I'm sure none of us are under any illusions that some form of robot / AI can and most likely will beat a human player at 40k. Hell I can probably throw together something that moves, shoots and assaults at random with a little shakey cup so it can roll it's own dice, and it'll eventually win a game in the technical sense of a victory.
Jin's argument seems to be, and apologies to Jin if I've misconstrued this, is that such a machine isn't really playing.
We've all had those games where everything has gone wrong, and a sizeable chunk of your army has been wiped out by an "overpowered" enemy unit for the fraction of the points cost. The kind of games that make you question the game in general. But we keep coming back, because something about the fluff, the models, or maybe just the thrill of that one time your one rending snap shot immobilized a charging death company dreadnought. Something keeps us invested, and keeps us going.
Artifical Intelligence doesn't have that (yet?), and until it does, in any games it does win it's kind of missed the point of playing.
There are probably people making a similar argument for Go at the moment, and it seems to be one of those perspectives that you either agree or disagree with, rather than disprove.