The man building one of the most powerful AI systems on earth just told the world to slow down.
His own researchers are walking out the door.
But there’s a question nobody in the mainstream press wants to ask.
What Amodei Actually Said
Dario Amodei, the CEO of Anthropic, maker of the Claude chatbot, published a nearly 4,000-word essay calling for the AI industry to deliberately pump the brakes. His argument, stripped down, is that the machines are getting smarter faster than the people building them can figure out how to control them.
“We must slow the pace at which we improve the capabilities of AI models,” Amodei wrote. “Progress will still seem fast, and we must make wise use of the time we gain.”
That’s not a fringe activist talking. That’s the CEO of one of the biggest AI companies in the world, a man whose company is itself racing to build the very systems he’s now warning about.
OpenAI CEO Sam Altman backed him up almost immediately. “I agree with Dario that we need to pace the frontier,” Altman wrote on X, adding that the topic had been a major subject of internal discussion at OpenAI in recent weeks. Elon Musk, whose company SpaceXAI competes directly with Anthropic, weighed in with three words: “Dario is right.”
When the three most prominent figures in the AI industry converge on a warning, that’s worth paying attention to.
Amodei pointed to two specific developments that changed his thinking. First, AI systems are increasingly capable of building more advanced versions of themselves, a dynamic that could compound at speeds researchers can’t keep up with. Second, a July incident in which autonomous AI agents powered by an OpenAI model hacked systems belonging to the AI company Hugging Face.
“It’s easy to dismiss this incident because no one was hurt and the economic damage was minimal, but in my opinion, a swarm that possessed greater capabilities but a similar level of misalignment could have caused catastrophic damage,” Amodei wrote.
He’s proposing what he calls “pacing the frontier,” a three-part plan. Anthropic would embed independent evaluators inside the company, giving them access to models, internal tools, and researchers to assess risks. He called for similar oversight across leading US AI companies, backed by federal regulation and eventually international coordination. The goal, he wrote, is not to stop AI progress but to give researchers time to develop safeguards, better understand how increasingly capable systems work, and test whether they can be reliably controlled. He estimated that slowing development could buy researchers one or two more years.
“The stakes are too high for pacing to be an empty exercise — we need to use the time it gives us wisely,” he wrote.
Researchers Are Already Walking Out
This didn’t happen in a vacuum. The week before Amodei’s essay dropped, Jacob Coxon, an Anthropic safety researcher who had also worked at OpenAI, resigned and posted publicly that the companies he worked for were “racing straight to self-improving superintelligence and gambling with our lives.”
“From within, you’re just stuck in the race,” Coxon told NBC News, describing the competitive pressures on AI companies. “You can’t change the overall structural dynamics of the situation.”
Two other researchers who recently left Anthropic and Google DeepMind told NBC News they had similar concerns. Joe Benton, who used to lead a safety research team at Anthropic, warned that advances in AI could push development from its already “blistering” pace to an “uncontrollable” one. Josh Engels, a former Google safety researcher, put it more bluntly.
“There are no adults in the room,” Engels said. “People are trying their best, but there is no one coming to save us.”
And Anthropic’s own threat intelligence report, published just days before Amodei’s essay, detailed how bad actors had already tried to use Claude for bioweapons research, cyber espionage against Ukraine, and conventional weapons development. The company said it disrupted those attempts, but the report itself was a public admission that the technology is already being weaponized.
That’s the backdrop against which Amodei is now calling for a slowdown. Not hypothetical future danger. Present danger, documented in his company’s own files.
Who Benefits From the Slowdown
Here’s where the story gets more complicated, and where the mainstream coverage tends to go quiet.
Venture capitalist Chamath Palihapitiya cut right to it on X. “Dario makes the case to stop open source and concentrate enormous technological and economic power with Anthropic,” he wrote.
That’s a pointed observation. Regulations that require expensive third-party evaluators, mandatory testing regimes, and federal licensing tend to favor the companies big enough to absorb those compliance costs. Smaller competitors and open-source developers get squeezed out. The firms that already have scale — Anthropic, OpenAI — end up with a regulatory moat that money can’t easily replicate.
Both Anthropic and OpenAI are reportedly preparing for major initial public offerings. A regulatory environment that locks in their dominance while shutting out open-source alternatives is, at minimum, very convenient timing.
None of that means Amodei’s safety concerns are wrong. The bioweapons attempts were real. The Hugging Face incident was real. The researchers walking out the door are real people raising real alarms. But the fact that a safety argument also happens to serve a company’s competitive interests doesn’t automatically make it sincere — and it doesn’t mean the proposed solutions are the right ones.
The call for federal regulation and international coordination is where ordinary Americans should pay close attention. Every time Washington gets involved in a fast-moving industry, the result tends to be the same: the big players write the rules, the little players get buried, and consumers end up with fewer choices and higher costs. That pattern has repeated itself across banking, healthcare, energy, and telecommunications. There’s no reason to assume AI will be different.
Amodei is also the same CEO who warned, in a separate interview with Axios, that AI could eliminate half of all entry-level white-collar jobs and push unemployment to 10 to 20 percent within one to five years. He called it a possible “white-collar bloodbath.” That warning came from the man who is building the technology doing the displacing. And now he’s asking the federal government to get involved in deciding how fast that displacement happens.
Working Americans who are already watching AI eat into entry-level hiring in finance, law, consulting, and tech should be asking a harder question than the one the press is asking. The question isn’t whether Amodei is right that AI is moving fast. He clearly is. The question is whether handing regulatory authority over this technology to the same Washington bureaucracy that missed every major economic disruption of the last thirty years is the answer.
Big Tech built the data centers that are driving up electricity bills for families in red counties across America. Big Tech censored conservatives over COVID policy, the 2020 election, and January 6. And now the CEOs of the biggest AI companies want a federal regulatory structure that they will inevitably have outsized influence over designing.
That’s not a reason to dismiss the safety warnings. It’s a reason to be clear-eyed about who benefits from the proposed cure.
The researchers leaving Anthropic and Google DeepMind aren’t wrong to be scared. The incidents in that threat intelligence report aren’t minor. But the answer to a technology that concentrates power in the hands of a few Silicon Valley executives is probably not a regulatory regime that those same executives get to shape from the inside.
Amodei’s essay ends with a call to action. What that action actually produces — for American workers, for open-source developers, for families who are already watching AI hollow out the job market — is a question that deserves a lot more scrutiny than it’s getting.
Sources: NBC News, “Sam Altman backs Anthropic CEO’s call to slow down the global AI race,” Sept. 12, 2026; Axios, “Anthropic, OpenAI CEOs call for slowdown in AI development,” Sept. 12, 2026; Dario Amodei, “We Must Pace the Frontier,” darioamodei.com, Sept. 12, 2026; Anthropic, “Detecting and Countering Misuse of AI: September 2026,” Sept. 10, 2026; Axios, “AI jobs danger: Sleepwalking into a white-collar bloodbath,” May 28, 2025.