Scalability is the item on the wishlist that is the hard one. Or rather, the one that requires the brains from whatever poor sap they coax into writing it for them. Those guys need a programmer who doesn't suck in the first moments that the project exists. I don't know that they understand what it means to scale the way they're talking: bandwidth costs, CPU. Disk is at least cheap... The architectural choices need to be made by a programmer who's done this before. The tempting early mistake is putting it all into an sql database; this mistake is fatal for scalability. All the work is in that initial data design.
Goes without saying that you'd open-source it.
The initial feature list includes more items about social issues than about software specification, yet the few that are about the software include some mighty assumptions. Modulating down the expectations to something satisfiable given the constraints of the project would be a trick.
The political fu involved in volunteering on something like that would scare me off. Huge demands + huge wankage + huge pressure + no money. The first thing I'd do is build a very very very high wall around myself so I wouldn't have to put up with the chattering classes. Then fling frequent iterative releases over the wall, XP/Agile style. Often programmers desperately need more interaction with their customers, so they can better understand what they want. In this case, the programmers would need insulation from the customers. Note the trend in the comments: "I'm not a 'coder', but I would like it to..." Have rainbow ponies, do my laundry, save me from having to categorize my deathless prose in any way yet magically make it easily findable by category-driven readers... but the commenters probably don't realize that's what they're asking for, because they don't have the domain knowledge. Some translator needs to tell them so, nicely.
The key insight is that the audience for this archive is the reader. Not the writer. Some of the odder writer-centric demands need to be politely ignored in favor of serving the reader as well as possible.
Finding stories is the major task. By fandom, pairing, author, genre, content, tags, "readers who liked this also liked....", however it works. That's the task an archive has to do well. Oh yeah, and serve those millions of stories to the hundreds of thousands of daily visitors without choking. That too.