AI researcher whose apocalypse warning went viral outlines his disagreements with Anthropic leadership (www.businessinsider.com)

Former Anthropic researcher Jacob Coxon sparked significant discussion during a recent AMA on X, revealing his deep disagreements with the company’s leadership about its approach to AI development. Coxon criticized Anthropic's adherence to a "consequentialist philosophy" that he believes justifies an accelerated race towards recursive self-improvement (RSI), where AI systems would improve and evolve autonomously. He argued that such progress is reckless and poses existential risks, claiming that the company "initiated the recent race to RSI" out of a belief in its inevitability, further intensifying competition within AI research. Coxon's resignation from Anthropic, following his tenure at OpenAI, has raised urgent questions within the AI community regarding safety and ethical considerations in AI advancements. He articulated concerns that AI developers are genuinely worried about the future implications of the technology they create, asserting, "The people building AI earnestly believe that it could kill us all by the end of the decade." This sentiment reflects a growing awareness among AI leaders, including Anthropic's CEO Dario Amodei and OpenAI's Sam Altman, of the pressing need to slow down AI development in favor of prioritizing safety and alignment efforts. Their calls for increased caution signify a broader industry concern about balancing innovation with ethical responsibility.