The pace of AI model development has become the sector's central safety argument, and at Dreamforce that argument broke open publicly. Nvidia chief executive Jensen Huang diverged from both Anthropic's Dario Amodei and the chief executive of OpenAI on AI safety at the conference, putting the hardware layer and the model layer on opposite sides of the question.
The split surfaced within days of Amodei publishing an essay urging the AI industry to slow the pace of model development. Amodei leads Anthropic, a lab built around AI safety research, and that essay had placed a clear position on record before the Dreamforce session opened.
Huang's divergence from both Amodei and OpenAI's chief executive maps onto where each company operates in the AI stack. Anthropic and OpenAI build and deploy the large models that face users directly and attract regulatory scrutiny when systems fail. Nvidia builds the compute infrastructure those models run on. Amodei's essay put the development-pace question to the full industry. Huang's Dreamforce appearance indicated his answer differs.
The comments from Huang and OpenAI's chief executive arrived at the conference within days of Amodei's essay. That timing made the disagreement between the two camps a public matter rather than an internal one.