It’s extremely good that Anthropic has not backed down — Ilya Sutzkever

In a notable show of cross-company solidarity, Ilya Sutskever, the recently departed co-founder and former chief scientist of OpenAI, has publicly commended rival AI lab Anthropic for its steadfast commitment to safety. Sutskever's statement, "It’s extremely good that Anthropic has not backed down," serves as a powerful endorsement from one of the field's most influential architects, coming at a time when commercial and competitive pressures are pushing many firms toward faster, less cautious deployment of powerful AI systems. The comment underscores a deepening philosophical divide within the industry between accelerationist and safety-first camps.

Sutskever's praise implicitly references Anthropic's Constitutional AI framework and its deliberate, measured approach to releasing models like Claude 3.5 Sonnet, which prioritizes alignment and controllability alongside capability. This public support from a former OpenAI leader validates Anthropic's strategy of building explicit safety constraints into its models' core architecture, even if it means potentially slower iteration cycles. The endorsement signals to investors, policymakers, and the broader tech community that rigorous safety practices are not a hindrance but a mark of responsible leadership, potentially influencing norms and regulatory expectations as AI capabilities continue to advance rapidly.