Note you can already do this using the hypothes.is tool. I have suggested using this in the past, but people were concerned that this make the discussion less unified, if you had to look in two different places. If this were integrated within the platform, that could solve this problem.
The discussion can be more unified (interpreted as organized with better-searchable ideas) if the comments are in-line and one does not need to search (the same) quotes and their responses in the comments. One would look in-line for comments relevant to the quotes that they like/seek to discuss or learn further perspectives on and under the article they would look for general comments. This is similar to how one would comment on a Google Docs draft that someone asked them to proofread.
Possibly, most commented on quotes could be highlighted - 'community highlighting.' By number of comments, their length, or post part upvote. Are there any bias confirmation/perpetuation on first-come basis risks?
I wonder what searchability (of annotations and linked notes) optimal for the Forum would be. Currently, it seems somewhat difficult to search articles by keyword by the Forum search function, because of the recommendation algorithm that may disproportionately show specific posts.
Super agree! I think a bunch more experimentation could be done with forum design, writing/reading design. Don't be afraid to break out of an inadequate equilibria where people have gotten used to a suboptimal thing, and now it's gonna cost a little to transition to something better.
This is a good idea!
You can already embed Manifold Markets, Elicit and Metaculus similarly (I didn't find where it was announced 🤔), although that's only part of the solution.
Can this be not only comments but also upvotes/downvotes (as you suggest with '+1'), questions, and polls relevant to specific parts, quotes, or sections of the post?
One could find it easier to orient themselves in the community responses to different parts of the text when they can hover over a highlighted part and see its karma and reactions. The reactions could also be categorized and users could choose to see only some type of reactions (e. g. not on typos or clarification questions or polls but yes on complementary or contradictory evidence, challenging questions, and idea advancement).
The community rather than the author should select segment that they wish to comment on. Otherwise, the author could 'hide' a contentious conclusion in a generally agreeable block of text. However, this has the disadvantage that someone can be responding to the key word in the sentence and another person to the entire sentence. Then, comments that could be consolidated would be split, which would reduce the text orientation efficiency.
I have not seen this on the EA Forum feature suggestion thread, which you may be interested in mentioning it on.
I Was listening to EA podcast "Hear This Idea" about biorisks and one thing they stressed was the importance of running the numbers and assigning probabilities and weights to components of a decision as these estimates can often make the difference between what is most effective.
Many posts on the EA forum contain many ideas(example), of which maybe you disagree with specifically one minor point of it. This disagreement may likely only be 5% of the reasoning to support one side of a hypothesis(ex. EA culture may be too big). Maybe because it is such a minor nitpick, and they may have many small minor nitpicks in a few various places, they may not comment at all due to the effort of having to provide the context of each scenario in their comment(such as if they wanted to comment on a line "Thanks for your anecdotal experience, but my experience differs..." or "I think this point was also articulated in [other community builder]'s post about their experience +1")
This may not be a big deal if we lose out on one person not weighing in on 5% of the reasoning of a position that has 80% confidence in sum. But if a few people each weigh in their 5% that 80% may drop to 65%, which may change the action that we end up prioritizing as it may be only worth it if it is at least 72% correct.
A reason why this may not want to be implemented is because it may be hard to read an article for the first time while the comments are present. Maybe by default inline comments can only be enabled when you reach the end of an article your viewing for the first time(in addition to a button at the end of the article/before the normal comment section to enable/disable inline comments manually). On aesthetics it may not be too big of a problem as I think people generally think google docs comments are well made. Although would love some others to weigh in on this.
Disagreements are interesting, even if minor such that they don't think the conclusion is wrong, but just think the confidence given in the conclusion is too high and think the conclusion should be more weekly held. This software upgrade may provide help to facilitate that.
Maybe change the title of this post to “Suggestion:…”? Unless I misunderstanding, the headline suggest that this has already been/is being integrated. However, when I read your post it seems clear that you are making this as a suggestion.
Thanks, updated