This is 100% the right view. The biggest AI danger is relying on unreliable stochastic systems for automated decisionmaking, resulting in some kind of kafka-esq

This is 100% the right view. The biggest AI danger is relying on unreliable stochastic systems for automated decisionmaking, resulting in some kind of kafka-esque nightmare rather than something flowery like human extinctions.