[whatthedangit]

Would you find it more agreeable for them to dismiss safety entirely?

[wagwang]

I would expect people who doom about AI existential risks to not train cutting edge models and give them agentic freedom.