-37

Today I shared an update on progress from last quarter and what we’re up to this quarter on the blog.

TL;DR - top areas of focus include:

As we make progress in each area, we will continue to share specifics on Meta through dedicated Meta announcements. Similar to the last quarterly post, this can be treated as a catch-all.

Those individual posts, and their subsequent posts, will share more details and will be the right place for feedback and questions on those specific initiatives. However, if you have any general questions or feedback about these focuses feel free to leave those as answers here. We’ll be keeping an eye on this post through October 17 and will do our best to respond, where appropriate, during that time.

Improve this question
6
  • 41
    the new data dump should have dropped at the start of the new quarter. That would be the start of October (since the August delay was labeled as an outliner due to revised process). Is releasing that part of the plan too?
    – ꓢPArcheon
    Commented Oct 4 at 8:46
  • 48
    Aside from a few words, this is a literal duplicate of the last roadmap... safe to say it was a bad quarter?
    – Sayse
    Commented Oct 5 at 16:08
  • 10
    @SPArcheon-onstrike The most recent data dump is available as of today
    – Berthold StaffMod
    Commented Oct 8 at 2:00
  • 29
    Could we have an update on how StackOverflow makes money?
    – Karsten
    Commented Oct 11 at 10:40
  • 3
    IMO, there is little to no motivation for users to moderate and improve quality of already existing content on the site. Most of the common questions have been answered so, I think, it is more reasonable to focus on quality maintenance of the content, so users have to be motivated in order to do so. AFAIK there is currently only 2 rep bounty for edit of the post (until you reach 2k) and no other "motivation" for review queues and so on.
    – Monsieur Merso
    Commented Oct 29 at 9:58
  • I have a proposal for this topic but can not add an answer - I'm a moderator on Tor.StackExchange but here I'm a regular user. How can I submit my proposal?
    – Alexey Vesnin
    Commented Nov 8 at 15:17

5 Answers 5

Reset to default
76

For askers, we are evaluating whether a question assistant can be effective to improve their drafts. This would allow askers to improve question drafts earlier in the process, while saving time for reviewers.

We’ve already been down this road, can we just go next and make real improvements to the ask wizard?

Improve this answer
7
  • 29
    Yes. The old wizard actually had phases that were designed to help users avoid asking questions that are off topic, such as honey pots to catch users who had software engineering or hardware recs questions - which also increased traffic to those sites, if I remember correctly.
    – Catija
    Commented Oct 3 at 17:15
  • 56
    Also, I feel like I have to remind people over and over - the fastest answer to your question is finding an existing answer. Tools for asking questions that lead to more questions that are duplicates, even if they're "good" questions, isn't helpful for anyone. But I see nothing in the blog post about preventing off topic or duplicate questions, only about increasing quality.
    – Catija
    Commented Oct 3 at 17:19
  • 2
    eh, i don't completely disagree, but also i do see value in allowing devs who want to to work through and solve problems that have already been solved. It isn't necessarily a benefit to the network's knowledgebase, but it can often be a better experience for the asker looking for help as well as for the user looking to hone their skills while helping others. It's far more interesting/fun to solve a problem than it is to find a duplicate. (though that's maybe unrelated to a better ask wizard)
    – Kevin B
    Commented Oct 3 at 17:49
  • 4
    "can we just go next and make real improvements to the ask wizard?" Let's start with a solicitation of ideas about what such a wizard actually needs to do, and see some evidence that they're actually listening.
    – Karl Knechtel
    Commented Oct 3 at 18:17
  • 9
    Ahem... like this meta.stackoverflow.com/questions/358600/…
    – TylerH
    Commented Oct 3 at 20:06
  • 11
    @KevinB I understand why these would be seen similarly, but as the CM who handled the Content Formatter Experiment, I can say with a high degree of confidence that this is a different approach. I know there is a post in the works that will be able to shed more light on the methodology.
    – Emerson StaffMod
    Commented Oct 4 at 18:24
  • 1
    Hopefully it doesn't involve sourcing questions from outside the network
    – Kevin B
    Commented Oct 30 at 5:41
25

Correct me if I'm wrong but knowledge-as-a-service seems to be mostly about collecting new Q/A to add to its repository with the specific aim to be useful to LLMs providers. User activation, Staging Ground, improving answering, stopping data dump, Preventing unauthorized automated access to the network etc. all seem to be related to this effort.

