Anthropic CEO's 'pace the frontier' proposal draws industry backing and Nvidia pushback
Anthropic CEO Dario Amodei has proposed a 'pace the frontier' approach to AI development, relying on independent safety evaluators and coordination among labs in democratic nations. The plan has gained some industry support but also drew criticism from Nvidia's Jensen Huang. TechCrunch's Equity podcast hosts discuss the feasibility of slowing AI and who would enforce such measures.
The proposal arrives just one week after an Anthropic researcher's dire warning about AI risks captured industry attention, giving Amodei's framework added urgency. The plan centers on independent safety evaluators who would assess frontier models, paired with voluntary coordination among AI labs operating within democratic nations. While some industry players have voiced support, Nvidia CEO Jensen Huang has pushed back against the approach. The Equity podcast hosts questioned whether competing labs can reach consensus on what "slowing down" actually means in practice, and which authority would enforce compliance if labs disagree.
This proposal could reshape how AI development is governed, affecting researchers, investors, and the public who rely on increasingly capable systems. If adopted, independent evaluators might create new accountability standards, but voluntary coordination among democratic labs alone may leave gaps, as labs elsewhere could continue unconstrained development. The debate highlights a growing tension between competitive pressure and safety concerns, with society potentially caught between innovation speed and precaution.