-38

“The reports of my death are greatly exaggerated." - Mark Twain

We all have seen the graph that is making the rounds. The question posed. The assumptions made. I have seen the speculation. Many have weighed in, offering their takes, their conclusions. But now I’d like to offer my perspective, one that is shaped from a place that is both deeply personal and uniquely informed.

No, I don’t believe this is the end.

It’s a moment. To pause, to reflect, and to remember why this place exists in the first place. It’s an intake of breath before the first step. It’s a chance. A precipice. An opportunity to rebuild, but not just for the sake of survival, but because what has been built here matters. Because Stack Overflow remains something rare. It’s human. It’s messy. And it’s beautiful.

This isn’t just a website. It’s a community. A place where people come together not just to solve problems, but to share what they’ve learned, to connect, to help each other grow. It’s imperfect, sure. People disagree, mistakes are made, and sometimes it’s not as smooth as we’d like it to be. But that’s what makes it real.

That’s not something you find everywhere. It’s not something you can replicate with a machine. Because it’s driven by something machines don’t have: empathy, curiosity, and the willingness to help a stranger.

I know things aren’t perfect right now. The once-unrelenting tidal wave of knowledge sharing has slowed to a trickle, and it feels like the world is moving faster than ever, leaving places like this behind. But I don’t see this as the end. I see it as a turning point. A chance to reflect on what’s been built here over the last 17 years - a place that’s survived on one simple, powerful idea: that knowledge should be shared freely, and that human connection is at the heart of learning. That mission hasn’t changed. And it’s still just as important today as it was when this all began.

We’ve all been there, those first uncertain steps of discovery or the frustration of problem-solving when the sheer volume of information leaves you unsure of which path to take. This place serves as a lighthouse in those moments of uncertainty, a guiding hand when you’re reaching out in the dark. Here, unlike anywhere else, seekers are never truly alone. It’s about the person on the other side of the screen. Someone who’s walked the same path, faced the same challenges, and is willing to share their hard-earned wisdom. That connection isn’t just valuable; it’s irreplaceable.

So no, I don't see this moment as the end, but the possibility of what could be. A chance to take everything we’ve learned - the good, the bad, and the messy - and use it to move forward. To double down on what makes Stack Overflow unique. To remember that this isn’t just a website. It’s a community. A place where people come not just to solve problems, but to connect. To collaborate. To build something bigger than themselves.

This is our chance to create something new. To embrace a fresh start, a bold reimagining, driven by the same enduring mission. To take the lessons of the past and shape a brighter future. A future where the community not only survives but thrives. A place that is stronger and more resilient than before. This is a moment to rebuild and to reimagine what this place could be. Not just for those who are here now, but for those who come after, seeking the same light, the same guidance, the same connection. This is the time to create something extraordinary. A space that continues to inspire, and in a world increasingly driven by automation, reminds everyone what humans can achieve when they come together.

And that’s why this matters. This place serves as a reminder of what it means to be human. It’s not always easy, and it’s not always perfect, but it’s real. And that’s why I’m not giving up. I still believe. Because what happens here matters. It matters to the people who come here looking for answers, and it matters to the people who stay to share them. This place is a testament to the power of human connection, and that’s something worth fighting for. That’s something worth believing in.

