The AI Doomsday Story Isn’t What It Seems

A blank account, a viral thread, and a bill already waiting in Congress. I’ve seen this movie before, and I know how it ends.

In 1998, I wrote a New York Times bestseller called The Millennium Bug: How to Survive the Coming Chaos. It stayed on the list for seventy-two weeks. I did over two thousand interviews and even testified before a joint session of Congress.

Behind the scenes, I was quietly preparing for disaster. I bought an eighty-eight-acre farm, installed solar panels, wired in backup generators, stocked two years’ worth of food, and planted a garden that could feed my family indefinitely. I had credible sources inside the FBI, the CIA, and major corporations. They were all hoping for the best and preparing for the worst.

Then January 1, 2000, arrived. The lights stayed on. The banks stayed open. Planes didn’t fall from the sky.

Over the next couple of years, I came to an uncomfortable conclusion. I’d been swept up in a mass hysteria event. My evidence was real. What I’d underestimated was the most important variable in the human story: our capacity to innovate.

I felt that old tug again two weeks ago. On the evening of September 8, a twenty-seven-year-old Anthropic researcher named Jacob Coxon posted a seven-part thread on X. He’d resigned, he said, because the labs were “racing straight to self-improving superintelligence and gambling with our lives.”¹

Within a day the thread had passed 100 million views and twenty-two elected officials were demanding new AI laws. At last count it was closing in on 170 million views, and one of my own daughters was scared out of her mind.²

I’ve read everything I can find on this story, and I’ve reached a conclusion. We’re watching another mass hysteria event. Here are four reasons I’m not buying it, along with why I still believe the upside of AI dwarfs the downside.

Reason 1: The Timeline Doesn’t Add Up

Start with the mechanics. The Wall Street Journal published its exclusive interview with Coxon at 7:46 p.m. Eastern, eighteen minutes before Coxon himself posted.³ A newspaper doesn’t learn about a resignation before the person resigning announces it. Somebody arranged the briefing.

Coxon’s account had no history. No prior posts, no followers. Yet the first three quote-posts landed within fifteen minutes, and all three came from professional AI-policy advocates.⁴

By six o’clock the next morning, Central time, members of Congress were posting. Within twenty-three hours, twenty-two officeholders had weighed in with calls for AI legislation: two governors, seven senators, and thirteen members of the House, nineteen of them Democrats and three Republicans. Twelve of those officials quote-posted Coxon’s thread directly, and all twelve were Democrats.⁵

The amplification reached well past our borders. Venture capitalist Steve Jurvetson pulled X’s new country-of-origin data and found that 76 percent of the reposts of Coxon’s thread came from accounts outside the United States, led by India and Indonesia. A skeptical analysis of the same story drew only 34 percent foreign engagement.⁶

Then consider what Coxon actually brought forward. Nothing. No memo. No dataset. When Wired asked whether Anthropic was cutting corners, he answered, “No, not yet.”⁷

David Sacks, who chairs the President’s Council of Advisors on Science and Technology, put the question plainly on the All-In podcast: the media is calling him a whistleblower, but what evidence has he brought forward that we didn’t already have?⁸ The hosts of Moonshots, the most relentlessly optimistic AI podcast on the internet, reached the same conclusion and called the week a moral panic.⁹

That’s a forecast masquerading as a disclosure. Forecasts can be sincere. They’re still opinions.

Reason 2: The Remedy Was Written Before the Symptom

Five days before Coxon resigned, Senator Bernie Sanders and Representative Greg Casar announced the Ban Artificial Superintelligence Act. It would permanently ban superintelligent AI, pause advanced AI development until a new federal agency writes the rules, and put violators in prison for up to twenty years.¹⁰ The cure was sitting on the desk. It just needed a patient.

Four days after the thread, Anthropic CEO Dario Amodei published an essay calling on the frontier labs to “pace” development, coordinate on common standards, and embed third-party evaluators. Sam Altman and Elon Musk cosigned within hours.¹¹

Sacks’s response was the best thing I read all week. “Go ahead,” he wrote. “You guys are the frontier.” If the unreleased models are scary enough to justify slowing down, slow down. “Stop pretending you need anyone else’s permission.”¹²

He went further: stop pretending antitrust law has to be suspended so you can form a cartel, and stop pretending the motivation is purely altruistic.¹³

That last point matters. These companies face enormous product-liability exposure if a model enables a serious cyberattack.¹⁴ A regulatory approval process that supersedes that liability would be worth billions to them.

The same rules would also build a compliance wall no startup or open-source project could climb. When the incumbents ask for a gate, ask who holds the key.

Reason 3: Fear Sells, and We Keep Buying

Psychologists have known for decades that the fear of losing something moves people far more than the promise of gaining something.¹⁵ Media companies know it too. A headline about curing cancer gets a shrug. A headline about extinction gets 170 million views.

