The stance held by Google engineers seems to be more subtle than that, although it is fairly clear: a weapon produced by an American company may go either to kill a civilian in a pointless war in a country nobody has ever heard of (high-probability), or to kill a Chinese soldier in a future war for the survival of our country (low-probability). The US government could address this by improving the mix ratio between pointless wars that benefit nobody and wars for the survival of our country, so that a weapon would be more likely to end up used for the latter than the former.