In short: it's an attempt to mobilize the "supply side" even more. Let's see if OpenScience.SE has the critical mass to live on its own. If not, moving to Academia.SE is definitely the best fallback option.
Also, sorry for replying with a considerable delay, I wanted to turn that issue over in my head a bit before writing down an answer. Here's my reasoning:
Assertion #1: A Q&A site with Open Science topics is valuable.
While wikis, fora, mailing lists, Google+ communities, Facebook pages, etc. devoted to Open Science all have their merits, an open Q&A site is beneficial as well.
For one thing, SEO: Q&A sites work well with search engines.
When a Google/Bing/Yahoo! user looks for an answer to a particular question,
their query is likely to resemble the corresponding question on a Q&A site (same keywords at least, if not grammar).
Also, a Q&A site focuses on solving particular problems and invites/encourages the general public to interact.
Assertion #2: A Q&A site devoted solely to Open Science will have a more committed "supply side".
Open Science community is more likely to be even more committed to building "their" ("our") site
than to contributing to a broader-scope site like Wikipedia, or Quora, or even Academia.SE.
Note: I truly appreciate the contributions of the community to the sites mentioned above, that's why I wrote "even more".
Also, visitors interested in Open Science will be more likely to explore OpenScience.SE rather than a more general site.
Problem: Lack of critical mass, fragmentation.
I understand the points raised by Piotr, by @vzn in another thread, and others through mailing lists. It is important to have a strong community and avoid fragmentation. There is indeed an overlap with Academia.SE and OpenData.SE. OpenScience.SE will only make sense if it is big enough (in terms of questions+answers, active contributors, traffic). Otherwise, it should merge into something bigger, in my opinion Academia.SE is the best fallback option.
Given the energy and the enthusiasm of the Open Science community ("supply") and increasing popularity of the topic ("demand"), I strongly believe OpenScience.SE has a potential to grow to a size comparable to Academia.SE or OpenData.SE.
Even if the proposal fails at some stage, we will be able to migrate our work so far to Academia.SE,
and perhaps some of the contributors, who wouldn't have responded to a "let's work on the [open-science] tag on Academia.SE" call,
will be hooked on the SE model by now and they will join Academia.SE.
Let's leverage the momentum to build OpenScience.SE. If we fail, let's merge into Academia.SE.