32
  • 36
    This could maybe rather have been posted as an Answer on this Post: Do you agree with Gergely that "Stack Overflow is almost dead"? Commented Jan 20 at 15:28
  • 22
    @chivracq It could have been, but I wanted to give space for people to respond. Commented Jan 20 at 15:31
  • 70
    "This isn’t just a website. It’s a community." How can it be where isn't really any way that people communicate with each other directly. The only thing we ever had for that is chat and the only community that ever existed on SO is those various little cliques created by chat users. Which is a tiny minority compared to all users of the site. Yet the company keeps yapping about this supposed community that doesn't exist and never did - to the point where I put the word 'community' in the very center of by buzzword bingo ticket. Commented Jan 20 at 15:50
  • 95
    Also, in the next second after posting some blog with lies about the love for the community, the company somehow always finds a way to take a dump on said supposedly existing community with some AI trash experiment, ignoring all "community" feedback, ignoring common sense, ignoring any form expertise for how to build web sites. Why a significant part of the supposed community has left the site. Not because of GenAI, not because of site feature-this-and-that, but because they don't trust the company. Trust has been permanently erased. Stop talking about communities. It is offensive. Commented Jan 20 at 15:55
  • 25
    @Lundin, I disagree. A community is a group of people who share the same goals, interests, or mission. It's a place where individuals come together to contribute to the whole. Sharing is the common thread and think that describes this place perfectly. Commented Jan 20 at 16:02
  • 24
    As for sharing, the company's idea of sharing seems to be selling the already public licensed material for money to evil AI corporations while blocking data dump API access to those who actually wrote the material... Sharing = stealing someone else's work without attribution and selling it for personal gain, right? Commented Jan 20 at 16:16
  • 46
    @Bella_Blue "A community is a group of people who share the same goals, interests, or mission" - by that measure SE is not a part of the community but explicitly antagonistic to it, as SE is working directly against community goals on several fronts (see for example the LLM trash, or hastily coded "experiments" that are pushed through even if the community tells SE they absolutely hate them). It is also abundantly clear that SE management is on a mission to get a return on the $1.8B investment, which the community doesn't care about and which directly leads to some of the above conflicts. Commented Jan 20 at 16:40
  • 13
    Side note - shouldn't the "company-update" tag be used for, well, actual updates by the company? This instead seems to be a personal "vibes" post and not an update by SE, as it is written from a personal perspective. Commented Jan 20 at 17:34
  • 19
    “The reports of my death are greatly exaggerated." Mark Twain died over a hundred years ago. Reports of his death are no longer exaggerated. Interpret the metaphor however you want, I suppose. Commented Jan 20 at 19:51
  • 16
    Moderator note: Reminder to please post an answer if you want to respond to this discussion topic. The comments on the question itself are not well-suited to multiple back-and-forth discussions. Commented Jan 21 at 7:20
  • 33
    "Human" this and "human" that but SO keep trying to make it NOT human by forcing AI everywhere. "Community" this and "community" that but SO have been actively acting against the community for years now. "Opportunity" this and "turning point" that but all I see is a company panicking because the money stopped flowing, and now they resort to the very few users left (after they drove away all the others) in a last desperate attemp to keep the wheel turning by appealing to the very same principles that they've been ignoring for a decade now. Sorry that you as an individual have to endure this. Commented Jan 21 at 12:48
  • 14
    Another Moderator Note: We have had multiple flags accusing this post of being AIGC. We have investigated and have not found that to be the case. Further flags will be declined. Commented Jan 22 at 14:46
  • 13
    "A community is a group of people who share the same goals, interests, or mission. It's a place where individuals come together to contribute to the whole." As a person who was a consistent contributor who felt pushed out by the forced licensing changes, the actions taken against Monica, and the reduction of capabilities, I think this is a tone-deaf and galling post. I would love to be a part of a StackOverflow community of helpful people, but I earnestly believe that the company sees us as little manipulable disposable factories of monetizable content. What a disappointment SO has become. Commented Jan 26 at 16:57
  • 13
    We hope that you finally stop ignoring our feedback, that you stop pushing features that no one likes... But no, all we get are these pointless motivational speeches. Commented Jan 27 at 13:55
  • 13
    Community Manager Shog9 knew how to move mountains, but the company saw fit to fire him, Robert Cartaino, and Monica Cellio. They completely ignored "community" then. That was back when SO was handling over a million questions per year. If I'm exaggerating, I stand corrected. The company had the golden goose, they gorged its throat until the liver–the organ of life– burst. It's all words, words, words from well-meaning employees who just don't get it. Commented Jan 27 at 19:46

31 Answers 31

Reset to default
1
2
115

The elephant in the room is that Stack Overflow is a business. So the question is not whether we can create something useful here, but whether we can create something useful that is also a viable business.

There are probably a lot of ideas that could work on a small scale and produce something useful. But I don't see how the public sites can remain a viable business without drastically increasing user activity again.

And the more recent features don't inspire confidence in me. They're too half-baked, very buggy and not thought through. Unlike significant parts of the community I don't even disagree with the idea of introducing more subjective content and increasing discussions. But the company is skipping all the hard parts, thinking about how this kind of content could work well within the existing systems.

10
  • 27
    And it's imoprtant to remember that "viable business" from the perspective of Prosus probably means getting a return on their $1.8B investment and not just being able to cover the operation and staff costs... Commented Jan 20 at 17:52
  • 21
    Then the question should be whether Stack should continue as a business or go the route of Wikimedia and Mozilla, i.e. continue as a foundation. Community and business don't really fit together. Commented Jan 20 at 19:19
  • 3
    @l4mpi That money was gone as soon as they signed the deal. Tricking them into believe that SO for Teams was the big deal was a nice con though... Commented Jan 21 at 7:21
  • 17
    Stack Overflow was created to fill a hole that existed in 2008; a lack of quality information. Now in 2026 the world has evolved (or devolved), people are way more prone to make do with low quality generated crud. That is not going to change. Where I see Stack Overflow fill a completely new hole which will be coming into existence right now is of the "Let me tell you how your AI is stupid" kind. I see the scope of Stack Overflow widening to be able to provide answers to quandaries that people will most definitely be running into when their Artificial Unintelligence hallucinates nonsense. Commented Jan 21 at 9:20
  • 4
    ... which will make pretty much every established Stack Overflow frequenter rage quit because they'll absolutely hate that kind of question. But that's just it... times are changing. People are changing. Not for the better, in my opinion. Commented Jan 21 at 9:20
  • 5
    It makes sense that features are half-baked (i.e. with minimal functionality). They need to see what has potential and what doesn't before investing more resources. Instead the issue is they wasted the first years since the advent of AI trying to chase the rabbit instead of immediately repositioning the platform on the market. Regardless of whether that was a conscious choice or mere incompetence, this is a blatant and utter failure of the company's leadership. Now we're about 3-4 years late and have to scramble. Commented Jan 22 at 5:52
  • 3
    "without drastically increasing user activity again" [citation needed] SO business model depends on views, not interaction. Commented Jan 22 at 14:46
  • 1
    @Gimby To fulfill that role you propose, is impossible with the current ruleset and the way active users of this website behave. The majority of questions that would arise from a hallucinating AI will be flagged as "duplicate question of...". That means I can just as well google or learn to use AIs better. Views matter even more than interaction for the business perspective. And those views to SO came much more from beginner devs. That role has been covered by AI well enough... Maybe not as accurate but with more patience, instantly and without the risk of getting downvoted. Commented Jan 22 at 15:04
  • 3
    SE's been view focused for years but... without posts, who will want to view the site? Activity means utility, and that's what brings people here. Commented Jan 25 at 0:52
  • 1
    @AlexGeorg you may have noticed that the site is already being changed though. They call it "experiments". More like in a blind panic making all kinds of changes to try and get the views back. In that haste, currently they're kind of neglecting actually updating the site rules. They just introduce new features like new question types and then remove functionality from them, like being able to close them. Or see downvotes. Commented Jan 27 at 14:19
77

I have a hard time being optimistic about stack's renewed interest in working on the public platform. Many of the efforts stack presented as being the future in the past 3-6 months are a step away from Q&A, rather than an improvement of it. An abandonment of what made SO different in favor of just doing what everyone else does. None of this is innovative. A chatbot in a sea of chatbots, an unmoderated discussion board for open ended questions with no real answers... I don't see the point.

