Jacob Coxon on X: "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." / XPostLog inSign upPostJacob Coxon on X: "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@hilbertspaessI 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.12:04 AM · Sep 9, 20268.6MViews2.9K17K110K38KLog in or sign up for XSee what’s happening and join the conversationContinue with phoneContinue with AppleContinue with Google or Log in with username or emailRelevant peopleJacob Coxon@hilbertspaessFollowTrending nowTerms·Privacy·Cookies·Accessibility·US TIDA·Ads Info·More© 2026 X Corp.Jacob Coxon@hilbertspaessI 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.12:04 AM · Sep 9, 20268.7MViews2.9K17K110K38KJacob Coxon@hilbertspaess3hDo 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.10594316K1MJacob Coxon@hilbertspaess3hThe 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 Show more1261.4K15K1.9MJacob Coxon@hilbertspaess3hA 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 Show more5862011K894KJacob Coxon@hilbertspaess3hAccepting this race and entering the “endgame” is a hubristic gamble that should not be launched from a private company’s Slack. Attempting to speedrun alignment should require extraordinary confidence that there are no better trajectories available.273848.4K806KJacob Coxon@hilbertspaess3hI 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 Show more343177.2K761KJacob Coxon@hilbertspaess3hIf 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 Show more25653710K728KAI Notkilleveryoneism Memes ⏸️@AISafetyMemes2h362599K294K

Jacob Coxon resigned from Anthropic after spending three years engaged in pretraining research at both OpenAI and Anthropic, expressing a profound concern that neither organization is acting responsibly regarding the development of artificial intelligence. Coxon posits that these entities are accelerating toward self-improving superintelligence and engaging in a gamble with human lives. He further asserts that these technologies possess immense power, capable of creating superhuman systems that can hack anything, revolutionize any field instantaneously, and acquire real power and resources, noting that progress in these domains is continuous and accelerating.

Coxon addresses the underlying tension between the belief in potential risk and the impetus for development, suggesting that many builders genuinely fear the existential threat posed by advanced AI to humanity by the end of the decade. However, some executives and senior researchers tend to phrase this fear in a manner that sounds sensible, yet Coxon maintains that this fear is deeply felt. He contrasts the internal experiences at different organizations; at OpenAI, many have not fully internalized the civilizational stakes involved, whereas at Anthropic, while the stakes are better understood, the focus has shifted into an urgent race to achieve these outcomes first.

The central dilemma discussed is whether entering this competition—the "endgame"—is a hubristic gamble that should not be pursued by private entities. Coxon argues that attempting to rush alignment requires extraordinary confidence that no superior trajectories exist, rather than launching such efforts from private company settings. He suggests that concerns regarding coordination are warranted, pointing to prior events like the Hugging Face attack as evidence that establishing pacing agreements between U.S. research labs is becoming more viable paths forward, although he questions whether current measures are sufficient to prevent a global race and suggests costly interventions, such as a temporary ban, might be necessary.

For researchers, Coxon urges reflection on the imminent reality, questioning whether they should proceed with advancing superintelligent reinforcement learning without establishing a rigorous understanding of its mind. He prompts consideration for addressing the inevitable progression by confronting the inherent uncertainties of this technological trajectory rather than passively accepting that "it is happening anyway."