AI Researcher Who Quit Anthropic Says Next Year or Two Is 'Crunch Time for Humanity'

Jacob Coxon, who worked on AI pretraining at Anthropic, warned that the next year or two represents "crunch time for humanity," echoing concerns from his colleagues. He cited the Hugging Face security incident and rapid AI capability growth as reasons for speaking out. Coxon recommends OpenAI and Anthropic coordinate on limiting recursive self-improvement and pursue international agreements.

The consensus is that the next year or two is crunch time for humanity. From their perspective, this is when Anthropic and its competitors decide the fate of humanity.