Comment on Anthropic researcher says more than 10% chance AI "could kill all humans"parentComments−madaxe_again2dI think we need to worry about paperclip maximisation as much as we need to worry about “Jon is having a bad day so brews up a novel plague”. There are likely ways it can go wrong that we haven’t even considered, too.
Comments
I think we need to worry about paperclip maximisation as much as we need to worry about “Jon is having a bad day so brews up a novel plague”. There are likely ways it can go wrong that we haven’t even considered, too.