3 ms·

I think many AI safety advocates, me included, would readily take these odds. We just think it currently looks more the other way around.