Maybe we should reflect on AI.

Are we choosing "(ROI) optimal" paths over "(minimal impact) sufficient paths?

<Usually, there are a lot of "local optimals" (traps or attractors) in any 'web of relations' (or systems).> 

;-)