DavidRooke comments on Disentangling arguments for the importance of AI safety

[DavidRooke]

Yes—we have increasingly powerful utility maximisers already and they are in many applications increasingly dangerous.