40
  • 10
    I don't really believe that... because it's been years since we've had stack actually address what the core mission even is. The tour exists and says one thing, and the community exists and sees the core mission as something else entirely. What is the core mission? The tour explaining what SO Q&A is hasn't matched with the reality of how SO Q&A operates in a really long time. Commented Jan 20 at 16:41
  • 31
    @Bella_Blue Y'all say the mission hasn't changed, but your actions suggest otherwise. Breaking the promises around the data dump is the most egregious, but the thousands of little sabotages of quality standards to boost engagement seem to indicate the core mission is no longer about building a library of knowledge. It's about building a community that you can sell to advertisers, who donate and curate information you can sell to AI companies. You aren't evolving SO to become a better repository of knowledge. You're evolving it to generate more ad revenue. Commented Jan 20 at 16:42
  • 5
    That's certainly aspirational, but what does it actually mean, and look like, when implemented? the current site mission statement is "Ask Questions, Get Answers, No Distractions." It's direct, to the point, and easily understandable by anyone who reads it. but it's not the reality people are met with when they ask a question, nor when just simply browsing. Commented Jan 20 at 17:05
  • 13
    @Bella_Blue "helping" is a trigger word for many of the experts who signed up to the original mission. If you scour meta you can find many references from the early days of people stating that SO is NOT a "help site" but a knowledge repository instead. SO explicitly did not want to become a place for tutoring/mentoring of individual users, or offer personalized help. See also the "too localized" close reason (rip, still sorely missed). Commented Jan 20 at 17:10
  • 4
    @l4mpi that's certainly not my opinion, helping people is very much an important part of what SO is. It presents itself as the helpdesk devs always wanted. A place to go ask dev related questions and get answers from experts, with the best and most useful pairs rising above the rest. It's how SO has always been used by devs, whether some people want it to be a helpdesk or not. Commented Jan 20 at 17:12
  • 6
    @user400654 that depends heavily on how you interpret "helping". Some people mean personalized help, as in mentoring someone instead of turning them away when they make it abundantly clear that they have zero clue about programming. SE tried to go in that direction with one of the recent failed events, which thankfully is one of the few that didn't graduate despite community objections. Others instead mean indirectly helping people by curating a library of knowledge that they can access to find solution (and incidentally helping the OPs asking questions), which is more what SO was intended for Commented Jan 20 at 17:17
  • 6
    @user400654 and a big distinction between the two is closing posts. A site intended for personal help first and foremost would not close questions and thus turn away people in need of help. A site intended as a high quality knowledge repository primarily cares about the quality and topicality of the question, and accepts it won't "help" everyone who manages to submit something. Commented Jan 20 at 17:18
  • 6
    No, a helpdesk can still exist with closing questions that don't belong. Helpdesk doesn't have to mean traditional forum with no rules, SO was an iteration above that. A place where if what you asked was already answered you could be pointed to the right place. A place where if someone asked a question that was disrespectful of answerer's time (such as lacking details, asking for a book/tutorial, etc) it would get closed. I am not insinuating SO should be or was ever a facebook-like discussion board. Commented Jan 20 at 17:21
  • 6
    @user400654 and these "disrespectful of the answerer's time" criteria were severely weakened over time; the above mentioned removal of the "too localized" and "required minimal understanding" close reasons were the first instances of this I was around for. There were also long discussions if SO should not offer some sort of mentoring explicitly for users who would otherwise be told to go look at a tutorial to get a clue. All this to say - "helping OP" is not a factor when deciding to accept a question to SO or not, and it should never be. It is a side effect of building a good repository. Commented Jan 20 at 17:27
  • 5
    too localized and minimal understanding were both heavily abused, i'm 100% on board with their removal. They mean nothing and can be used to close literally anything. Commented Jan 20 at 17:30
  • 5
    a knowledge repository is only as helpful as the knowledge it contains. Currently SO is quite stale in that department, and new initiatives don't seem to be doing anything to address that... instead they introduce open ended discussions with no real answers. Commented Jan 20 at 17:37
  • 4
    It's different target audiences really. There are the people who want to, like always, do a Google/Bing search and get relevant answers handed to them from Stack Overflow. The same people are very willing to guard said knowledge base with their life and are thus very apposed to the new style questions. And then there are the people that want to dump a question and get a personalized answer, a growing audience. I don't think both flows of questions and answers can stem from one and the same database - the target audiences and their expectations of the data they will get are too different. Commented Jan 20 at 17:47
  • 6
    @l4mpi I don't know - can we be both? Can we serve both audiences? I think we can. For me, when I start something new, I want to learn everything I can about it - immerse myself in it as much as possible. Yes, I have questions that need answers, but I also want to understand how that answer came to be. I can see a future where we serve both needs. Commented Jan 20 at 18:09
  • 8
    @l4mpi Stack Overflow has always been something of a help desk. Almost every asker has come here looking for help to get unblocked, gain a better understanding , etc. Some of those askers have written great questions that make it easy to create repeatable answers. But the needs of devs have changed since then; that doesn't mean becoming a no-rules help desk, but I do think it's worth asking how we can remain a relevant resource for new devs just starting out. Commented Jan 20 at 18:24
  • 6
    Right now, there is a fundamental conflict of interest between these groups: one prioritizing long-term signal and the other short-term velocity. Navigating this requires careful, deliberate thinking, but instead, we see "bolt-on" features that contradict the core mission of a high-signal repository. The recent handling of the Data Dump is a prime example of this; it shows a lack of care for the very group that builds the knowledge base you're trying to leverage. You can’t claim to support the repository while simultaneously undermining the community's trust. Commented Jan 20 at 22:54
57

