shoyshow

AI researchers and executives are warning that increasingly capable systems could eventually become impossible to control. Former Anthropic and OpenAI safety researcher Jacob Coxon says this concern exists inside the companies developing the technology. Investor Brad Gerstner argues that the warnings are exaggerated and notes that AI companies already invest heavily in safety, cooperate with the federal government and publish reports on misuse. I agree that AI development will not stop—and America cannot simply surrender the race to China. But when insiders acknowledge a credible risk from systems capable of improving themselves and operating in the physical world, we should take that warning seriously without treating any speculative percentage as established fact.