Report
Anthropic CEO defends AI safety strategy while acknowledging mistakes
Dario Amodei says Anthropic withheld general access to Mythos-class models until June, when stronger cyber safeguards were in place.
TLDR
Anthropic CEO Dario Amodei acknowledges some mistakes raised by a critic but defends the company’s “race to the top” approach to AI safety. He says Anthropic made commercially costly choices, including restrictive biological safeguards and delaying general access to Mythos-class models. He also defends changes to its Responsible Scaling Policy, arguing that its newer focus on transparency and external review has been valuable.
Combined views
1.7M
44 Sources, first seen ago
Comments