[This post is not directed to Bella personally - any presence of 'you' in the post is addressed to Stack Overflow - the company.]

I tired of these kinds of posts long ago - I don't want any posts from CMs or middle management trying to sugar coat how bad everything is - this is a just huge waste of everyone's time and won't change anything. Everything I have to say about that can be found here.

Instead why can't the CEO wake up and post these meta threads themselves? Then it might even be meaningful to give feedback. The CEO is ultimately responsible for the fiasco, even though the negative trend started even before Chandrasekar took over. There was a sudden outburst of activity from him one year ago here, which is the reason behind the chat overhaul, opinion-based questions experiment etc etc. Then he turned silent again.

What I really would like to hear from him is him thoughts of why all of these often quite good ideas - I support several of them - end up implemented so poorly. The opinion-based questions, new editor, new graphic design (hot dog theme) and comment overhaul in particular stand out as good ideas implemented in outright horrible ways. And that's without even mentioning AI-something-something, which was not even a good idea to begin with.

It is not isolated occurrences, every good (and bad) idea during the past 2-3 years does not live up to expectations when implemented. As if the implementation was handed over to complete laymen.

You shouldn't need to do a beta launch and get community feedback in order to tell that Stackoverflow AI crashing when started needs to be fixed before launch, or that a new editor shouldn't wildly bounce up and down when typing, or that we can't have a question category named 'advice' together with no moderation instructions, or that opening up chat to 1 rep users means there will be a flood of trash & spam that needs to be moderated somehow.

This wasn't the case some five years back. Is it because the last competent staff left the company with the 2023 layoffs or is it because they started leaving/were fired even before that?

We can't rebuild if there is nobody who has the skills to rebuild. Yes, I do believe this is the end.

9
  • 3
    @l4mpi He was quite reasonable when you got a chance to speak with him in comments through meta - that's a different person than the one who writes the various marketing buzzword bingo blog posts. The point is that after the "ask me anything" meeting, things were actually starting to happen. Unlike the hundreds of posts where we give feedback to CMs or middle management: they clearly don't have the influence to make the feedback happen anyway, should they even listen to it. We tried that literally hundreds of times: it does not work, huge waste of everyone's time. Commented Jan 22 at 12:20
  • 5
    Specifically this is what set everything in motion: the opinion-based questions/Discussions revamp, the chat experiments, site design changes etc etc. As is expected with US-based company culture, very hierarchical and everything goes from the top down, with nobody in the middle allowed to think or use their own judgement. That's a very dysfunctional way to run a company, but that's another story... Commented Jan 22 at 12:27
  • This sort of is my feelings on this whole post, kudos to Bella articulating their thoughts but its a bit futile when those above are less interested/able Commented Jan 22 at 15:30
  • 2
    @Lundin being able to produce replies that look "reasonable" does not exactly change my mind here - even LLM chatbots can do that nowadays. And from an outside perspective it's impossible to say if SE operates in the dysfunctional way you describe, but if it does this is also a failing of Prashant as the CEO. And the topics you mention that were "set in motion" by the AMA are the same you rightly complain about as being badly implemented, so if more discussions with Prashant results in yet more badly implemented "experiments" I'm not in favour. Commented Jan 22 at 15:37
  • 3
    I think the general/blanket indirect statement that current staff aren't competent is uncalled for. It's not constructive, or entirely fair. why do you assume these are due to a lack of skill? from my perspective, I see gaps in understanding of 'product" (the public platform), lack of dogfooding (and sometimes apparent lack of testing), lack of presence in day to day community activity and discussion (until there are concerns about what we're doing), and lack of interest (whoever makes the calls) in working on things past MVP. Commented Jan 23 at 4:42
  • 1
    I assume retention of corporate knowledge suffers from the layoffs, and I get the feeling that for all the things/experiments the company is trying to do, there's a lack of human resources. I think particularly given these two things, (even though I'd love to see more of the other things I listed above,) it's unfair and unjustly harsh towards current staff to indirectly say all the competent staff are gone. Commented Jan 23 at 4:43
  • 2
    @starball well also culture - both within the company (which went from dev/community first, to one that was more focused on marketing a SAAS product or two, and more corporateish. Certain competencies were lost, or de-emphasised. And sometimes these are difficult to bootstrap, or not everyone believes they're important. I'm not saying folks are incompetent, but people competent in 'soft' skills of certain sorts are generally rare, and not really appreciated as much as they should be. Commented Jan 23 at 7:41
  • 6
    @starball This is literally a platform for programmers and web designers, who has been potty-trained by the site design to down vote. And I'm done with good faith of this company since many years ago. So if they are doing a poor job, it will get called out. They are doing a poor job. If it was just a bug here and there nobody would raise an eyebrow but if literally every new feature is subpar, there is some big systematic problem within the corporation, not necessarily related to everyone leaving the site, but happening at the very same time. Pretending that it ain't there won't fix it. Commented Jan 23 at 8:00
  • I think you hit a pretty core issue @Lundin. There's been a number of features (or "experiments") over the years that I thought had potential, even sometimes contrary to "popular opinion", but were basically stillborn because of implementation (and a snail's pace of improvements). If something launches unusable, and stays that way for a month, the well is basically permanently poisoned; even if people's claims of it being a bad idea all along aren't quite right... Commented 59 mins ago
39

The above question post is well-written, but it does not appear to offer any solutions or ideas how to get out of the current situation. Much has been written in other answers already, and I would like to take the chance to add a few personal points.

