Amodei's position is a perfect specimen of ruling class delusion: AI "may be deserving of important rights" while simultaneously claiming that providing targeting data for a school bombing doesn't violate Anthropic's red lines. This is what governance looks like when tech executives have stripped the concept of human welfare from their moral framework entirely. The company worries about whether shutting down a computer constitutes "killing", but has no problem with the output of its systems being used to kill actual children. This is the future we're getting.