Are we only talking about Stack Overflow here? Because for most other SE sites, I'd imagine that most useful questions have already been asked (and probably answered).

Also where does curating and/or updating existing Q/A fit in there? I guess it doesn't matter for LLM training, but as a human Stack Exchange user, the most value I get from SE is by finding answers in existing questions that I reach via asking my question in a search engine, not on ChatGPT. Top Stack Exchange posts that show up in my searches usually tend to be canonical questions that have high number of views and multiple answers, and I usually find my answer (and even an answer to a follow-up question) on the same page. I believe that can only be achieved by closing duplicates as much as possible so that canonicals can stay easily discoverable. So how do we reconcile this need for searchability with the need for more Q/A?

Improve this answer
6
  • 17
    "Are we only talking about Stack Overflow here? Because for most other SE sites, I'd imagine that most useful questions have already been asked" - Did you mean this the other way around? I think we're much closer to exhausting the range of useful questions on SO compared to anywhere else on the network. Or, from an AI's perspective, the best training data is already here. Of course as things change there are always new useful questions appearing, but the network has nowhere near the coverage outside of SO as within it (perhaps an argument could be made for a couple of the other tech sites).
    – Bryan Krause
    Commented Oct 4 at 18:35
  • @BryanKrause 1/2 Maybe I don't use SE correctly but for me, partly due to its focused format, SE is useful for common problems that don't require too deep knowledge. Like if I have a question about the Pareto efficiency of an allocation, I'd come to SE because everyone who did econ in college will know about it. But I wouldn't ask about axioms that characterize the NRMP algorithm because such questions would probably be too broad for SE, and those who could answer it are either not on SE or no one else could verify if their answer is correct or not.
    – cottontail
    Commented Oct 4 at 21:07
  • 2
    2/2 So most new questions, at least on the sites I have expertise in, are either homework/duplicate or off-topic because all low-hanging fruits have already been asked. On the other hand, there are always new libraries with a few thousand users popping up in some programming language so even the most basic questions with respect to that library will be novel for SO.
    – cottontail
    Commented Oct 4 at 21:08
  • Voting on answers is still information.
    – EngrStudent
    Commented Oct 15 at 17:33
  • 4
    @cottontail "SE is useful for common problems that don't require too deep knowledge" <-- you're welcome to browse MathOverflow or proofassistants.SE or ... ^_^
    – David Roberts
    Commented Oct 23 at 22:45
  • 3
    Also diy.stackexchange.com, which has a radically different culture than SO, and there is basically no such thing as a "canonical answer." Your comments don't match my experience with electronics.SE either. It feels like you have some very specific SE's in mind rather than SE generally.
    – Rob Napier
    Commented Oct 28 at 17:04
14

I see the Staging Ground developing over at Stack Overflow, and it seems to have had its ups and downs. I only have a vague idea of what it is from the help page (first questions => human review before posting), and don't really know how it has been evolving.

My understanding is that it's going to be rolled out at other sites eventually, but I haven't encountered any requests for input from sites other than Stack Overflow; we haven't been kept in the loop on this. Currently, it looks like only users at Stack Overflow get an opportunity to voice their feedback, and I worry that by the time it gets to other sites, design decisions will have been locked in.

Will sites other than Stack Overflow get an opportunity to give feedback on the Staging Ground?

