It seems to me that the humane approach is endorsing a tactic of avoiding assessing everything in terms of the endgame; think of Tit-For-Tat, whose tactic completely ignores the opponent's predicted response. Of course, you need to think tactically in order to choose a tactic, but it could be counterproductive to continue taking it into account, perhaps due to limited time, computational resources, or something more arcane like Omega reading your mind.

Or, of course, because you run on hostile hardware.

(I do not necessarily endorse the "humane" position in that fictional debate.)