My main problem here is that he gives a horrible variation for white, including a flat out blunder (Nh5????), as justification for his choice. e5 is the 'obvious' response to Nc5 and it's completely unclear how black has accomplished anything as the variations after d6 are quite complex, but it leads to alot of positions where black has an isolated pawn and a knight versus bishop endgame.
Whereas the complications after e5 Nd3 Nc5 are also very interesting and far from clear, and labeling 1. .. e5 with a ? just reeks of laziness. Although I will at least agree that playing e5 with the plan of following it up with f6 seems like a very superficial idea.
And yeah, I think computers are pretty useless in positions like these unless you give them massive amounts of time per move. Their main use is just a blunder checker especially when you have a habit of playing provocative moves.
But when their evaluations are within tenths of a point of each other in a strategic position, it means nothing.