ExLinkedIner (@ExLinkedIner)

An honest question for all. AI pretraining researcher Jacob Coxon resigned from AI startup Anthropic (having previously worked at OpenAI) and issued a public warning about the unchecked pace of AI development. "Gambling with Our Lives": Coxon publicly stated on X (formerly Twitter) that neither OpenAI nor Anthropic is acting responsibly. He argued that both tech giants are racing toward self-improving superintelligence without established safety controls. Existential Threat: He warned that AI could potentially "kill us all by the end of the decade". He emphasized that while executives soften their language publicly, many privately express severe fear about catastrophic risks. The Indian Express+ 1 Commercial Race Dynamic: Coxon noted that while Anthropic understands the existential risks better than most, the company continues to push forward because leadership believes a competitor will reach superintelligence first if they slow down. Following Coxon's announcement, Anthropic safety researcher Evan Hubinger publicly validated aspects of his concerns. Hubinger stated on X that Anthropic does treat existential threat scenarios seriously, estimating a >10% chance within the next decade that runaway superintelligent AI could pose a fatal threat to humanity. He acknowledged that while current risk levels remain low, the industry does not yet have a clear, proven framework to align superintelligence before it emerges. For me note the phrase "within the next decade". What happens after that? And so much more to wonder about.