433

I am here today to update you on the beta site and its future.

First, thank you for all of the feedback - your thoughts are appreciated and have significantly impacted the decisions that we’ve made about how we move forward.

TL;DR - We will be retiring the beta site shortly and will be removing the button to get to it and ceasing support for it.

We will not be migrating the unified posting experience to the main site. Not migrating the unified experience to the main site will obviate the need to solve the issue around the conversion of comments and answers to “replies”, because that was tied to this unified post experience.

We will be massaging the subjective/opinion based content that was submitted in the beta site to the Q&A format, but we are still figuring out exactly how this will work. We will need to do this with sensitivity to the fact that “replies” were more conversational than answers would be.

We will not be migrating most of the design elements from beta at this time: for instance, the homepage, the navigation, the left and right sidebars, and the wider layout.

The retirement of the beta site will happen in the coming weeks, and we’ll communicate more as we get closer.

I think it’s important to recognize the hard work of everyone who put their energy into the beta experience - it did exactly what a beta should do: it explored what new directions and iterations we can take on the site, even if that means eliminating ideas that don’t work. While there’s no question that there were some fumbles in the rollout, I think that ultimately we’ve learned a great deal about how to make major, substantive changes to the site in a way that respects the desires and wishes of all who built and continue to contribute to this incredibly important resource.

Note: This is not an April Fools joke.

56
  • 168
    Okay..., woaw!, that's a good news...! First time I think the Company is listening to negative feedback from the Community... I'm impressed...! Commented Apr 1 at 21:15
  • 16
    Thank you for giving us an update on what the plan is going forward. 🙏 Commented Apr 1 at 21:16
  • 47
    Was genuinely quite worried this was an April Fools joke. Thank you! Out of curiosity, what made you decide to change? Commented Apr 1 at 22:22
  • 21
    @Starship - the beta was always that - a beta. A project to be learned from. Some of what we learned was great, and some of it we learned would need to be reworked. We aren't "changing our mind", exactly, because we had never settled on what would deploy. Commented Apr 1 at 22:56
  • 23
    @JayRiggs I do too. But we're in an incredibly resource constrained era, where every developer that we have is on projects that are mission critical. We have zero free hands for the more lighthearted stuff, currently. hopefully that changes. Commented Apr 1 at 22:57
  • 16
    Perhaps the ---friends we made--- the beta was the april fools day joke after all... Commented Apr 1 at 23:30
  • 47
    My cynical side thinks they laid off most of the people working on this project and then couldn't get a vibe coded replacement working. Commented Apr 2 at 0:01
  • 31
    Would be interested in knowing why the CTO barely responded to any questions in the meta topic he posted. Commented Apr 2 at 0:28
  • 19
    @Philippe Thank you for being active and present in this discussion. It is a positive step, and it is noticed and appreciated. Commented Apr 2 at 0:59
  • 31
    pls give us the old icon back. pls. Commented Apr 2 at 5:54
  • 27
    All the beta did was show how abhorrent your respect for the community is. The CTO didn't care to respond and when you got rid of the CM that was probably responsible to reply, as far as I could see you didn't make any effort to find someone to take over Commented Apr 2 at 8:28
  • 16
    @OrangeDog although there’s no way I can prove it, I assure you that you’re wrong. In the meetings we had about this community response was primary. Stats were secondary, and where they contradicted we went with community opinion. Commented Apr 2 at 9:27
  • 20
    @Philippe The old logo was way better. Now new one could be a tad more bearable if you would keep the orange in it. The same goes for fav icon. And black logo on orange box looks awful. Generally, the box doesn't look good, regardless of colors used. Commented Apr 3 at 11:14
  • 35
    7 days later and the button is still there. When is 'shortly'?? Commented Apr 7 at 11:30
  • 14
    If taking a few weeks to add a line of CSS or JavaScript is "as quickly as possible", it really does not reflect well on SO's development processes. It feels so disrespectful to users to keep inviting them to try a thing you have already decided they will never get to experience for real. Commented Apr 12 at 2:37

8 Answers 8

Reset to default
215

Credit where it's due: A difficult, but correct, decision

I have been very vocal about my opposition to this Beta, but I want to offer credit where it is due. It takes courage for a leadership team to look at a project of this scale, recognize it as a "fumble," and choose to pivot entirely based on community pushback.

Seeing that feedback actually translate into a hard stop, rather than just another "synthesis" of minor tweaks, is a genuine surprise and a breath of fresh air.

Hats off to the team for making the tough call to protect the platform's original mission. Thank you for listening.

