It's semi-automated scripting from RoG. I'm trying to work out the most useful method(s) of detecting duplicates and cross-referencing. In this case I didn't put the number of comments, and marked the referenced items as "related". Compare with:

http://news.ycombinator.com/item?id=2005790

and

http://news.ycombinator.com/item?id=2004764

where one is marked strictly as a duplicate, the other lists related, along with a note of how many comments.

I'm working on this. There's a sort-of A/B testing, since I'm able to go back and see which cross-references get up-votes, and which get down-votes. Currently I'm getting precious little of either.