AI safety’s divide over slowing down versus engineering safeguards

A September 16 post describes a split: Anthropic’s Dario Amodei is calling for slower frontier AI development, while Jensen Huang argues that safety should be treated as an engineering problem rather than a reason to slow down. The commentator says agents are becoming more autonomous and labs are reporting behavior they did not fully anticipate. They remain bullish on AI but argue that moving fast only works if the ability to test, monitor and control these systems keeps pace.