I can't help but reflect on my previous discussion regarding "viable" MVPs and experimental guardrails. In that post, I argued that shipping experiments without essentials, and without a clear rollback plan, undermines trust and wastes community effort. Had the "Concrete Questions" and "Checklist" I proposed there been applied to this Beta from the start, a great deal of organizational resources and community frustration might have been saved.

20
  • 55
    Thanks you, and we appreciate the sentiment. There will be one heck of a retro on this one, I assure you. Lessons learned must be documented. I'll raise your checklist and questions there.. Commented Apr 1 at 23:48
  • 7
    @Philippe: I suppose it would too much to hope that one of the ideas which comes up in the retro is maybe reconsidering some other changes which have faced massive negative feedback in recent years and giving them a second look?(Sorry, I just had to ask...) Commented Apr 2 at 20:15
  • 2
    Does the Meta community fully represent the broader Stack Overflow / public-internet-software-engineering user base, or does it tend to reflect the views of those who are most active and vocal there? I don’t have a firm position on the beta site itself, but it seems reasonable to consider that the net/ecosystem has evolved over time, and the platform may need to evolve with it. This may be an unpopular perspective within Meta, but it’s possible that it resonates more widely among users who are less represented in these discussions—or who may not engage with Meta at all. Just 2 cents. Commented Apr 3 at 10:54
  • 3
    @Menelaos Meta is biased and not representative for sure, but that doesn't mean it's wrong. The platform should evolve in the right direction. The company may not know what that direction is. All we have is a string of failures. All we know it's what doesn't work so far. What else can I say? It's all a bit in the open and not resolved by far. Commented Apr 3 at 11:14
  • 12
    @Μenelaοs generalisms like "the platform may need to evolve" are not helpful. At best, it's just meaningless. At worse, it's trying to paint the community as resistant to change for the sake of it. People on Meta have been requesting and discussing ways the platform to evolve for well over a decade at this point. So, we can immediately dismiss the potential fallacy that stagnation is what the meta community supports. The most common criticism levied against the company has been for no change where it's needed, and for completely senseless decisions the result of which is seen as harmful Commented Apr 3 at 11:15
  • 5
    @Μenelaοs also, it's important to remember that ultimately SO set out with a mission. There is expectation that it should stay on course. There is room for change but it's very important the evolution shouldn't fail the mission of SO. Change for the sake of change is also as undesirable as stagnation. Commented Apr 3 at 11:18
  • 3
    @VLAZ "dismiss the potential fallacy that stagnation is what the meta community supports" It's no fallacy. Typically long time users are set in their ways and most resistant to change. All the changes that were discussed in the past were rather incremental. If you had to build a similar system from scratch it would probably look different. And given the decreasing activity, it's helpful to remind people that still things have to change. We just made one step back from the abyss but that doesn't mean that the original direction isn't also doomed. Commented Apr 3 at 11:22
  • 7
    @NoDataDumpNoContribution I didn't suggest any change is dooming SO. I'm saying that 1. People aren't opposed to change just because it's change. 2 Change for the sake of change can deviate from the purpose of SO. On point 2.: I've used SO, I've also used Reddit for the same purpose - finding information. SO, is not the best but it works a lot better than Reddit does. Thus I can say for sure that Reddit-like UI that is the open ended questions does not help but hinders finding information. Thus, the change does bring harm to the purpose of SO which is to expose information. Commented Apr 3 at 11:30
  • 3
    @VLAZ I understand the point you’re making, and I agree that the Meta community has been actively discussing and pushing for meaningful change over the years. My concern is slightly different: even when a direction is internally consistent, ethical, and well-reasoned, there’s still a risk that staying firmly on that path may lead to unintended long-term consequences. History has shown that even very strong platforms can gradually decline when they fail to adapt effectively to changing environments. Hopefully that’s not the case here—but it’s something I think is worth keeping in mind. Commented Apr 3 at 11:40
  • 3
    At the very least, I think we can all agree that we care about the site—that’s why we’re here. Even if we don’t always see eye to eye on policies, direction, or underlying approaches. Commented Apr 3 at 11:45
  • 2
    @VLAZ Reddit and SO are simply two different purposes. Reddit is entertainment, SO is knowledge generation and retrieval. Unfortunately, knowledge retrieval was taken over by LLMs with huge problems for knowledge generation, but entertainment is still thriving. The Reddit-UI is perfect for the Reddit-purpose. The SO-UI is not perfect for the SO-purpose. Commented Apr 3 at 13:22
  • 4
    @NoDataDumpNoContribution I've no idea why you're replying to me. It seems like random musing unrelated to my general points. I tried to point out a problematic statement and explain why it could be problematic. You're here seemingly trying to convince me of something I'm neither arguing against, nor have been invested in. All your points are "the platform needs to change". Great. I don't care. It's not a comment discussion I'm interested in participating in. Commented Apr 3 at 13:34
  • 5
    @einpoklum - that would not seem to be germane to a retro on this project. Whether it should be done independently is another topic to consider for sure, but no, that won’t be part of this retro. Commented Apr 5 at 1:34
  • 8
    @Μenelaοs - that’s a valid argument and one that sometimes keeps me up at night - who represents the voice of the user who doesn’t engage here? To compensate for that, Piper and her Design and Research teams do a ton of research and talk to users who are not active here and try to represent them internally, as does my own community team. Commented Apr 5 at 1:37
  • 3
    @Philippe: I think there is a big thing here that's very easy to understand. "Don't lose your distinctions." Such a plan rarely ends well. Commented Apr 6 at 15:06
52

