There's been discussion of this under various topics on Meta. I want to make a few broad observations to start pushing things in a more concrete direction.
My personal view is that we should not include people looking for legal help or advice within our explicit target audience. I would prefer this site to be analogous to StackOverflow rather than Super User: technical discussion between people active or interested in the field (whether professional or amateur), rather than a forum for end-user support. I think I'm in the minority here.
Attracting genuine experts is going to be an enormous boot-strapping problem. I've been relatively active because I'm in a pretty unusual position: (1) I'm an experienced programmer so I'm familiar with the benefits of a StackOverflow-type service; (2) I'm in the midst of studying for the bar exam, so a lot of things are "on topic" for me in a way they won't be in a few months; and (3) I'm not (yet) trying to make my billable hours targets, and my opportunity cost is relatively low, especially when I can justify things as "bar review." With a few exceptions, my sense is that the majority of trained lawyers are in exactly the opposite situation: they have no reason to expect any professional value from the site; most of the questions will be irrelevant to their interests (and there are other reasons not to write about their core topics); and they've got pretty high opportunity costs.
The above two observations are mutually reinforcing: lawyers will be less interested in a consumer-oriented site, and a consumer-oriented site will be less useful in the absence of lawyers.
That said, I think there are a few areas where this site will be potentially very productive even in the short term.
This is a good forum for answering the sorts of questions that librarians answer. Most prominently, helping people find laws, cases, and other resources. Right now, this sort of online resource is widely dispersed, most notably on various law schools' library websites. Both lawyers and end-users can benefit from this, and the dangers of being "wrong" are relatively low, except with answers like "no, there's no law prohibiting that," which are easy to avoid and censure.
We can also productively explain specific statutes, cases, etc. I have in mind translating legalese into plain English. Right now a lot of the resources in this area are bad, wrong, or hopelessly over-technical. Resources for law students would fit in here as well; much as I hate case briefs, I suspect we can constructively answer questions asking for explanations of one aspect of a particular case. And we'll likely be able to attract enthusiastic law students to write such answers: it'd be good writing practice for them.
[This is not an exhaustive list of what I think we can do productively; I just ran out of time to work on this answer.]