Dario Amodei has been CEO of Anthropic throughout, and is active and opinionated on these issues. I assume Dario and his team have changed approach. One of my readers just commented on LinkedIn that the Anthropic team now believe that they better understand how the model will behave under general principles rather than specific values-based rules. That’s very interesting, and defensible, but I worry about unintended / misaligned consequences of putting the model in charge.
Sigh. Did anyone change in Anthropic's leadership team that could explain this change of heart?
Dario Amodei has been CEO of Anthropic throughout, and is active and opinionated on these issues. I assume Dario and his team have changed approach. One of my readers just commented on LinkedIn that the Anthropic team now believe that they better understand how the model will behave under general principles rather than specific values-based rules. That’s very interesting, and defensible, but I worry about unintended / misaligned consequences of putting the model in charge.