I told an intelligent, well-educated friend about Less Wrong, so she googled, and got "Less Wrong is an online community for people who want to apply the discovery of biases like the conjunction fallacy, the affect heuristic, and scope insensitivity in order to fix their own thinking." and gave up immediately because she'd never heard of the biases.
While hers might not be the best possible attitude, I can't see that we win anything by driving people away with obscure language.
Possible improved introduction: "Less Wrong is a community for people who would like to think more clearly in order to improve their own and other people's lives, and to make major disasters less likely."
I think the value of attracting users to LW has a power law distribution. Both Luke Muehlhauser and Nate Soares were "discovered" through their writing for LW, and both went on to be MIRI's executive director. I think the core target audience for LW should be extremely intelligent people who, for one reason or another, haven't managed to find an IRL cluster of other extremely intelligent people to be friends with. (Obviously I welcome math grad students at Caltech who want to contribute, but I think they'll be substantially less motivated to find a community than someone of equivalent intellectual caliber who decided school was bullshit and dropped out at 17.)
Given that, I think LW's marketing should be optimized for very smart people with finely tuned bullshit detectors who may be relatively uneducated but are probably budding autodidacts. (That's part of the reason I added prominent links to the best textbooks thread/Anki when I wrote the about page.)
That sounds like an excellent approach.
However it will make unhappy a bunch of people here who want to carry the torch of rationality into the masses and start raising waterlines :-/