Jacob Coxon

Jacob Coxon is a British AI researcher, aged 27 in September 2026, who worked on model pretraining at OpenAI and then at Anthropic and resigned from Anthropic on September 8, 2026 with a series of posts on X warning about the pace of frontier AI development.¹²

Background

The Washington Post reported that Coxon was about three years into AI research, first at OpenAI and then at Anthropic, which he joined in May 2026 according to his colleague Ethan Perez, who helped recruit him. The same report states that he is originally from Britain, has a mathematics background, competed in an international mathematics competition for secondary school students in 2016 and 2017, and attended University of Cambridge. Coxon told Axios that he gave up his Anthropic equity to leave and still held OpenAI equity.²

Resignation posts

CNBC reported on September 9, 2026 that Coxon quit Anthropic on Tuesday and wrote on X that Anthropic and OpenAI were "gambling with our lives." He wrote that the people building AI "earnestly believe that it could kill us all by the end of the decade," that "Neither company is acting responsibly," and that "They are racing straight to self-improving superintelligence." The post also said "These will soon be superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources." CNBC reported that the post had been viewed more than 70 million times.¹ The Pirate Wires report of September 24, 2026 gives the view count of the thread as 173.8 million.³

The CNBC headline reads "Researcher says AI has more than 10% chance of 'killing all humans.'" In the body of the article the ">10%" figure appears in a post by Evan Hubinger, an alignment lead at Anthropic, who wrote that "we really do earnestly believe AI could kill all humans" and "I personally think it is >10% within the next decade." Hubinger added that he believed "Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to."¹ A spokesperson for Anthropic told the Washington Post that "We have always been transparent that AI will bring both enormous benefits and unprecedented risks."²

Reception

The Washington Post reported that Elon Musk wrote "Seems like a setup" in response to the posts and that Coxon replied with a selfie and the words "I'm real and these are my real beliefs." Senator Bernie Sanders invited colleagues to a private briefing on the dangers of AI the following week, and Representative Lori Trahan wrote that "Safety researchers are resigning, powerful AI models are breaking out of their labs, and companies are racing ahead anyway."¹²

Timing and media coverage

The Washington Post reported that Coxon's resignation was first reported by The Wall Street Journal and that his account on X had little prior public presence. Coxon told the Post that he did not work with any organizations to promote the announcement before posting it, that he spoke with a Journal reporter and then, after posting, "got a group chat of about 10 people to retweet and send to their friends." He said the group included the founder of Encode and that he had also spoken with Daniel Kokotajlo.² StartupHub reported that the Journal's exclusive went up 18 minutes before the thread.⁴

Coxon appeared on Fox News on "Special Report With Bret Baier" on Wednesday, September 9, 2026, where he said "What I see is two things: very intelligent AI systems, and hard proof that we don't know how to control them," and that together these "basically spells disaster, as far as I'm concerned."² The Post Millennial reported that he later said he had "misspoke" when he told Baier he had not worked with third parties, and quoted his statement that the only coordination was to "reach a WSJ journalist via friends, and got a groupchat to retweet my post."⁵

Pirate Wires, in a piece by Hunter Ryerson dated September 24, 2026 and headed "EXCLUSIVE: Jacob Coxon Worked With Doomer PR Firm," stated that Coxon announced his resignation "in an exclusive WSJ interview" and alleged that the firm representing Eliezer Yudkowsky was helping him book interviews.³ The Post Millennial's summary of the report states that the firm, DEY.Ideas+Influence, pitched Coxon for interviews the day after he went public, that Pirate Wires saw an email between Coxon and the firm and cited two people familiar with the matter, and that the firm had also booked an interview for Nate Soares of the Machine Intelligence Research Institute timed to Coxon's departure.⁵ Everything-PR, a public relations trade site, reported that Soares said on X that he had introduced Coxon to the firm after the original post passed 100,000 likes and that Pirate Wires appended a correction noting the ambiguity in the timeline.⁶