Private notes on LW?
Lately I’ve been noticing what a powerup is to read things in google docs, where I can take whatever notes I want as in-line comments without worrying about looking dumb or confusing. In changes my relationship to confusing passages, where I feel much more affordance to think through what exactly is confusing about it.
As a general reading-habit, “copy it into google docs” is a pretty good habit. But I (and I think others on LW team although for slightly different reasons) have been thinking about building a feature directly into LW to facilitate it.
One version of it might explicitly be “private notes” that are optimized as such.
Another version of it might basically just take the side-comment button we already have and add a “private comments” option that lets you set the comment to “everyone”, “only you”, “you + author” (for giving the author feedback in a way that’s more private than a comment but having more context included than a DM) [edit: also, sharing the comment with arbitrary people is a fairly obvious feature here]
Curious what people think about this and what options they’d expect themselves to use.
I’m maybe specifically wondering whether people expect a UI that’s oriented around “arbitrary sharing” would feel good enough as a personal note-taking thing.
Include a “specific people (specify)” read permission too. Then it can be used to write private replies to comments.
Often, private replies are the entire solution to discussions becoming personal, it’s a lot easier to take a hit to the ego, or acknowledge personal failings, when there isn’t an audience. These side conversations are a major part of the solution to performative narcissism and atomization, even though, of course, you don’t see them taking place, in public.
Relatedly, I’d really like to be able to attach private notes to author’s names. There are pairs of people on LW with names I find it easy to mistake, and being able to look at the author of a post or comment and see a self-note “This is the user who is really insightful about X” or “Don’t start arguing with this person, it takes forever and goes nowhere” etc would be very helpful.
I’ve used https://hypothes.is for taking notes on sites that don’t necessarily support that, mostly to share with others. It might be easier than the Google Docs approach.
I’m reluctant to trust personal notes with third parties. They might get leaked or lost or paywalled. I’d be more willing to copy articles into something like Joplin than Google Docs. Public comments at least get archived by archive.org, and weren’t supposed to be private to begin with.
I do like the idea of being able to DM an author with context. I’d probably use it for things like pointing out spelling or grammar mistakes. I’m reluctant to make public comments about those things, because I don’t want to embarrass the author, or seem like I’m trying to steal status points, but as an author, I’d prefer to correct such things in my own posts, and would find them faster if they were pointed out. The potential for abuse seems not really worse than the existing DM system.
I’d like to be able to make groups, such as other users who seem interested in AI in China and have had good approaches to the topic in the past. If there is a reliable way to automatically add a wide variety of specific people to such a list, e.g. people who write posts and comments with tags (or even key words like “china” or “intelligence agency”) that indicates that they frequently write about a specific domain like AI policy.
I’d like to be able to make comments that new users, e.g. ones with the sprout icon, cannot see until they no longer have the sprout icon. If AI safety blows up then the number of people on Lesswrong will blow up too, which means an increase in the absolute number of people on LW who I don’t trust (absolute, not relative). Karma-based restrictions would be neat but I don’t know how to make that work.
I’d like to be able to be able to make certain kinds of comments that exclude certain users from a list, and I add people to that list whenever they engage in bad-actor behavior. There should be a duration option e.g. 6 months, and/or it should be very easy to put people on the bad-actor list and even easier to take them off the list, or easily reduce the number of remaining months on the list if I see indicators that they aren’t bad actors anymore or weren’t as bad as I thought.
This comment reminded me: I get a lot of value from Twitter DMs and groupchats. More value than I get from the actual feed, in fact, which—according to my revealed preferences—is worth multiple hours per day. Groupchats on LessWrong have promise.
Note LessWrong has group chat – it’s in the conversation options button after you start a chat with one person.
Why would LW need a group-chat (or more developed DM) function? If you want non-public conversations with select LW members, can’t you do that today, on twitter, discord, slack, or e-mail?
Yes, of course you could. The concern is user friction.
To paraphrase Douglas Adams, I object partly because it is a debasement of open discussion, but mostly because I don’t get invited to those sorts of parties.
One thing that would be really cool is if you could share your notes on a particular document such that you could produce a sharable annotated version of a document.
For some posts, it may become standard to read Bob’s annotations on Katie’s document rather than to read the document raw.
Maybe consider making it super easy (one click easy) to export LW posts to google docs?
I rarely take private notes, but I do a lot of highlighting. Currently via Readwise Reader, which requires a monthly subscription; before that by highlighting passages in Kindle ebooks etc. The way highlighting is implemented often allows you to attach a note to your highlight, which would correspond to your private notes.
One challenge with highlighting and digital note-taking is: what happens after you finish reading? Can you export your highlights? Do you have to do that manually, or does it happen automatically? Or, if you can’t export your highlights, can you at least see them all in one place?
If you consistently take highlights on one site or device, you probably want to do the same in others. But do they even all support highlighting? And even if so, how do you collect all those highlights in one central spot? Etc.
By default, if those things aren’t possible, then you might take lots of highlights or notes, and yet never look at them again after the fact. At least that’s been my own experience.
Wait. When I comment on a Google Doc, that comment seems to be public (unless I make a private copy of the doc, which based on this, I will do more often—this seems like a very useful habit).
I wouldn’t push for significant work on LW to support private comments, notes, etc. to me, LW is primarily about public posts and comments, and that’s where I’d like it to focus.
There may be a middle-ground feature, like comment drafts or the like, which COULD remain private, but is intended and useful for improving the quality and intentionality of published content.
Another thing that would be multi-purpose handy would be “view markup” or “export in compatible format”, which both makes it easier to understand LW, and easier to copy into a private notes repo if that’s your thing, and easier to archive copies of posts you want to keep offline or independent of the site.
Some data points:
When I initially bought Rationality from AI to Zombies, I took a bunch of notes in the pdf file. Not for every post, but a lot of them.
Then when I bought the first two book versions, I did something similar.
I spent time going through Multiagent Models of Mind and took a bunch of notes for it.
So yeah: for various periods of time when I have material that I think is really important and want to spend time going deep on, I would use private notes.
I might also use it in a more day-to-day sense. Like when I pull up LW and find a post I like. I might take some private notes in that situation. I’m not sure.
That said, I wouldn’t find it very valuable. I don’t think I’d do it too frequently, and when I do feel like doing it, using some other solution like Google Docs seems fine.
A vague background belief I’m forming is “actually, people should do this all the time when trying to read anything particularly complicated” (which maybe isn’t all LW posts but is a significant number of them), so part of the idea here is to nudge people towards that being a more salient option.
(Btw, the question-posing reacts (like what you’ve just given to my comment) have broken workflow: Where do I reply, especially if the react-giver wasn’t in the conversation? There’s not going to be a notification. Even as it is, with you already being in the conversation, I might want to reply in a wrong place from the point of view of the threading. Also, as an obvious workaround, do mentions work with Markdown?)
(the thing I expected you to do was reply to the same thread, I’ll continue the rest of the convo over there)
Sure, it’s general feedback on the reacts feature, not coordination on a particular conversation.
Things that are not written down keep nagging at attention, which is useful for understanding them better. Also forces developing greater fluency to be able to work with them without looking anything up. So I intentionally avoid writing down important/confusing/complicated things, giving them some time to do more work first.
It’s different for instrumentally useful or routine calculations, where either training data is plentiful, you don’t care about improving capabilities of doing the calculations, or quickly getting a result is more important.
To elaborate in response to some reacts: There is a noticeable effect where writing something down eases the mind, so that you no longer try as hard to retain awareness of it. This is less important when you can formulate a task/problem/question and then slot answering it into your workflow. But with confusing things where even formulating the issue is an unsolved problem, it’s important to retain the context that only exists in the mind at that point and is hard to recreate from a poor description. Writing something about it down causes losing traction on it, so it’s useful to hold off for a bit, perhaps until the next day.
I guess the react I actually wanted was “this doesn’t seem true for me, and seems to be making a kinda strong claim?”. “Disagree” felt a bit strong/weird in this context, since I don’t doubt that this is true for you. Maybe I want an “overstated” react. (I could have written that as a reply, but one of the things I wanted was to flag the objection to the line for people who just read the opening comment but aren’t necessarily delving into the thread)
Your original sentence might be true in some contexts for me, but in the scenarios I have in mind, I don’t usefully boot up the confusing thing into my attention to process or anything, it just slips away completely. And for a dense post, where typically each paragraph is somewhat hard to parse or I have a lot of interesting thoughts about, having some kind of working memory augmentation is really important. Little flags about where I’m confused or what my brain was doing in a specific spot helps give me some handholds to recover my attention after I get stuck for awhile.