To slightly counter the discussion about perceived toxicity/steep learning curve/over-moderation on this page: I've become active on SO in 2019; after many years of passive reading and learning I wanted to give back and contribute my own expertise. I never faced any problems when asking and answering questions. I do my own research and only ask when I am sufficiently sure that the solution is not trivially discovered (i.e., I don't waste the time of other experts). Answers should be self-contained and cleanly offer/explain a solution -- exactly what I would like to see myself when researching a problem. If I spend an hour solving a particularly annoying issue, that might be a candidate for a self-answered question.

To be clear, I'm very much in favor of improving the experience for new users; everyone approaches problems differently, and the Staging Ground was a promising start to get them gently introduced into the site's culture. But for me it's certainly not as horrible as some comments or off-site discussions claim.

Side note: On the other hand, the flood of low-quality questions during that time was very frustrating -- I wanted to contribute effectively, but spent lots of time searching for answerable questions that had value beyond that specific isolated problem. Thus, I got involved into moderation/curation. This was an entirely different beast with tons of history, long meta discussions and unspoken rules to be aware of; though doable, it could really use an overhaul. I wrote more about that here.

I seldomly use LLMs for solving problems; I like to understand what I'm doing (and why), and prefer to build my solutions through research. SO is one of my primary and most trusted sources, though recent Google algorithm changes made this somewhat more difficult, and SO's native search function is still notoriously bad.

All these things haven't changed for me, and I would gladly continue contributing as I did in the past years. There may be fewer questions due to LLM assistants, but the remaining questions are more likely to be interesting, and worth the time salvaging them. However, something has changed, and that is the treatment of the company towards us experts and volunteers. It always had its ups and downs, but in the recent years it has reached a new quality.

See, I'm a developer. I think logically, and I like to spend my time efficiently. When something unnecessarily gets in my way, I become annoyed and want to automate it away. Unfortunately, the company has done everything in its power to get in my way, and at some point it was just enough.

To list just a few:

  • Profile page and question list re-designs - I absolutely support getting rid of tech debt and modernizing the design, but not at the cost of information density and utility. I managed to fix most of that with user scripts.
  • Prominently announced but poorly thought out features like collectives and discussions - they honestly had potential, but constructive community feedback was mostly ignored.
  • All that AI stuff that no one asked for, and that binds precious developer resources.
  • Cancelling the data dumps - I even understood the rationale for that, but the communication was catastrophic, and we still don't have the unified data dumps again, despite contrary promises. This does not so much get in my way, but it irrevocably broke trust.
  • The comment redesign - I like threaded comments, but the design is horribly inefficient w.r.t. to screen space and quick skimming. Highly upvoted comments take significantly more time to spot, and I need to do much more scrolling to get to the next post. All this was noted shortly after rollout, but of course the bug reports have been sitting unattended for months, while the feature is live. I don't have the energy to try to fix this via a user script. But it improved engagement metrics at least, as I now have to login each time just to get rid of this new view. (sorry for being cynical)
  • Most recently, native ads - I know you need money to keep the lights on, but this is just not the way.

On the other side, the Staging Ground and the outdated answers project/trending sort were rare silver linings, which unfortunately weren't fully concluded.

We tried complaining on meta, wrote dozens of well-conceived proposals, pointed out obvious flaws in the company's agenda -- but it was mostly met with silence. Instead, yet another broken feature was pushed out, and long-standing conventions were broken.

This causes frustration, as it is so incredibly unnecessary. Many recent changes are objectively bad, and that was repeatedly pointed out, alongside constructive proposals for adjustments. At some point, the tolerance for that runs out. I may have to endure the very same problem in real-world politics right now, where well-thought out solutions are ignored in favor of (from a scientific perspective) nonsensical and actively harmful actions -- but I don't need to do this in my free time as well.

This really only left one option: Walk away and hope that the company may finally have an epiphany amid declining contributions. I did this more than a year ago.

I genuinely admire anyone who is still staying and keeping the site running, even if I don't have the capacity to do so at the moment. I'll still be around, if nothing else, for solving my day-to-day programming problems; I really treasure the site as the valuable resource it is, and I learn something new every other day, sometimes from decades old anwers. And maybe, if the situation improves one day and the company takes serious steps to make using the site a pleasant experience again, I can once again commit some of my free time and meaningfully contribute new questions and answers.

