I agree with your sentiment, but it's plain ignorance to believe that those "war simulations" or any other software developed by the military is always intended to raise the kill count and not prevent casualties. To make my point, I would bet the invention of C made the military more effective in some fashion or other, but that's only a consequence of solving the more basic problems C was intended to address. I think it's hard to argue that the problems IBM was solving, or those that the military solves, are really any different in that respect. They solve one problem and then the solution can and will be adapted for other purposes, and it's often hard to judge whether that's a net good or not.