It seems like the internet in general, and social media broadly, e.g. LessWrong, has this problem that there's a lot of data. The data has to be filtered somehow. The main way data is filtered is by (1) recency, (2) backlinks, and (3) upvotes.
One problem is that these filters don't distinguish between stuff that's not upvoted or backlinked because it's not so good / interesting / important / novel / true / useful / etc., vs. stuff that's not upvoted or backlinked because it was hard to understand (or, easy to misunderstand as not good) and therefore didn't meet the threshold for being signal boosted. In other words, we aggregate signals of "I don't know about this yet" roughly the same way we aggregate signals of "This isn't good enough that it should be remembered much".
This seems like potentially a big problem, or rather, a big opportunity. Maybe there's good ideas that didn't spread much that we could revive. Maybe ideas would develop faster if people could expect that other people would remember "there's an open question here" rather than just not remembering. But IDK how big of an opportunity it is, so I wonder:
What are some examples of works (of any kind whatsoever) that were / are hard to understand, and so didn't get signal boosted when they were new enough to have attention directed to them to become understood, but instead got lost in the shuffle?
One of the famous examples around here is [Infrabayesianism](https://www.alignmentforum.org/posts/83DimRqppcaoyYAsy/job-offering-help-communicate-infrabayesianism). Maybe some of [Donald Hobson's work](https://www.lesswrong.com/users/donald-hobson) gives examples. What else?