Upcoming: Open Questions

post by Raemon · 2018-11-24T01:39:33.385Z · LW · GW · 7 comments

Contents

  Goals for Open Questions
    3. Create a lower-barrier-to-entry action for new users.
  What Counts as Answered?
  Any questions?
None
7 comments

(note: everything in this post is still subject to change)

I wanted to give people some sense of our next priorities (and why you've seen fewer public-facing features lately). The LessWrong team's current project is Open Questions (or "Q&A" for short)

This will be a new format for posts, and probably a new section of the site. Users can post questions, or respond to them with detailed answers, similar to sites such as Quora, StackExchange or MathOverflow.

Many different types of questions will be fine (ranging from noob questions, to open problems in AI alignment, to just trying to understand a given phenomena).

The aspiration here is to aim for rigor in our question-answers (a la a place like MathOverflow), but to encourage questions in somewhat murkier domains, where the answer may not be clear cut, and "progress" on a question may look more like refactoring into a different question, or breaking it up into small chunks and hacking away at the edges [LW · GW].

Goals for Open Questions

1. Develop more "demand driven" content on LessWrong. Right now, people basically come up with ideas for posts based on their own curiosity or interest, but they don't have much sense of how much other people will care. With Open Questions, you know that at least one person actually cares about understanding something (and if their question is highly upvoted, many other people do as well).

2. Tweak the site to focus more explicitly on intellectual progress. Right now, people show up on LessWrong with a vague desire to learn, but not a clear goal. This means that the default discussions often end up feeling more like "nerds hanging out on the internet" than "nerds gaining skills, learning or solving interesting problems."

Q&A will aim to:

Ultimately, we'd like to Q&A section to not just be useful for answering one-of questions but for laying out actual research agendas. (Possibly with a higher-level organization that groups questions together, similar to how sequences group posts).

3. Create a lower-barrier-to-entry action for new users.

We intend for "asking a question" to feel a bit more accessible than making a post.

What Counts as Answered?

This is going to be a bit trickier, since we're intentionally having this focus on domains where answers are less clear-cut, and people may realistically disagree. Our current guess is that the question-author gets to mark the answer that best fits the paradigm-in-which-they-intended-the-question.

(But, questions should also be as well defined as possible, such that it's more clear what should count as an answer).

This seems like something we're likely to reconsider as we think more and get more user feedback.

Any questions?

We're still in the process of hammering out the exact implementation details, both on how an individual question+answers page will work, and how they will fit into the rest of the site. Right now we're wrapping up a basic prototype. We don't have an exactly timetable for this, but it seemed good to keep people in the loop.

We're interested in people's thoughts on the idea so far. Any questions about Open Questions?

7 comments

Comments sorted by top scores.

comment by Adam Zerner (adamzerner) · 2018-12-09T20:16:08.561Z · LW(p) · GW(p)

I'm really excited about this! (Just realized that it's actually been released. Woo hoo!)

comment by namespace (ingres) · 2018-11-24T04:54:17.174Z · LW(p) · GW(p)

We’re interested in people’s thoughts on the idea so far. Any questions about Open Questions?

Can I post any Open Questions I have right now with a title like:

"[Open Question] Bla bla bla bla?"

Replies from: habryka4
comment by habryka (habryka4) · 2018-11-24T06:10:14.870Z · LW(p) · GW(p)

(My guess is you wanted to write "Can't I post any Open Questions I have right now...", so I will respond to that, but let me know in case I misunderstood)

Yep, you can, but I think there are a few reasons why people don't, and why doing so wouldn't get you all the benefit of having a more dedicated Q&A system:

  • In terms of UI, I think you want to give people affordances for asking questions, which we right now don't do (simple things like having a smaller text-field that doesn't scream for 3 pages of content, plus a top-level button that says "Ask a Question"). Having dedicated UI for this, with centrally placed buttons, I expect will significantly increase the degree to which people think of interesting questions to ask.
  • I think question and answer pairs have the opportunity to be better indexed and longer-lived references for important ideas than blog posts are. Similarly to how StackExchange questions often have discussions over multiple years, I think a Q&A format on LW more naturally lends itself to creating more timeless content. But to take advantage of that, you need to have some UI that makes that historical content accessible, and that encourages users to point to previous times a questions was asked, instead of opening a new thread.
  • One of the ideas that has been central to a lot of our thinking on this is the question of how we can measure and incentivize long-term intellectual progress on important issues. In mathematics and many other sciences, well-known "Open Problems" (such as the Millenium Problems, or the problem of squaring the circle, or the problem of dark matter) have often served well at organizing research fields and helped ensure that fields continue to make real intellectual progress. Similarly, the goal with this system is to allow users to create common knowledge about important open questions, and to help people organize all the progress and research that has been done on those questions so far, and I think the current post-system isn't amazing at doing that.
  • I am interested in experimenting with incentives for answering questions, in the form of Karma, but also in the form of prices and financial rewards (similar to Paul's "Could we see distant aliens [LW · GW]" question + prize). I think to do that you need to be able to distinguish discussion from the answer, have the ability to mark things as closed, and some infrastructure for people to attach various bounties to questions in a way that doesn't really make sense for posts. (In a bunch of user-interviews we've found that many people would be interested in doing similar things to what Paul did in that thread, but were unsure about whether this would be welcome on LW, or didn't want to deal with all the logistics associated with awarding prices, both of which I think we can address)
  • In terms of voting-patterns, I think right now people tend to upvote things if they think the content of a post is good, and told them something new, which is relatively rare for open questions, so I think they don't tend to get as much attention as they deserve. I think it helps to create a separate mental category for users for which they can ask reliably themselves "is this a question to which I would like to know the answer to?" as the primary question to ask before upvoting.
  • I think right now if you ask an Open Question, and your question does get a bunch of attention, the answer will come in the form of a giant comment-thread that will probably be 5 times as long as it has to be, which is probably good enough for you as the person asking the question, but is much harder to link to and reference in the future. One of the goals with the new format is to encourage members to distill the content of a discussion into a new top-level answer that includes all the relevant info that you need to know, without needing to dive into the depth of a comment thread. A lot of the goal with the Open Questions system is to distill content into more modular forms, with the goal that this will make referencing past discussion and building on past intellectual progress easier.
Replies from: ingres
comment by namespace (ingres) · 2018-11-24T06:28:54.769Z · LW(p) · GW(p)

(My guess is you wanted to write “Can’t I post any Open Questions I have right now...“, so I will respond to that, but let me know in case I misunderstood)

Nope. My question was literally just whether I can post some open questions I have right now to LessWrong, this sounds like an excellent direction for the website to take.

Replies from: Raemon
comment by Raemon · 2018-11-24T08:57:19.837Z · LW(p) · GW(p)

Heh. I interpreted your question the other way, and my off the cuff answer is "Yes, you can, although it wouldn't automatically get converted into the new format. It would probably be pretty easy to convert into the new format though. But, there's enough pieces still up in the air that I can't make promises about it."

comment by Donald Hobson (donald-hobson) · 2018-11-24T18:52:25.390Z · LW(p) · GW(p)

I think the ideal system would be a subquestion Tree. Someone asks a big root question, and people post splits, that divide the question into several smaller ones. People can answer splits of subdivide them. Splits can have a suggested dependence of one split on others. Questions can have many different splits, as people can come up with many different ways of solving the problem.

Example of 5 different users working together to solve a problem. Lines labeled with the same user letter are expected to be provided by the same user.

Root Question: What is (2+3)*(4+5)+6*7 (User Z)

Split 1: I know that 2+3=5 (User A)

SubQ 1: What is 4+5 (User A)

Answer: 9 (User B)

SubQ 2: What is 6*7 (User A)

Answer: 42 (User F)

SubQ 3: What is 5*(SubQ 1) (User A)

Answer: 45 (User G)

SubQ 4: What is (SubQ 3)+(SubQ 2) (User A)

Split 2: I know 6*7=42 (User C)

SubQ 1: What is (2+3)*(4+5) (User C)

Subsplit 1: Expand brackets (2+3)*(4+5)=2*4+2*5+3*4+3*5 (User D)

SubSubQ 1: What are 2*4, 2*5, 3*4, 3*5? (User D)

Answer: 2*4=8, 2*5=10, 3*4=12, 3*5=15 (User E)

SubSubQ 2: Whats the sum of the Answers to SubSubQ 1: (User D)

SubQ 2: What is (SubQ 1)+42 (User C)

Replies from: Raemon
comment by Raemon · 2018-11-24T23:35:15.682Z · LW(p) · GW(p)

We've talked about things roughly along these lines. I'm a lot less confident about how the details of this would hash out, but some kind of related questions system that helps you break apart hard problems into easier ones seems like something we'd eventually want.

The UI could easily get too complex and would probably require some effort to get right.