Comment on Anthropic admits AI 'not perfectly aligned' with human valuesparentComments−fittingoppositeOP10dI guess the problem is that we humans can believe in contradicting values at the same time.
Comments
I guess the problem is that we humans can believe in contradicting values at the same time.