Thank you for listening to the community here. The direction that the Beta wanted to take the site(s) was a significant point of contention for many, and I must admit, I am pleased that that path is no longer the plan. I also want to say thank you to the Staff Members that voiced the same concerns within the company and finally those that listened to them and that the decision was that the direction of travel was, in fact, the wrong one. This, I hope, will restore some amount of trust that the community has with the company.

I do, also, truly hope that this gives more time for the developers to focus on tools and features that might both be better for the site, and network's development, and will be well received by the community. I, for one, hope that opinion-based content will, for example, finally get some curation tools for both the community and the moderator team (we all know delete, and only delete, is not a good tool).

Though this was quite the polarising, and probably negative, experience for all involved, I do really hope that lessons were learned, and that going forward the experience from this experiment drives better decisions. Many of the users here really do want continued success of the site, and most are the ears on the ground.

I don't know what the company's plans for the future are now, but hopefully we'll see some news on that in the coming months, once the dust settles, and a constructive and well received plan appears.

5
  • 3
    They laid of many of their staff, traffic is way down, and user activity has collapsed to 1% of its peak. I wouldn't expect any significant development any time soon. Commented Apr 7 at 6:21
  • 3
    Adding features that don't break existing workflows, and are beneficial to the community, however, is a good way to get more activity, @SteveBennett . Opinion-based content (OBC) could actually be a good thing to the communities, if it were well implemented well. Commented Apr 8 at 15:07
  • I'm sure it's true that those things would slightly increase activity, but in the face of the absolutely massive disruption caused by AI, it would hardly move the needle for them from a business perspective. Commented Apr 8 at 22:35
  • It might just be me, but I think the AI age has caused their set back. People mostly do a lot of reading by searching multiple sites including StackOverflow for possible solutions to an issue. But now, you could just ask an AI who had scanned these pages and ended up summarizing it. Hence no need to go visit these sites anymore. Commented Apr 11 at 8:11
  • @Sojimaxi Yes, that is the dominant theory for the massive decline in SO activity in the last 2 years. See meta.stackoverflow.com/questions/438596/… . Commented Apr 12 at 10:58
40

Ah so - I do hope the lessons hopefully learnt might carry forward on future and ongoing endeavours. The writing was very much on the wall, with large, deep letters on fire in terms of community feedback on this and the redesign. It was also clear there was some awareness from staff of problems, but it slipped between the metaphorical couch cushions along with some spare change, and maybe my keys.

I think 'who are we designing for' and 'does this fit the core mission of Q&A' need to be important questions. Trying to turn the site into reddit does not make sense, unless y'all were trying to pull a new coke on purpose.

Some of the things that have been tried during the AI brown pants era don't really fit these, and hopefully the lessons learnt will colour what we see moving forward.