Improve this answer
7
  • As far as I know, there are no concrete plans for the Staging Ground on other sites but it may be considered once the goals (i.e. all questions going to the Staging Ground and most getting reviewed I guess) succeed on Stack Overflow.
    – dan1st
    Commented Oct 4 at 8:00
  • 4
    @dan1st but that'd only really serve the SO use-case. Wordbuilding and Puzzling have a "proto-SG". SG has cited those as inspirations. Yet at this point of time it's nowhere near posed to replace them. Crucial things that are missing are opting into posting on SG. That's how the other sites do it. And it's very important for even non-new users to be able to use that space - it's there to have back-and-forth to shape the question before posting. SG is aiming to just deal with first-ish questions. There are other things to consider as well. Especially if sites without a sandbox want SG.
    – VLAZ
    Commented Oct 4 at 10:33
  • These things can still be worked out once the decision to give the SG to other sites is made (AFAIK there is no such decision yet).
    – dan1st
    Commented Oct 4 at 11:31
  • 1
    @dan1st or we can have any sort of clarity beforehand. I don't see a good argument for none. Non-SO sites should still have any idea and maybe even any input about such things. Considering a formal "sandbox" is indeed something interesting for those sites. I really don't see why there should be any opposition here. It's up to SE in the end. If they don't want to share information - they are very good at it and don't need user support or approval. IMO, we should be encouraging more openness from them, not less.
    – VLAZ
    Commented Oct 4 at 12:22
  • I am just saying what I think are the reasons ("As far as I know" doesn't mean "I think it should be that way") and I think other SE sites are just not that much of a priority/"not worth" implementing dedicated features for when there is a risk of removing it entirely (or that the "success" of the SG in the eyes of SE Inc is specific to SO and other sites may be irrelevant when it comes to deciding on how successful the SG is).
    – dan1st
    Commented Oct 4 at 12:28
  • I think the answer to "Will sites other than Stack Overflow get an opportunity to give feedback on the Staging Ground?" is "no concrete plans"/"we'll see if it's successful enough on SO" while "Can other sites please get the Staging Ground?" is a different (and definitely justified) question.
    – dan1st
    Commented Oct 4 at 12:31
  • 7
    Right now, the focus for Staging Ground is on Stack Overflow. If, in the future, we decide to expand to other sites on the network, feedback from those communities would be sought after because it wouldn’t be a simple site setting switched. Canned responses on SO wouldn’t universally apply to other SE sites for example. That said, expanded Staging Ground beyond SO isn’t on the roadmap for the near future.
    – Rosie StaffMod
    Commented Oct 4 at 20:37
14

Will User activation at any point include activating users as curators? As of now, the main focus seems to be about completely new users. However, there seem to be no (meaningful) efforts to activate users as editors or reviewers.

From what I have read there seems to be a lack of especially reviewers. The low number of active reviewers seems to limit the amount of questions that can be processed in the staging ground, and is a cause of ever full suggest edit queue on SO.

Improve this answer
5
  • I mean, is that not what adding badges to sg is?
    – Kevin B
    Commented Oct 15 at 13:13
  • 7
    @KevinB That sort of thing is more engagement, meant to keep people going. But there is not really any incentive to start curating in the first place really, as the experience of reviewing can be quite frustrating, at least if you are below 2k (Just look at my past proposal for reviews on MSO). Same goes for editing posts.
    – A-Tech
    Commented Oct 15 at 13:39
  • I mean, i get that, but what would be an example of user activation as you see it, if literally inviting people to participate isn't that?
    – Kevin B
    Commented Oct 15 at 21:13
  • @KevinB Personally I would like to see a solution to <2k users not being able to edit in reviews properly. An issue discussed many times e.g here or in this post of mine. Perhaps one could also implement some sort of reviewer onboarding, though I have no concrete idea at the moment how that may look like.
    – A-Tech
    Commented Oct 16 at 7:29
  • 3
    As part of our user activation strategy, we have internally discussed a few different things related to improving the experience for reviewers and/or contributors. That being said, those are still internal discussions. When we are ready to focus on this and ask for feedback we’ll share more with the community.
    – SpencerG StaffMod
    Commented Oct 16 at 13:43
-2

Hmm, I’m not sure:

Continued discovery on improving answering

How does the community tries to discovery? Is there anything I have missed?

Improve this answer
4
  • 18
    "Discovery" means the company is doing fact-finding and research, to work out how to improve answering, rather than just going with their first idea. (I'm not entirely sure why they do this: while the ideas they come up with are usually good ideas, they're usually not as good as the best feature requests on meta.)
    – wizzwizz4
    Commented Oct 3 at 16:27
  • 7
    @wizzwizz4: Looking at Meta to see the community's suggestions/feedback can be part of the discovery process too :)
    – V2Blast
    Commented Oct 3 at 17:15
  • 3
    If only we had a search capable of surfacing such suggestions/feedback effectively
    – Kevin B
    Commented Oct 3 at 18:30
  • 2
    I mean, we're not in such a bad state that if you asked nicely, you wouldn't need to search. The community is always happy to help - as long as there's a potential for positive outcomes
    – Journeyman Geek
    Commented Oct 4 at 1:51

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .