I find it interesting that most discussions I've seen about this exact topic are about Arc and closures... instead of about the fact that this may well be an interesting programming thing to do but it's a moronic user experience thing to do.
Getting user experience right depends on the users. I wouldn't use this technique in an online store. Random online shoppers would be confused by expired links, and you'd lose sales. But HN users aren't confused by them. What HN users care about is the quality of the stuff on the site.
Since I can't work full time on HN, I focus on the things that matter most. What I spend my time thinking about is e.g. detecting voting rings. Those affect what you see on the frontpage, which is what users of this site care most about.
That said: I come for the community - and the community has obviously noticed that the site occasionally throws up an annoying 'error'. The fact that you've done something cool programatically has no bearing on what I get from HN.
So far I'd rate the user experience of the site around 3/5 and the content 5/5. You don't need to work any more on the content unless it starts dropping!
I could care less if this issue got fixed. I have never once felt, "man, I'd definitely jump to another site if it didn't have this expired linky thing happen".
(Now, politics stories on the front page, on the other hand... I've often wished for a site with as good a crowd as HN but without the politics...)
I'd rather have any kind of cool new feature, like messaging, than have this fixed.
TIL Windows has always been an amazing user experience. Just look at their numbers.
> But HN users aren't confused by them.
"About 2,200,000 results" -- says they are.
I have a question: What advice would you give to one of your YC startups if they were having this same issue?
"Ah, your users won't be confused. Ignore all evidence that says they are."
I can tell you exactly why people continue to use Hacker News. It's because it's YOUR SITE. They put up with the broken web design where links die after a few minutes.
Your site won't get beat out by a competitor because it has you, and you fund people. So people continuing to use the site is orthogonal to whether the user experience is any good. You have lots of feedback indicating that it isn't.
Aren't you the one who says "listen to your users"?