17
  • 3
    The lessons learned are less "how not to screw the site" and more "how to keep screwing the site in a way that experts keep contributing to it". Commented Apr 1 at 23:31
  • 1
    TBF - the latter is a longer term, ongoing issue that needs fixing. Its also so big a problem I didn't think I wanted to address that here right now, and focus on potential learnings here. Commented Apr 1 at 23:34
  • 1
    Learning community-friendly things does not fit the MO of the company at all. They have their agenda and do what they can to effect it with as many core contributors sticking around as possible. I guarantee "we should've listened because they were right" is not a "learning" here. At best "we should've listened and made the redditification more subtle". Commented Apr 1 at 23:36
  • 1
    Yeah, but that's talked about in the second and third paragraphs. Commented Apr 1 at 23:46
  • 3
    Also if we're going to do drastic changes, being responsive to the community, and aware of the effects decisions have on the community is a great place to start. Commented Apr 2 at 4:04
  • 2
    That would be a great place to start, but there's no indication we're there. They canned this experiment and trying to sell it as doing it for the right reasons. It would be foolish of them not to do that. But the leopard can't change its shorts. Commented Apr 2 at 8:00
  • 8
    @AndrasDeak--СлаваУкраїні- I get it. You don’t trust us. But I would like to think that we have been pretty receptive to feedback about this. I can’t undo the past, but I can encourage my folks to do their jobs and do them well and hope that people notice and attitudes change. Commented Apr 2 at 9:20
  • 11
    @Philippe you can choose to think that you have been pretty receptive but that won't make that true. There has been a flood of feedback from us, for many years, provably falling on deaf ears. This change here is the first action aligning with our screams (trivial poop polish bugfixes not withstanding), and it's not provably due to any of our feedback. So many times we've been ignored and flat-out lied to. Please do surprise me and implement a 180 degree turn in long-standing company strategy. I'll be happy to see your actions speak a thousand words. But until then I won't hold my breath. Commented Apr 2 at 10:17
  • 12
    @Philippe "I would like to think that we have been pretty receptive to feedback" - good joke. In the face of overwhelmingly negative feedback, SE response was extremely lacking, culminating in Jodys damage-control meta post who was so invested and "receptive" that he commented on two of the 42 answers a week later and then stopped visiting (no exact dates but his profile shows "last seen more than a week ago" for a while now). Otherwise it's been mostly radio silence. Promising "change" without following through has been a pattern for ages as well, so like Andras I'm not holding my breath. Commented Apr 2 at 10:30
  • 1
    @l4mpi - Jody was "last seen" about an hour ago. Commented Apr 2 at 15:53
  • 7
    @Philippe "But I would like to think that we have been pretty receptive to feedback about this." Maybe but we wouldn't know because you don't give feedback back. There are little to no explanations in terms of stats or insights into your thinking. You seem to constantly march in a direction the community isn't fond of says a very long string of negatively scored Q&As. Nothing really indicates that receptiveness to feedback, except maybe the final shutting down notice of experiments. But this is only on the level of "avoided total failure", which isn't so satisfying. I wonder why it's like this Commented Apr 3 at 8:13
  • 8
    @Philippe I'm going to respectfully disagree there. There's many cases where frontline staff seem to understand our issues, but it gets overturned or ignored higher up, even with the process of redesign. I'd argue this whole process is a perfect example of it, and there's others. If the company was good at being receptive to feedback, we'd at least be in a better place with respect to the community, instead of constantly being at odds at it going back a long way. Its definitely something that needs to be worked on, and is an ongoing issue going way back. Commented Apr 3 at 8:26
  • 7
    Instead, we're basically perpetually in survival mode, and with conflicting goals, in a very unhealthy way. Something in the process has gone bad over the years - both in terms of the willingness to communicate, and rebuilding a healthy staff/community culture. Sometimes it feels like the onus keeps being put on the community but... we also do need the company to communicate with us, and internally much better than they are now. Commented Apr 3 at 8:45
  • 3
    The lesson I learned was that no lessons are ever learned. I await the next stackquake. Commented Apr 4 at 1:26
  • 3
    @PresidentJamesK.Polk As somewhat of an optimist, even if that label is somewhat faded and full of holes - I'd say its yet to be seen if lessons are learnt, and if so if it means positive changes or fresh dumpsterfires. Commented Apr 4 at 15:17
35

I really appreciate that the decision to change direction is taken, especially after investing that many resources. I think this is in general a good sign.

Please let me offer some suggestions what might be the lessons:

  • Listen to your community much more often, earlier and more intensely.
  • The step size of the change was too big. You could have easily done only half of it (keep curation and downvotes for example). Maybe there exists an optimal number of changes that typical are more informative in the end.
  • Coupling unrelated design and rules changes together adds unnecessary additional friction and is a bad idea. For example the new icon is independent of how curation works. Maybe one could have introduced unrelated changes in steps to disentangle feedback more easily.
  • There need to be good rationales for a change, "something has to change" is not sufficient. Always clearly name the problem and why you think the proposed solution will help.
  • No curation is not the solution. Curation in some form is always needed. Doesn't mean in the current form, but in some form. Laissez faire doesn't work with Q&A.
  • As a corollary to the last point: promising to add curation later also doesn't work. You need a clear idea and implementation of how content will be checked right from the beginning.
  • If everything can be an opinion, people simply will put in less effort resulting in low quality. You want people to put in high effort. How to do that? Let's discuss with us.
  • Engagement is simply not a useful performance indicator. High quality engagement might be, but how to define that? Let's discuss with us.
  • Fake buttons are a dark pattern. Don't assume we are stupid enough to not see through it. We saw right through it.
  • Chat is nice but storing knowledge in chat like structures doesn't work. Offer chat for those that want it.
  • Forums don't work for knowledge generation either. This was already known, and has been again (and unnecessarily) verified here.
  • Answers and comments on them however work nicely together.
  • Question categories aren't that bad, although they add additional complexity. In any case they need much more guidance and definitely also curation. The community must be able to change the question type for wrongly sorted questions. Maybe even let the community decide what an appropriate question type is.
  • And if you don't start using your own product, you will never understand it.

I could probably go on if I think a bit longer about it, but that should be enough for now. There are many dimensions to it. To me it looks like a capital blunder has taken place.

Are there alternatives for this beta? What else could have been done? Look for example at Focus on answering questions may help to mitigate the decline of traffic. And it can be coupled with making quality generic answers and particularly the answers in it from already three years ago.

8
  • 15
    "Listen to your community ..." I don't know how much they actually listened and read. What was striking though, is that they asked for feedback and invited for a conversation, and there was lots of feedback but only little response, it was no real "conversation". I hope this is part of the lessons learned. Commented Apr 2 at 12:39
  • 1
    @463035818_is_not_an_ai Listening is more than just reading, at least thoughtful consideration of the community input. Feedback would be even better but not strictly needed. They kind of made the impression that there is some interest in conversation before running away. Either they feared that meta bites or the engagement is shaky. Who knows when next we will hear about the alternative plan. I hope there is an alternative plan. It's still true that something needs to change for the network to survive. Commented Apr 2 at 22:19
  • 2
    "Coupling design and rules changes together adds unnecessary additional friction and is a bad idea." <- Disagree with this point, because aspects of the design cater to certain kinds of interaction, which follow from the combination of rules and policies. Commented Apr 3 at 21:09
  • 4
    "if you don't ... us[e] your own product, you will never understand it." <- indeed, eating one's own dogfood is very important. Commented Apr 3 at 21:12
  • 9
    @NoDataDumpNoContribution - thanks for synthesizing this. I agree with the list you’ve created, at least at the 10,000 foot level. I will spend some time thinking a bit more deeply about each, but it’s safe to say that this is a good synthesis of what we could learn from this project. Commented Apr 5 at 1:43
  • @Philippe Thanks for reading. It's just a list, it may be partly or fully wrong. Even with potentially avoiding all these points, there are still quite a lot of possibilities of how SO/SE sites could look like. Maybe less closing and more banners or filtering capabilities. Maybe clamping of scores below a certain threshold. Maybe focusing more on creating useful content. Maximize for non-duplicate, successfully answered question/answer pairs as metric for example. Content creators must quickly get to know what is wrong with their content. People must find useful answers quickly. Commented Apr 6 at 18:03
  • 5
    "Forums don't work for knowledge generation either" - this one has always fascinated me. Meta is kind of in the same boat. On a forum, or say the responses below a news item, everything exists in a bubble. People argue, make assumptions, hate on each other and perhaps even come to an agreement. And then in the next similar forum post / meta post / news item... it's as if the previous one did not exist and everyone starts from exactly the same opinions and assumptions again and history repeats itself. The internet persona has no memory, it just wants to argue. Commented Apr 7 at 12:01
  • @Gimby It's even worse. I could easily argue even with just myself. Give me a topic and I write a pro/contra piece for it. Bike shedding? Loosing yourself in details? Preferring theory over practice (like just trying it out)? Or the general arbitrariness of the world? Commented Apr 7 at 14:43
28

it did exactly what a beta should do

Oh please.

The normal definition of a beta, as used everywhere else in the software industry, is a mature prototype that might still have too many issues to get released, but will definitely be released once those issues are fixed.

If you don't understand the above, then you really don't understand why the whole thing had such major backlash. But I think you do understand what a beta is, you are just trying to save face by retroactively pretending that this was just some "proof of concept"/"demo"/"experiment" to test the waters. That was not how it was released. It was released like this:

1. Soft beta launch
2. Full beta launch
3. Full release

And because of this, a whole lot of users, including myself, thought that this was the final straw that killed the site. The damage is done, and you can't undo it. I do not intend to participate here any longer. I have pretty much stopped all Q&A participation ever since this mess was rolled out.

