It would take maybe 3 days for someone to come up with an algorithm, implement it and send it to pg ready-for-use. All he has to do is ask.

Yup, I agree with you entirely. It would be great, and it's absolutely trivial, anyone could do it. Off you go then, and let us know when it's done. The HN source is available for download, and while no doubt the one PG runs is slightly different I have no doubt that whatever you write can easily be integrated.

It doesn't even need to be automatic. You can just have people say "this" is the same story as "that" - amalgamate.

Shouldn't even be three days - bet you could do it in one.

And yet the concept is rife with potential problems. Which article & title becomes the default, first submitted or highest ranking? What happens if articles get combined that shouldn't be combined? Is there an undo? Who controls that? How simple is merging the discussions?

I didn't mean to imply that this is an easy problem to solve. My point is that there are very smart people on HN willing to figure this out for the rest of us.

Lack of HN features is surely an Arc conspiracy: http://news.ycombinator.com/item?id=830171

the back-end to HN is open-source, so fork and build in your extensions and see if he'll commit it as a patch back to the main trunk

But he also has to want it.

kinda like techmeme?

That would be reddit.

Already exited. It's a black line on the balance sheet. No conflict of interest. Go Paul!