@Andrew: doesn't flagging as "offensive, abusive, or hate speech" have the same effect as flagging as "spam"?

Yes, from the point of view of the software automatically deleting the question (after 6 spam/offensive flags) or removing it from the home page (after 3). I think the distinction is entirely for the benefit of 10k+ rep users who can see spam/offensive flags.

Erdos-Renyi is probably not a very good model for your situation, though.

Edit: For those who are interested, here is a transcription.

The large component of the fuck graph

Define a graph G to be the simple graph whose vertices are the people in the world such that two vertices are connected if an only if the two people have had sex together.

The graph G of course has many isolated vertices, and it has many components of size two. There are also many small components (of size 10 or smaller, say) among teenagers.

Graphs that describe social networks often admit small world phenomena, see for example Jon Kleinberg's article in the Proceedings of the ICM 2006. Does this mean that G has a large component where basically everybody (except those cases mentioned above) belongs to? If yes, what is the maximal distance between two people?

If not (i.e. the graph splits into many components due to age restrictions, geographic distances, etc.), what are typical sizes of the components? 100, 1000, 10000?

A rough summary of the comments is that some people felt the question is mathematical because it might be asking about a probabilistic model of G, while some people felt that the question is non-mathematical because it is asking about G. My sympathies lie with the latter; again, consult the sociology or epidemiology literature.