7
  • 1
    Looks like they failed to take into account that that phrase would mean something very different to them than to us. Commented Apr 10 at 8:59
  • 10
    @MisterMiyagi Except they didn't. Everything in the release post on meta indicates that they know exactly what a beta is and the plan was to push through all of these poorly-designed disasters. And then trying to gloss it over and pretend that it wasn't the plan at all. Commented Apr 10 at 10:10
  • 1
    It's entirely plausible that people who are capable of producing something as technologically horrendous as that beta would have no idea what the word beta implies. Commented Apr 10 at 12:45
  • @Lundin Nah, "launch" and "release" just don’t mean what we think they do. It’s not like they are starting this now – they’ve been pretty much upfront about their words being meaningless. Commented Apr 10 at 12:49
  • @KevinKrumwiede Then maybe a company working with software with the target audience being programmers could consider hiring just 1 person who knows the meaning of various software lingo... Not having one (who's allowed to speak) is actually an even worse state of affairs. Commented Apr 10 at 12:52
  • People who actually know what they're doing create a hostile work environment. Commented Apr 10 at 13:37
  • 2
    It’s been weeks (14 days) and the ability to click on a link to the beta website still exists for some reason. Commented Apr 14 at 20:36
23

This paragraph seems a bit confusing to me:

We will be massaging the subjective/opinion based content that was submitted in the beta site to the q&a format, but we are still figuring out exactly how this will work. We will need to do this with sensitivity to the fact that “replies” were more conversational than answers would be.

I was under the impression these were just the open ended questions that we also have being asked on main currently, so there shouldn't need to be any massaging of content there unless we're also doing away with the open ended questions feature, am I misunderstanding what was actually happening here? Or are open ended questions also going away (in their current form) somehow? Perhaps just a behind the scenes thing?

18
  • 26
    When a user on the beta site left a "reply" to an OBC question, it showed up on the "classic" site as an answer. They aren't the same thing, by a long shot. What we're trying to do is take the ones that are ANSWERS and show them that way, and make the others comments or something similar. Commented Apr 1 at 22:59
  • 5
    Subjective/opinion "questions" took "replies" stored as answers & allowed them to be nested & did not allow comments or reasonable curation. Thus many replies are poor answers or are really comments. Commented Apr 1 at 23:10
  • 3
    ^^ What @philipxy said. Commented Apr 1 at 23:21
  • 2
    @Philippe meaning you're retiring opinion-based questions experiment as well? Commented Apr 1 at 23:27
  • 5
    @M-- No, we are not. However, we're retiring the interface for them that was used on beta. We will retain them as they currently are on the "classic" site - with a drop down. Commented Apr 1 at 23:46
  • 3
    that's what seems confusing to me. the beta's OBQ didn't have answers (yet?), just like live's OBQ don't... so why would there need to be an effort to make replies into answers, if we aren't converting OBQ's to regular Q&A? can't they just stay in their existing format if that format isn't going away? Commented Apr 2 at 5:24
  • @user400654 I am not sure, but what I gathered is that even replies (comments) to a reply (answer) on beta were registered as a top-level reply (answer). Or maybe it's not limited to OBQs, and all questions are like that. Commented Apr 2 at 5:50
  • @Philippe "They aren't the same thing, by a long shot" - great that SE finally figured that out. Why did it take you so long? Commented Apr 2 at 8:20
  • 14
    @l4mpi, there’s a serious misconception that staff is dumb. It’s not a matter of “finally”figuring something out. But there are product cycles and review processes and a need to gather data to support rolling things back. Well contrary to the prevailing narrative most of the folks here are pretty smart. And we realize that we have to establish process and follow it, or risk having people make decisions all willy nilly - which we would be criticized for (rightly). Commented Apr 2 at 9:14
  • 8
    (Cont) - when you stop to consider it, the cycle from deployment to announced deprecation here is dramatically faster than we have done in the past. I choose to believe that’s improvement. Commented Apr 2 at 9:18
  • 21
    @Philippe not sure why you jump to "staff is dumb" when nobody said anything about that. From my PoV it's more an issue of "decision makers at SE push ahead with bad ideas even if they come with sirens, warning lights and a -500 meta post". In this case, there was lots of feedback from the very start of the opinion-based experiment that lack of distinction between answers and comments is a very bad idea. NOW you post that answers and "replies" are not the same - but it's incomprehensible why SE did not figure that out before and took it into account when designing the beta. Commented Apr 2 at 10:20
  • 8
    @Philippe Further to l4mpi's comment: groups can make really poor decisions even when each individual group member is extremely competent. This has been known for a very long time: Stack Overflow was literally announced with the tagline "None of Us is as Dumb as All of Us". It is difficult to communicate this without upsetting people (I certainly find it an upsetting fact, not least because I'm a member of several groups!), but it's important to keep in mind. Commented Apr 2 at 14:50
  • 17
    @Philippe If I may, I agree with you that the product cycle was faster & that this is an achievement. But I also share the community's concern about how we even ended up here in the first place with something that didn't just seem poorly designed or ill-advised, but contained elements that seemed fundamentally contrary to the overall vision & purpose of the site (e.g. not optimizing answer pages for readers). Needing to "gather data to support rolling things back" is 100% justified, full stop; I think we're just also left wondering about the processes that supported the roll out, too. Commented Apr 2 at 15:58
  • 4
    On the frontend, opinion-based questions already had their own slightly distinct UI – even on the current (non-beta) site. On the backend, they functioned similarly to regular Q&A, in that the questions were questions and the "replies" were actually answers. The redesign changed the frontend so that both opinion-based questions and regular Q&A were displayed the same way as one another, looking kinda like a Reddit thread. Commented Apr 3 at 5:37
  • 7
    ...That said, I am a bit confused by the phrasing that the opinion-based questions need "massaging" (design-wise) to fit into the current site – unless Philippe's post is implying is that the separate UI for them on the current site is also going away, and thus those posts will be getting the same UI as regular Q&A somehow. But that's not how that reads to me. Commented Apr 3 at 5:38
12

