1. You're right in thinking that bootstrapping users will be much harder than coding the site. A dating site requires not just a critical mass of users, but also a critical mass using it in a certain location with enough people that there are internal compatibilities.
2. We loosely categorized our users into 3 groups. (a) those explicitly interesting in dating. (b) those who would consider it, but who weren't ready to make the decision, and (c) those who would never sign up for our dating features. Our early strategy was to (try to) nail all 3 groups, arguing that there was mobility between b and a, and that group c could still be used to spread around dating-related toys without realizing that's what they were doing. All this led us to personality tests and a bunch of other goofy things that attempted to be viral but could also enhance a profile if converted. On average, 10% of our early users became online daters on our site, and the other 90% were spreading the word in one way or another, but we never heard from them again. As we got bigger, we slowly shifted our message towards the dating side of things, as those "growth hacks" diluted our message.
Of course this is not the only strategy. Consider (1) Tinder, a very simple dating app, which has become a phenomenon. I would say it's hard just to invent this phenonmen, though... And (2) pay dating services, which can collect membership fees and then use the money to market. It's an expensive game which requires a lot of cash and marketing knowledge.
Disclaimer: I left OkCupid a year ago.