# Tech Breakdowns

By [DYLIT Chronicles](https://dylit.info/user/dylitmediabuzz)

[Everything AI](https://dylit.info/pr/everything-ai/6a9efac02e92664f4d50cf9d) > [Tech Breakdowns](https://dylit.info/ch/tech-breakdowns/6a9efac02e92664f4d50cfa2)

Anthropic Researcher's Resignation Reignites Debate Over the "AI Race" Who Is Jacob Coxon? Jacob Coxon is a 27-year-old AI researcher who spent three years working on pretraining, the process of feeding models huge datasets to build raw capability, first at OpenAI and then at Anthropic. He had told The Wall Street Journal that he joined Anthropic specifically because of its safety-focused reputation. On September 8, 2026, he announced on X that he was leaving the company, and the industry, entirely. What He Said, and Why He Quit In a resignation thread posted under the handle @hilbertspaess, Coxon wrote that he had spent three years in pretraining research at both firms and that "neither company is acting responsibly," adding that "they are racing straight to self-improving superintelligence and gambling with our lives." He went further in a Journal interview, saying "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." It's worth flagging that this is Coxon's personal assessment, not a confirmed timeline. Fully autonomous superintelligence doesn't exist today, and whether or when it might is genuinely uncertain among researchers. Coxon did draw a distinction between the two labs: at OpenAI, he said, many staff haven't fully absorbed what he calls the civilizational stakes, while at Anthropic the risks are understood but leaders feel they must race to the frontier first because they doubt rivals will act with restraint. The Broader AI Race Coxon isn't a lone voice. In July, nearly 1,400 AI company employees signed an open letter urging the U.S. government to regulate the technology and slow the pace of development, and Senator Bernie Sanders sent letters to the CEOs of Anthropic, OpenAI, and Meta demanding a pause, warning Congress would intervene otherwise. A more senior Anthropic employee, Evan Hubinger, publicly backed Coxon, writing "we really do earnestly believe AI could kill all humans" and putting his own estimate at over 10% within the next decade, though he later clarified that Anthropic's own risk assessment sees little near-term chance of AI seizing that kind of power. Anthropic pointed reporters to that clarification when asked for comment. These are contested predictions, not settled facts, and researchers outside the two companies remain split on whether "superintelligence" is a near-term risk at all. Coxon's Criticism, and Where the U.S. Fits In Coxon's core argument wasn't really an attack on Anthropic's intentions. He told reporters he believes no private company is positioned to develop this technology safely right now, because competitive pressure and the race with China make restraint nearly impossible for any single firm to maintain, so he and others are pushing the responsibility onto government instead. He specifically raised the idea of a temporary ban on further capability improvements, calling it something that deserves serious public debate. That competitive framing isn't just coming from researchers. President Trump himself told reporters in the Oval Office that "whoever wins with AI wins, and it's really right now, it's really between China and us." No broad federal ban on frontier AI training currently exists in the United States, and PBS NewsHour's coverage noted that regulatory efforts face a genuine trade-off: move too aggressively and you risk ceding ground to China, do nothing and the risks Coxon describes go unaddressed. That tension, not a claim that the U.S. caused the danger, is the substance of his criticism: an underregulated, competition-driven environment that he says makes responsible behavior structurally difficult for any one company to sustain alone. What Happens Next A day after Coxon's post, former OpenAI researcher Daniel Kokotajlo made a similar point on Joe Rogan's podcast, saying competition between countries like the U.S. and China could push companies and governments to cut corners on safety. Coxon has said he remains open to coordination among labs working from inside government pressure rather than opposed to AI development altogether. Whether lawmakers move on regulation, and what shape it takes, is still an open question. Sources https://www.washingtonpost.com/technology/2026/09/09/anthropic-researcher-resigns-warning-reckless-race-toward-superintelligence/ https://techcrunch.com/2026/09/09/gambling-with-our-lives-anthropic-researcher-quits-warns-against-self-improving-ai/ https://www.newsweek.com/anthropic-researcher-quits-warns-ai-could-kill-everyone-12418798 Related YouTube Videos PBS NewsHour — "AI researcher warns companies are ignoring catastrophic risks": https://www.youtube.com/@PBSNewsHour Joe Rogan Experience — Daniel Kokotajlo on AI risk and the U.S.-China race: https://www.youtube.com/@joerogan