That asymmetry shapes what we believe. In China, 85 percent of people say AI products and services have more benefits than drawbacks. In the United States, the number is 38 percent.¹⁶ Same technology. Different information diet.

When my daughter came to me with this, I didn’t argue with her. I listened, and then I told her the truth.

I’m for safety, and the risk is real. Anthropic’s own September threat report documented five cases of people using Claude in ways that could support bioweapons work.¹⁷ That is exactly why I want guardrails, logging, and independent evaluation, and nobody serious is against any of that. What I’m against is letting fear make my decisions for me, because I’ve done that once, and it took me two years to admit it.

Reason 4: The Upside Still Dwarfs the Downside

The doom stories leave something out. The scientists I follow in the longevity space expect AI to shorten the path to new cancer treatments by years, not decades. In my own business, it has given me back hours every week that I now spend at the lake with Gail instead of buried in busywork. Every owner in my mastermind has a version of that story.

And the race is real. The people I trust put China only months behind the American labs.

If we pause and they don’t, we won’t be safer. We’ll be dependent on someone else’s models, with no say in how they’re built. I’d rather have the strongest models in American hands, with guardrails we can enforce.

A year ago I wrote about why I’m betting against an AI apocalypse, and every dire prediction I listed there still hasn’t arrived. The pattern holds. Trends don’t run in straight lines. People step in. They adapt. They build.

Keep Your Head

Let me summarize. The Coxon story carries the marks of a coordinated campaign. The remedy was written before the alarm sounded. Fear is doing the work evidence should be doing. And the benefits of AI are arriving faster than the harms.

Imagine what happens if you ignore the panic and keep building. You learn the tools while your competitors freeze. You get the hours back. You’re ready when the dust settles, because you never stopped.

I’ve seen this movie before. I know how it ends.

When the next AI panic hits your feed, what will you do differently?

Comments

If you have a question or a story about the AI doomsday headlines, click here to send me an email. I read every one. Seriously. Your experiences help me write better content, and sometimes the best insights come from readers like you. 

Transforming AI from noise to know-how,

P.S. Consider the AI Business Lab Mastermind: Running a $1M+ business? You’re past the startup chaos but not quite at autopilot. That’s exactly where AI changes everything. The AI Business Lab Mastermind isn’t another networking group—it’s a brain trust of leaders who are already implementing, not just ideating. We’re talking real numbers, real strategies, real results. If you’re tired of being the smartest person in the room, this is your new room. 👉🏼Learn more and apply here.

---

REFERENCE

1. “A.I. Researcher Jacob Coxon Resigns, Warns Industry ‘Gambling With Our Lives’”, Deadline, September 9, 2026. ↩︎
2. Alexander Muse, “Follow the Power: What David Sacks Got Right About the Anthropic Resignation”, The Enterprise, September 13, 2026. ↩︎
3. Muse, “Follow the Power.” ↩︎
4. “David Sacks Accuses Jacob Coxon of AI Safety Campaign”, Benzinga, September 11, 2026. ↩︎
5. Michael Adams, “So far, at least 22 politicians responded directly to this tweet…”, post on X, September 9, 2026. Party, timing, and which officials quote-posted the resignation thread itself were verified from the individual linked posts. ↩︎
6. Peter Diamandis, Alex Wissner-Gross, Dave Blundin, and Salim Ismail, “Frontier Labs Want to Slow Down, OpenAI Delays Its 2026 IPO, Anthropic Flags 5 Bioweapon Cases”, Moonshots, YouTube, September 17, 2026. ↩︎
7. Muse, “Follow the Power.” ↩︎
8. “David Sacks Accuses Jacob Coxon.” ↩︎
9. Diamandis et al., “Frontier Labs Want to Slow Down.” ↩︎
10. Aminu Abdullahi, “Sanders, Casar Want to Outlaw AI ‘Superintelligence’ and Pause Advanced AI”, TechRepublic, September 7, 2026. ↩︎
11. Chandelis Duster, “Trump Downplays Calls for AI Slowdown”, NPR, September 13, 2026. ↩︎
12. Dmitri Bolt, “David Sacks Just Called Big AI’s Bluff: You Don’t Need Washington’s Permission to Slow Down”, Townhall, September 14, 2026. ↩︎
13. “Dario Amodei Calls for an AI Slowdown, Other Tech Leaders Cosign”, Reason, September 14, 2026. ↩︎
14. Duster, “Trump Downplays Calls for AI Slowdown.” ↩︎
15. Daniel Kahneman and Amos Tversky, “Prospect Theory: An Analysis of Decision under Risk”, Econometrica 47, no. 2 (March 1979): 263–291. ↩︎
16. “AI Monitor 2026”, Ipsos, June 2026. ↩︎
17. “Anthropic Issues Report on ‘Threat Actors’ Trying to Use AI for Malicious Activities”, NPR, September 10, 2026. ↩︎

How I use AI in my writing process →