A minor issue but will you change back the favicon color back to what it was before? The current color is too similar to Reddit icon's color and it is making it difficult to identify Stack Overflow. Frankly, whenever I search a problem online, I generally avoid Reddit results because they are an unreliable mess. The current favicon is inadvertently making me (and I assume other users as well) skip Stack Overflow results as well.

For reference, the following is how it looks like on a google search:

google search result

The favicon was changed right around the beta release so I assume they are connected; that's why I posted here.

3
  • 4
    See Logo updates to Stack Overflow's visual identity Commented Apr 10 at 19:04
  • 1
    One hopes the feedback on redesigns is taken, much as the feedback here was Commented Apr 11 at 0:33
  • 5
    It wouldn't surprise me if some casual SO users see that favicon and think: "That doesn't look like Stack Overflow. It must be some scam site". Commented Apr 11 at 10:22
-43

At last... some wisdom. Wise decision.

Now, in my view, to avoid the shutdown of SO you have to do 2 things:

  1. Fight toxicity to its core: No downvotes without (nicely) commenting, punish harsh comment authors, make a radical "be nice" policy. Do not tolerate harshness anymore. No one likes to be mistreated and any advantage that SO may have will be nothing if the price is receiving harshness. Especially if there are other options of solving the problem.
  2. Fully embrace AI: Create an automatic AI generated answer and let users decide if it's bad or not, like any other user. Normal answers continue to be accepted as always. AI answer can receive comments.

Why add an AI answer? Because if you don't put the AI answer there, people will go to an AI of choice first. If it solves the problem, SO will not even be considered. But if people know there'll be an automatic AI answer, people will also know that users will comment and up/downvote the AI answer and that's a differential from a sole AI answer: people's comments and validation.

SO still has a differential that no AI has: the possibility of showing one's knowledge to serve as a proof to possible employers. And if you put the AI answer there it will have another differential over direct AI questioning: human validation of AI answers.

But none of these will stand against continued harshness. The main problem to be addressed continues to be harshness.

