What would be the difference between starting from "scratch", creating a new 'rational' type of spirituality and responding to past spirituality?

Here's my crack at this: I take both sides in this to be arguing that we should pursue something like spirituality. Call it elevation. Adam Frank and timtyler seem to be saying that the most well-developed, existing understanding of elevation comes from religion; the quickest way to secular elevation is by appropriating the good parts of spirituality. Eliezer, perhaps taking a more long-term view, wants to build a much more solid foundation. I think both projects would come up with the same result if they succeed. The big question is which is more likely to be successful and how quickly.

Consider designing a word processor. There is probably code already out there that you can use to achieve your goal, but maybe it's buggy or written in an outdated language. Depending on the exact state of the code, it might be quicker to refactor or it might be quicker to begin from the bottom up. Either way, the end result is going to share some features with the original application.

I don't think it is fair to call a proposal for secular elevation a "war on spirituality" any more than building new software is a war on old applications or general relativity was a war on classical mechanics. This is merely a striving for something better.

I'm afraid you completely lost me in your last paragraph. There is always some probability we are radically wrong about the universe, but what would it even mean for the things you speculate about to be true?