They know they'll be helping Facebook rape the world's privacy in currently unimaginable ways, but they've got a world-class team doing some serious fucking software engineering, so they don't really care.
Much like for a "quant" on Wall Street, ethical concerns take a back-seat to getting to grapple with lovely, sciency challenges.
Let's take a look:
> I think it is going to be positive, but clearly many disagree.
Positive for you, or positive for the world's sheeple?
> Much of the ranting has been emotional or tribal
Ah, those silly emotional people, group-thinking and being all irrational (unlike Carmack about his technical work). What's there to be emotional about? It's not like there's anything wrong with a police state (like the US or UK) knowing everything about you!
> but I am interested in reading coherent viewpoints about objective outcomes.
He wants us to make a compelling case for how the state knowing absolutely everything about you could possibly be a bad thing. But it has to be objective - you know, exactly like his own view on the matter.
> What are the hazards?
How could Carmack possibly have a clue about the hazards, what with Snowden's revelations and all? It's like, impossible.
> What should be done to guard against them?
How about making sure governments all around the world don't abuse the information they gather on everyone? -Oh wait, you can't.
> What are the tests for failure?
Oh I don't know. Find yourself in a gulag for thinking/saying/doing things the government finds vaguely uncomfortable? --> "Ohhhhhhhh, right. This is a bit of a problem."