I've always been kind of curious how the people who fill threads with quotes work. Do they have macro hot keys? A spreadsheet of quotes? (In some cases, the first google result for a quote is a page explaining that the quote is fake, so either they have some system that avoids google or they just ignore results they don't like.)
I've always wondered that as well, not being one that remembers quotable material in enough detail to be able to put it into use easily in most cases.
For myself, usually I remember someone said something in a vague way that I think it relevant (often a prior HN submission of some sort), and I start using google and hn.algolia.com to start looking stuff up until I find the article. comment or submission I thought I remembered or I give up. My quotes are generally less quotes and more notes about what I think are interesting references to the discussion at hand. I imagine if even only one out of fifty people reading do that for the whole submission, it still might add up rather quickly.
Of the things I do reference, I generally find I'm likely to use them multiple times over a few months. As things are in recent memory, I see more connections. Sort of like after you learn a new concept you see places to use it all over. It feels like it's just topical all of a sudden, but I suspect it's mostly just that prior to learning it those times weren't strong enough to leave an impression in memory, making it feel like you didn't hear about it before when actually you did.
It then fetches the Hacker News URL, analyzes the text and finds a relevant quote that contrasts the average sentiment of the comments and then posts it.
I think it's more of a birthday paradox type of deal...you get enough people in a thread, it becomes increasingly likely that at least one knows of a long-ass quote that is relevant to the situation.
The birthday paradox might not be a good technical analogy: the number of ways to collide grows quadratically, while the number of people who can individually think of quotes grows linearly.
