Both changing and not changing Bitcoin present equal risk of unknown unknown, it is inherent in the very nature of what that is.
That's an interesting point. I am not sure I agree. Does the risk of "unknown unknowns" not decrease over time? Intuitively, I would think so but maybe I am missing something.
I would think you're right.
Their point seems to be that # softfork UUs = infinity = # nofork UUs = infinity - N, where N is some finite number of risks that time has reduced.
Why would it decrease over time? I could see the amount of unknowns increasing over time just as much as it could go down.
I agree with the idea that you can't not do something because of unknowns. That doesn't mean you have to do it though. You simple take caution because there might be a lot of unknowns. e.g. in the Agile development world if there are a lot of "unknowns" you just increase the story point therefore signaling that the story/task is more difficult and could take longer.
Why would it decrease over time?
I gather because more becomes known about the current bitcoin system overtime.