(Btw, I recently changed my mind about the react functionality. Initially I felt pretty “meh” about it. Now I’m really liking it!)
Ah, right. Hmmm. That updates me a good deal, but I’m not quite sure it changes my mind on whether a) I think I’d use it consistently or b) whether I anticipate others using it consistently.
Sadly, I strongly suspect that a very large majority of LW usage is shallow, along the lines of casually browsing Reddit. However, it seems that the private notes are only useful for deep usage, along the lines of reading a paper in an academic journal.
Personally, my usage is to check LW multiple times a day very shallowly, and then once in a while when I see something important I go a little deeper on it (probably not enough to use private notes; at least not extensively), and then even more infrequently when something seems very important, I go very deep on it (here I would use private notes). And this actually does seem like a pretty good approach to allocating my time. I’m not sure though.
I agree a lot of use is shallow, and, that seems fine. I don’t really care what percentage of the site is shallow, but I care about how much of it is deep, in absolute terms. So I’d be aiming to just encourage a bit more deep reading on the margin, and optimize the use of it.
(It occurs to me now that the admins tracking overall, anonymized metrics for “how much private notes do people take?” might be an interesting mechanism for measuring “deep reading”)
That makes sense. And I agree with what I think you’re implying: that the question of how worthwhile it’d be to build this feature depends on how much deep usage there is in absolute terms. I don’t have a great intuition for how much there is or where the threshold would be, but I’m moving closer towards thinking that it’d be worthwhile.
I would definitely routinely use a private side-comment feature, and also a you+author feature! I don’t like offering corrections for typos in the general comments because it messes up the flow of reading them for content.
I could imagine a few use-cases of sharing with arbitrary people, but haven’t had a time where I wished I had the feature yet.
I think that most of the people who would take notes on LW posts are the same people who would benefit from, and may use, a general note taking system. A system like Obsidian or Notion or whatever would be used for a bunch of stuff, LW posts included. In that sense, I think it’s unlikely that they’d want a special way to note-take just for LW, when it’d probably be easier and more standardized to use their existing note taking system.
If you do end up going for it, an “Export Notes” feature would be nice, in an easily importable format.
FWIW, I have a general notes system (in the genre of Obsidian / Notion / Roam / LogSeq / etc.) and also a Zotero full of downloaded articles and books. I keep almost everything in my general notes system … but I also sometimes attach individual notes to things in Zotero. These days those Zotero notes are usually one of the following two categories: (1) a note that just says “See my general notes”, indicating that I have written something about this article into my general notes system (and then I can go find it by searching for the author’s name, if I want to), or (2) a list of acronyms / jargon that are specific to that article and that I don’t already know. I can imagine using LW private notes for those same two things, just like my Zotero notes, if such a feature existed. It’s very low priority for me, though—I’m getting by just fine as is.
if one goes in this direction, then a natural next step might be the ability to organize private multi-party discussion...
the main use case would be as follows, in my opinion: there are plenty of sensitive technical topics in the field of AI existential safety; in particular, many important approaches and techniques are of dual use, they can be used to improve safety and they can be used to boost capabilities; “security via relative obscurity” is one possible approach here, but LessWrong does not currently have any mechanisms to support additional levels of “security via relative obscurity”...
the question is: what would be a downside of having something like that? let’s ponder this...
hm maybe you have a private note version of a post, but each inline comment can optionally be sent to a kind of granular permissions version of shortform, to gradually open it up to your inner circle before putting it on regular shortform.
For the same reason, I usually prefer PDFs, because I can annotate them however I want, and they won’t change. I would gladly welcome the possibility of taking notes on lesswrong.
Have you ever printed-a-LW-post as a PDF and then annotated it? Curious how that’d go for you.
No, when I need to take notes on lesswrong, I open a note taking app and copy-paste the paragraph because it’s quicker. Now that you make me think about it, I should try out printing it to pdf, and keep my lesswrong pdfs organized.