Talk:Human Compatible

Unwaranted criticism in relation to superintelligence
The description of the book seems a bit misleading in that it suggests that Stuart's arguments depend upon the possibility of either artificial human-like intelligence or superintelligence. However, one of Stuart's main points is that machines with only a little intelligence could still pose a threat to humans. The quoted criticisms are also misleading with regard to this same issue, as well as seeming to misunderstand the notion of "perverse instantiation", which is much more a problem for Weak AI systems than of Strong or Human-like AI. Sifonios (talk) 14:05, 2 February 2023 (UTC)