategy of terrifying their own customers.

The strategy of “doom trolling”

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

Never miss an episode of Deep Questions with Cal Newport

Get every new episode summarized in your inbox — free, ~5 minutes to read.

No spam. Unsubscribe anytime.

What you'll learn

1 (00:00) **Cal’s Initial Excitement & the Turn to Dread** - Cal describes his initial enthusiasm for ChatGPT’s capabilities, comparing it to the iPhone’s pinch-to-zoom, and his frustration when the discourse immediately shifted to fear and hype.
2 (00:45) **The Early Doom Narrative & Cal’s Response** - Cal recalls the alarming New York Times op-ed by Harari, Harris, and Raskin (calling AI an “alien intelligence”), his own New Yorker piece pushing back, and his attempts to explain AI to senators.
3 (02:07) **The Pivot: AI Companies Embrace Doom Trolling** - Cal argues the real problem began when AI companies themselves started scaring their customers, citing Sam Altman’s “risk of extinction” letter and Dario Amodei’s 25% chance of “really, really badly” outcomes.
4 (02:46) **Specific Doom Predictions from CEOs** - Cal details Amodei’s claim that 50% of entry-level white-collar jobs will be automated in 1-5 years, and Altman’s funding of UBI experiments based on fears of economic collapse.
5 (03:26) **The Anthropic “Mythos” Case Study** - Cal recounts Anthropic’s report claiming their new model was too dangerous to release, the resulting “terrifying” media coverage, and how they quietly released it six weeks later, suggesting it was a publicity stunt.
6 (04:14) **The Latest Anthropic Report & The “China” Excuse** - Cal describes Anthropic’s report on AI recursively improving itself, complete with scary animations, and their conclusion that there’s nothing they can do about it because of China.
7 (04:52) **The Human Cost of Doom Trolling** - Cal reads a heartbreaking email from a software developer whose mental health has been ruined by constant predictions of replacement, comparing the anxiety to the COVID-19 pandemic.