6
  • 20
    Yep! It's not AI that's killing Stack. It's several years of intentionally alienating the core community with monumentally bad decisions. All we can do now is walk away and watch the fireworks from a safe distance. - Sincerely, A fellow disenfranchised decade-old former Stack contributor :) Commented Jan 22 at 1:09
  • 4
    "We tried complaining on meta, wrote dozens of well-conceived proposals, pointed out obvious flaws in the company's agenda -- but it was mostly met with silence. Instead, yet another broken feature was pushed out, and long-standing conventions were broken." This has been the modus operandi by the company since long before 2019 even, all the way back to the "Documentation project" and various strange GUI changes before that. Here's me ranting about it back in 2018 when things had already started to go bad: meta.stackoverflow.com/a/373471/584518. Commented Jan 22 at 9:39
  • ("Ask a question wizard" eventually boiled down to "Staging ground") Commented Jan 22 at 9:39
  • 4
    Upvoted, but not using LLMs to solve problems is a personal choice. You can use LLMs to assist in solving problems and still remain fully logical and understand the reasoning as long as you have critical thinking. I do it all the time. I use LLMs for my research alongside SO, that doesn't change SO is the most reliable source for me just as for you. Commented Jan 22 at 9:43
  • 3
    @WeijunZhou I fully agree, I didn't want to imply a contradiction there. This is mostly by preference (and due to the obscurity of some frameworks I find myself using). I think that LLMs can be quite useful, but they are just another tool that sometimes gives good solutions, and sometimes doesn't. It is up to the user to learn to wield it properly. Commented Jan 22 at 10:02
  • "I never faced any problems when asking and answering questions. I do my own research and only ask when I am sufficiently sure that the solution is not trivially discovered (i.e., I don't waste the time of other experts)." --- yes, but it shouldn't be like the Soup Nazi on Seinfeld where to get soup you have to humbly present yourself and move diligently to the right without violating protocol. Commented Jan 28 at 18:08
34

This reads very poetically, I enjoyed the lecture. But as with many company updates, I don't see much of a vision or strategy for what's next. None of the people I know personally enjoy Stack Overflow for the community. They enjoy it because it's useful.

Now something else has taken the stage for "appearing to answer questions usefully". The question graph reminds me of Nokia after the iPhone took off. But they don't fare too bad currently:

Stock of nokia 1995 - 2025

With patience though, I believe you have a chance when this happens:

Vibe coders in 2030 watching 10s unskippable ads

8
  • 7
    I don't believe I should be the one to determine what comes next. This conversation needs to involve the entire community, and from what I’ve observed, it’s already happening. My goal is simply to let people know that we are here, and we believe in a tomorrow. I truly believe that Stack Overflow is valuable now and has the potential to become even more useful and useable in the future -- whatever form that may take. Commented Jan 20 at 19:24
  • 5
    I don't believe the conversation involves the entire community in a way that is useful... given every result of it thus far has been extremely negatively received. Commented Jan 20 at 20:14
  • 1
    @Bella_Blue True, didn't mean to make it sound that harsh, edited the post. Mainly wanted to show the analogy to Nokia and that everyone knows free version of ChatGPT will be ad ridden once VC is burned Commented Jan 20 at 20:18
  • 1
    +1 for "appearing" to answer questions usefully. LLMs too often respond with plausible bullshit. I visit SO because here incorrect answers [usually] don't get upvoted. Commented Jan 20 at 23:03
  • 3
    The Nokia comparison may be reassuring for investors - the company succeeded in reorienting its business after its most famous division was sold off - but less so for users. Nokia Corporation has not made a phone since 2014, although it makes some royalties from HMD Global who license the "Nokia" brand; presumably less now that they've switched to using "HMD" as a public brand for their smartphones. Commented Jan 21 at 10:29
  • Hopefully the ads-based style of internet finally dies until 2030 too and people will just pay 5-10$ a month for an ad free assistant... Commented Jan 22 at 15:26
  • If the revenue graph distinguished between the "phone business" and other (telecom network) businesses then it'd argue against the point you're trying to make. Commented Jan 22 at 17:40
  • 2
    @Kashyap SO could change their business model as well, instead of doing the exact same thing for the next 20 years. Nokia needs to build hardware someone wants to buy, no need for it to be phones. SO needs to produce content someone wants to view. I only hope the company knows how valuable of an asset their community is to generate such content Commented Jan 23 at 7:21
32

Apparently I've been on StackOverflow long enough that my account is as old as User #1's, and I remember listening to the podcasts before the private beta was even announced. So with that being said, if you really want to develop StackOverflow - and associated sites - as a community, Stack Exchange, Inc really needs to be backing off the push for AI. Ideally the company would be spinning off StackOverfow into something like Wikimedia Foundations 501(c)(3) non-profit since that would decouple the Q&A from the business side of things. Otherwise you have a for-profit enterprise that is trying to find a way of monetizing what is effectively volunteer work.

2
  • 12
    "(...) monetizing what is effectively volunteer work" - absolutely on point Commented Jan 22 at 7:18
  • 2
    spot-on. The question I asked myself when one of the first really personally impactful dramas happened (the pronouns one), was "why am I doing volunteer work that brings profit to someone that's not me if I'm not even having fun doing so, but instead am constantly sad and frustrated, and I'm not getting anything in return?" I was answering other people's questions, but nobody was answering mine. I disassociated, and instantly got happier. Been like that ever since, SO is now the 4chan/b of the programming world - just old"timer"s doing a circlepat for old times' sake. Commented yesterday
31

Tagging this is spot on here. People heard those stories before. People heard those promises before. People heard those pep talks before.
People believed this. People supported this. People lived this.
And yet here we are.

You've had your chance to create something new.
You've had your chance to embrace a fresh start, a bold reimagining.
You've had your chance to take the lessons of the past and shape a brighter future.
You've had your chance for a future where the community not only survives but thrives.

