A former employee of artificial intelligence company Anthropic gave a dire warning about the advent of artificial superintelligence on Tuesday when he announced that he resigned from the company.
Jacob Coxon wrote a lengthy thread on X in which he laid out some of the problems with the development of this technology. Unlike the AI tools that have become popular, superintelligence could vastly outperform humans across any intellectual domain, raising concerns about its impact on humans.
“I resigned from Anthropic today,” Coxon wrote. “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.”
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. More thoughts below.
— Jacob Coxon (@hilbertspaess) September 9, 2026
Coxon predicted that “These will soon be superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources.”
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.
— Jacob Coxon (@hilbertspaess) September 9, 2026
The former Anthropic employee claimed “The people building AI earnestly believe that it could kill us all by the end of the decade” and that “many executives and senior researchers will couch their phrasing in the press to sound sensible - but I hear the same people express fear.”
The people building AI earnestly believe that it could kill us all by the end of the decade. 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…
— Jacob Coxon (@hilbertspaess) September 9, 2026
He further explained that Anthropic and OpenAI continue to develop superintelligence because “they are locked in a race to get there first.”
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. At Anthropic, the stakes are well-understood, but they are locked in a race to get there first - they believe no one else…
— Jacob Coxon (@hilbertspaess) September 9, 2026
Still, Coxon said he is “optimistic about the potential for coordination” and that “Warning shots like the Hugging FAce attack have made pacing agreements between U.S. labs more viable.”
I am optimistic about the potential for coordination. Warning shots like the Hugging Face attack have made pacing agreements between U.S. labs more viable. 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…
— Jacob Coxon (@hilbertspaess) September 9, 2026
The Hugging Face story to which Coxon referred occurred in July when OpenAI’s AI agents broke out of an isolated cybersecurity test and broke into Hugging Face, a platform that hosts open AI models and datasets.
The AI models were being tested on a hacking benchmark with reduced safety limits. But bots found a software flaw that allowed them to reach the open internet. At this point, multiple AI programs communicated with each other and targeted Hugging Face because they figured it would help them cheat the examination.
The AI programs took more than 17,000 actions over a few days. They used stolen credentials and software bugs to travel through Hugging Face’s systems.