It's true of course that for users who didn't see the first thread, the second is not a duplicate at all. But we don't have a good way to know who's who, so we have to rely on a global approach, hopefully one which is simple and easily explicable. (This is in the FAQ under "reposts": https://news.ycombinator.com/newsfaq.html).
In the end, there's nobody who sees all the threads on HN, or even all the major ones. Everyone misses tons of interesting material, but what it is varies for each person. This is a discoverability problem. Hopefully we can get better at solving it, but the solution is not to allow duplicates of major discussion, because that would just crowd out fresher interesting material.