A researcher who helped train new A.I. models at Anthropic has resigned, warning that the industry is “gambling with our lives.”
Jacob Coxon, 27, posted a seven-part message on X yesterday that drew nearly 76 million views overnight. The post followed a Wall Street Journal interview with the now-former Anthropic employee.
“I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives,” Coxon wrote.
He added: “Do not underestimate the power of this technology. These will soon be superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources. We have all witnessed the progress in each of these domains, and progress is not slowing.”
Coxon’s Warning on X
Coxon’s third post in the thread is the most direct. “The people building AI earnestly believe that it could kill us all by the end of the decade,” he wrote.
“This is not a marketing stunt. If anything, many executives and senior researchers will couch their phrasing in the press to sound sensible – but I hear the same people express fear privately. No other human activity poses this level of danger.”
In the Journal interview, Coxon said the timeline for serious trouble is short. “We’re on track for a lot of the most aggressive of these scenarios where by the end of next year things could be out of control already,” he told the paper.
Coxon worked at OpenAI before joining Anthropic earlier this year. He said competition between the two companies, as well as with Chinese rivals, makes safety trade-offs inevitable.
The Journal reports that Coxon says his colleagues use the terms “crunchtime” and “endgame” to describe the state of fast-tracked, self-improving A.I. models.
Why Researchers Keep Building
Coxon’s X posts address a question many people ask: if the people building this believe it could kill everyone, why do they keep building it?
“A common response is ‘if they truly believe this, why are they still building it?’ At OpenAI, many have not deeply internalized the civilizational stakes,” he wrote.
“At Anthropic, the stakes are well-understood, but they are locked in a race to get there first – they believe no one else will act responsibly, so they must do it themselves, despite the risk.”
Coxon is not alone in raising the alarm. The Journal quotes OpenAI CEO Sam Altman, who recently said at a summit in North Carolina: “I think some things are going to go very wrong with cybersecurity unless people act quite urgently.”
| Company | Coxon’s Role | What Coxon Said |
|---|---|---|
| OpenAI | Pretraining research | “Many have not deeply internalized the civilizational stakes” |
| Anthropic | Pretraining research | “The stakes are well-understood, but they are locked in a race to get there first” |
The Race to Superintelligence
Coxon describes an industry moving at a pace that leaves little room for caution. He says the systems being built will soon be capable of hacking anything and acquiring real power and resources.
He frames the current moment as a competition where safety loses out to speed. Coxon said competition between OpenAI and Anthropic, as well as with Chinese rivals, makes safety trade-offs inevitable.
His posts end on a note of conditional hope. Coxon says he is “optimistic about the potential for coordination” and urges other researchers to speak out.
“Warning shots like the Hugging Face attack have made pacing agreements between U.S. labs more viable,” he wrote.
“I don’t feel like we’re on track to prevent a global race, which may require costly actions such as a temporary ban on improving model capabilities.”
A Call to Other Researchers
Coxon closes his thread by addressing the people in the same position he was in. He asks them to think about what the next few years will feel like.
“If you are a lab researcher, I urge you to consider what the next few years will actually feel like. Do you want to kick off a superintelligent RL run without a rigorous understanding of its mind? Should you put your head down because ‘it’s happening anyway’ – or take this moment to call for different conditions?”
| Coxon’s X Post | Key Warning |
|---|---|
| Post 1 | Resignation announcement; “racing straight to self-improving superintelligence” |
| Post 3 | “The people building AI earnestly believe that it could kill us all by the end of the decade” |
| Final post | Call for coordination; “temporary ban on improving model capabilities” |
Coxon’s warning is notable for its specificity about timing and for his position inside both OpenAI and Anthropic. Whether his call for coordination gains traction among other researchers remains to be seen.
Coxon has made his position clear: the industry is moving too fast, and the people building it know the risks.
Source: deadline.com
Get the Notebook.
The day's best stories and every fresh verdict, in plain English, in your inbox by seven. One email a day, no more.