37
  • 8
    They already tried automatic AI-generated answers. It did not go well, either in terms of the quality of answer or in terms of community reception. Commented Apr 8 at 21:38
  • 8
    "But if people know there'll be an automatic AI answer, people will also know that users will comment and up/downvote the AI answer and that's a differential from a sole AI answer: people's comments and validation." Why would enough people invest time doing this? We already have a blanket ban because GenAI output is so hard and exhausting to validate. We already have only a sliver of curators left because validating content here is such a thankless task. We would have an extremely harsh environment with your suggestion 1., making validation even less attractive. Why would people do this? Commented Apr 9 at 3:32
  • 21
    Its somewhat ironic that the words you use are "punish", "radical", "no tolerance" to push a "be nice" policy. You want to fight harshness with harshness? Thats not going to work. Commented Apr 9 at 8:02
  • 9
    The reasons for not requiring comments alongside downvotes have been adequately explained already. Commented Apr 9 at 9:10
  • 2
    "Fully embrace AI: Create an automatic AI generated answer ..." Fully supported. I wonder why SO never did it. Sounds like the obvious thing to do in this age. It would also immediately have killed any incentive for people plagiarizing and copying in their own AI generated content. Commented Apr 9 at 13:51
  • 5
    @NoDataDumpNoContribution "It would also immediately have killed any incentive for people plagiarizing and copying in their own AI generated content." Huh, how so? People post AI answers as pretend-human answers. Having a not-human answer doesn’t make that redundant. It also doesn’t make them detectable, seeing the range of answers GenAI can provide (to the point where I wonder which GenAI answer SO should post…). Commented Apr 9 at 14:33
  • 6
    It is impressive that the shutdown of beta is considered a good thing even from the perspective of those who hold unpopular views on meta. It is also telling that even from their perspective, beta failed to deliver what it promised to deliver, including "more welcoming" and "more engagement". Commented Apr 10 at 4:58
  • 12
    This is actually quite demonstrative on itself. People commented, and then you got defensive and turned the comment section into a long diatribe on "being nice". If "nice" comments are required, we'll only see more and more of this kind of diatribes where people try to argue that they are correct instead of focusing on improving the post itself, and those who commented will get a long series of ping like this. This alone is reason strong enough not to mandate comments for downvotes. Commented Apr 10 at 12:24
  • 11
    If I want an AI generated answer, I'm just going to instantly ask a GPT in my local Cursor chat where it has full context on the codebase I'm working on, or in a first-class AI app like ChatGPT like I already do a bunch of times a day. I can get a response that exactly matches my use case. I'm not going to translate my question into generic google terms, sift through results, maybe land on Stack Overflow, find some similar-looking question and an AI answer and try to engage with that. If I'm on Stack Overflow, it's 100% because I want to know what the humans think about the problem. Commented Apr 10 at 14:47
  • 7
    Even if Stack Overflow had an app and AI-based chat (which it has, in browser, which is so 2023) or even an IDE integration/agent mode, it'd offer zero additional value on top of what GPT apps can already do right now. The product would die, except instead of turning into a sub-par Reddit clone nobody uses (i.e. this beta experiment), it turns into a sub-par Claude Code nobody uses. It's true SO is in big trouble, but I don't see any obvious path forward than the "we've got human answers when AI fails you" business model which is at least a unique service. Commented Apr 10 at 14:56
  • 3
    I understand that you value "be nice" alot. I think everbody does that, its the way to get there that makes a difference. I understand that you are convinced that it needs exclusion and punishment. Thats ok, it just doesn't resonate with my notion of "be nice". I find it odd that you talk about countries/cultures that must be confronted with ostracism. My notion of "be nice" is based on inclusion. I find it odd that you refer me to panic trends, as if this was an ok argument to justify drastic measures. I am convinced that "be nice" primarily requires to be nice, if not I consider it a fail Commented Apr 10 at 17:09
  • 6
    "I don't remember ANY time I received a downvote and understanded why I was being downvoted." And this is exactly the disconnect. You are not being downvoted, the answer is being downvoted. Even on Meta I've occasionally downvoted posts from people I usually agree with if I happened to disagree with them on that particular topic, because I was downvoting the answer and not the person. Commented Apr 10 at 18:57
  • 5
    Waste of time to generate the code to provide an AI answer, it won’t be helpful, and will take away from actual answers. I avoid harsh treatment by not explaining votes, your attitude and this answer, is exactly the reason it’s necessary NOT to explain votes. Commented Apr 12 at 5:03
  • 3
    Why do you think human users would waste their time voting on AI answers? We come here because it's fun, and it's fun because we're learning and interacting with other humans. If visiting these sites becomes an exercise in grading the output of AIs, that will no longer be fun. So what you are suggesting is simply another, probably less good, AI bot. Unless you can think of some reason to convince humans to spend their free time voting on the AI's output, I don't see how this could work. I know I would no longer be interested, for example. Commented Apr 15 at 8:31
  • 3
    "I don't remember ANY time I received a downvote and understanded why I was being downvoted." Ignoring the issue on the wording of "you" vs "your post" being downvoted, you have clearly been told, multiple times, that why someone would have downvoted your post, so you should, by the time you made that comment, understand why your post here is downvoted, esp. if you aim for improvement. If you are expecting someone explicitly stating "I downvoted because...", it is the wrong expectation, and I don't see how explicitly disclosing that the commentor is the downvoter can help with anything. Commented Apr 16 at 16:13

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.