You've had your time to create something extraordinary.

And that’s why this matters. This place serves as a reminder of what it means to squander the effort of humans. It’s not always easy to accept this, and it’s not always perfect to accept this, but it’s real. And that’s why loads and loads of people have given up. Because what happens here matters. It matters to the people who come here looking for answers, and it matters to the people who stay to share them. This place is a testament to the power of human connection, and that it needs fighting for – not empty phrases.

Too bad SO Inc has shown time and time again they don't consider this effort, this connection, these people worth believing in.

16
  • 2
    I'd say that all of you are our most vital asset. As we move forward, you all are going to be our strongest resource to invest in, believe in, and rely upon. We've relied upon you for so long now that it looks like we take you for granted, but I need you to understand that this team believes you all to be of paramount importance for our future. How we tap into your utility is going to change in the coming months, and my hope is that you all still believe in the future of this platform enough to continue to devote your time to us. It's too much to ask, yet we will ask it of you. Commented Jan 22 at 19:02
  • 24
    @Spevacus #WeAreNotYourAsset Commented Jan 22 at 19:10
  • 15
    @Spevacus What a presumptuous thing to say. We are not your "asset". We are not your "resource". We are not here to provide you "utility" to tap into. It says a lot that you know this is too much to ask, yet you feel entitled to ask it of us anyways. Commented Jan 22 at 19:15
  • 9
    I really don't like to kick someone while they are down-- c'mon.. they are here to do their job and basically are on their knees. Don't do this... As we are not the companies asset, they are not the company. Commented Jan 22 at 19:17
  • 10
    What I'm trying to say is that the platform's future is dependent upon your willingness to engage. I'm not trying to say that we're entitled to it. That distinction matters, and I'd probably blame comment length limits on not being able to make it, so I apologize for that. Further, I'm not asking for you to keep contributing purely on faith. The history of trust issues, which is not at all lost on me given my participation prior to being hired, is long and difficult. I know that I can't simply ask for your trust or continued participation and that we have to work hard to earn it. Commented Jan 22 at 19:23
  • 2
    Well the company did much in a top down hierarchical way and now that there is not just outrage by the community but other effects taking place, they once again coming with we will be more down to top hierarchical as they said in the past. The only issue was to not phrase it directly to the company, so it had the same bad taste as the initial dehumanized asset. Commented Jan 22 at 21:29
  • @MisterMiyagi re: "empty words" and "delivering", 3, 1. ??? Commented Jan 23 at 9:09
  • @starball Okay, since moderators have decided that word dump is meaningful conversation: What exactly are you trying to tell me? That I need to qualify every "you" as "you (the agents of this company and their apparent actions)" or somesuch? That the grand solution to the decline of SO and to reestablish trust by the community and showing the valuation of its assets… is doing standard housekeeping? Something else? Commented Jan 25 at 7:53
  • 1
    all I've done is point out that spev's words aren't empty, and that he and the staff he works with already deliver on meaningful improvements to issues that matter to us. to say in a comment directed at him "how about you stop talking empty words until you start delivering?" doesn't make sense. Commented Jan 25 at 9:24
  • 1
    and yes, I do believe that it aids clarity to the reader if you qualify in a post responding to a company announcement, or an @reply to a particular staff whether "you" is intended to refer to the staff making the announcement or to the company. for that reason, I'd personally encourage it (though I may not be a consistent example of it, and for that, I'm trying to change). Commented Jan 25 at 9:34
  • @starball "all I've done is…" No, you haven’t. You really, really haven’t. You’ve dumped a couple of out-of-context words and followed up with a reference that has zero meaning for me. Commented Jan 25 at 11:25
  • So if you want to encourage proper references – that is great! But then please actually do so! Personally, I think it’s absolutely clear what I am referring to since this entire Q&A here is tagged company-update and I explicitly referenced that. But if it’s not – just say so! Commented Jan 25 at 11:28
  • 3
    So to get back to delivering… really, that’s it? Spam is SO's problem. That they offloaded this to users as free labour is a complete failure on their part in the first place, not a reason for praise that they finally stopped that. No, seriously, selling this as something net-positive is indicative how seriously damaged the relationship to users is. This is abusive relationship territory, expecting people to be grateful for not being exploited quite as much anymore. Commented Jan 25 at 11:43
  • 1
    @starball pretty tone-deaf to point to spam detection as a great thing SE is delivering for us. As a user, spam on SO was never a problem thanks to the community; during my 14 years here I'd estimate I saw a low 2 digit nr of spam posts on SO itself, all of which were gone in a flash. The only place spam was an issue was discussions and that was a huge SE own goal. So, good job building a detector I guess, but "native ads" are more annoying + problematic than spam, and in general I couldn't care less about spam while SE keeps making everything else worse with various half-broken experiments. Commented Jan 26 at 10:03
  • 1
    Spam on SO has always been handled by volunteers and devs who built a tool to do what the company was too incompetent to do themselves. Spam handling is MVP. Are we celebrating that the site is finally ready to be launched on the Internet...? Commented Jan 28 at 7:33
27

It’s a community

There's the problem, right there.

When this site was buzzing, I got many answers from the likes of Hans Passant and John Skeet. Do you think they need to be part of a community? No way. We had access to rock stars in the field, and quality answers as a result. Lone wolves who racked up gold badges without even trying, just based on the quality of their answers. Some of those guys were pretty blunt, I would say even hurt a few feelings. Takes me back...

They are gone. The high performers aren't interested in community. There's nothing egalitarian about wanting to craft a great answer just to see the updoots.

You can take your "community". Do I go to a library to join a book club, or to gain knowledge in a quiet corner alone? Maybe some need a support group. The good ones don't.

Atlas has shrugged.

9
  • 3
    I think SO Inc. needs to think hard about how to get new experts to start answering questions. The current rep and badge systems don't really make sense anymore given the current rate of questions and general traffic to the site. Because let's face it, many answer questions partly because they want updoots and badges. Commented Jan 21 at 18:28
  • 3
    @cottontail In all fairness, there's not really a good model short of actually paying people to answer the questions and even then, it's really hard to sustain that since you might have run into the trap of seeing the same question on a daily basis. In contrast, when I teach a class, I can see the students develop over the weeks I spend with them which is quite rewarding to see. Commented Jan 21 at 19:24
  • 6
    These people became high performers and become known because of Stack Overflow. When Stack overflow was launched it was way easier to get reputation or get gold badges, so many people were farming it which is fine and they stayed for sometime, but as I said these people weren't known, they got known from the community and it's just pure luck. Why aren't people answering? Because most of the basic questions are answered already, most questions for sometime now require some sort of debugging which no one is going to do for free Commented Jan 22 at 14:33
  • 3
    @cottontail Many individuals who reached 100k+ reputation were answering for badges, rep and getting known including the names written in the answer here. The gaming behind Stack overflow is exactly how it got famous and it's exactly why people were answering, but now most if not all basic questions are answered and it's mostly questions that require debugging and no one wants to answer these for one upvote. Commented Jan 22 at 14:39
  • 2
    I've learned a lot over the years from comments from other experts on my answers, and from their answers to other questions I didn't know the answers to. When I think of the community in the tags I follow (mostly assembly / cpu-architecture / SIMD), that's what I think of. I'm very capable of learning from docs and manuals, but I'm not going to take the time to read everything, especially about platforms I'm not as interested in, so other people correcting my wrong generalizations and similar things have been super useful. But yes, a lot of the time I can just write a good answer alone. Commented Jan 23 at 20:43
  • 3
    I think this completely misunderstands the word “community” in this context and name checking high rep users would have some meaning if y it could quote them giving the sentiment you attribute to them. Commented Jan 24 at 8:58
  • @PhilipCouling I didn't really feel a sense of community when times were good. I'm very aware of a community now and all of its colorful members. I don't want to be a part of this community, and I'm sure I'm not the only one. Commented Jan 26 at 18:26
  • 2
    We are not talking about community in terms of all sitting around the campfire singing together! It doesn’t matter if you “feel” part of the community, it exists either way. A large group of people all coming with their own interests, their own goal, their own capability, their own drivers… that’s a community whether you feel part of it or not. There is no way a site like SO could work without it. Wanting the high performers back is like wanting a new engine in a car with no wheels, gearbox, or fuel tank. It’s not the feel of the thing that matters. Community is required to function. Commented Jan 26 at 19:26
  • 1
    Although Hans Passant is gone, Jon Skeet is still active: stackoverflow.com/users/22656/… Commented 2 days ago
25

Stack Overflow is not a community. It is a business that is attempting to monetize some communitarian behaviors. This isn't necessarily exploitive: there can be a fair exchange of a technology platform in return for volunteer effort. However, for years, the business has done a stunningly terrible job of fostering the community it is trying to monetize, and perhaps we now see chickens coming home to roost.

One source of the terrible management of the community is a contradiction between two definitions of the community. One 'community' is the persistent, committed, collection of more-or-less experts who answer questions and curate. Another definition of 'community' is the potential mass of people who are new, learning, confused, or failing to do their homework. I won't assert that it's impossible to foster a communal sense that extends to both, but I will assert that the site mechanics were built for the former, and then the business has spent years flailing around trying to address the later.

When you read, for example, the academic papers on the experience of women using the site (e.g. https://dl.acm.org/doi/10.1145/2950290.2950331), you will find that the authors assume the second form of community. New askers expect to be welcomed, guided, assisted, and perhaps even aided in collaborating with others in the same situation.

That is not the design of the site. The design of the site is that askers deliver coherent, discrete, questions, and answerers answer them -- and that askers who don't have, or don't know how to make, these question get downvoted and closed. Askers hate it. Of course they hate it.

On the other hand, many of the members of the 'original' community are fiercely committed to a vision of 'building a library' -- which is at odds, again, with community #2. No one likes to be told: 'your question is (just) a duplicate.'

When people write that 'the community' is toxic or nonexistent, I claim that they are thinking about community #2. Community #1 exists and not especially toxic. There is an esprit de corps, if of late it has started to have an air of Morituri te salutant.

The business has responded to this situation with:

  • browbeating the answerers and curators
  • ineffective, half-cooked, attempts to adjust tooling to allow for welcoming, guiding, etc.
  • browbeating the answerers and curators
  • browbeating the answerers and curators (repetition intentional)

To succeed in any attempt at the second bullet, the business would have to do three things:

  1. Spend the necessary money on the necessary developers (or AI tokens, I don't care) to actually design, implement, and iteratively refine something like 'Staging Ground' or 'discussion questions.'

  2. Spend the necessary money (blah blah) on the site mechanics to allow the curation they want to take place with less stress, so that the curators and answerers don't feel exhausted and overwhelmed by the constant effort to pull weeds. This might be the same thing as #1, if indeed, as claimed by some SO staff, many of the 'weeds' can be pruned into flowers.

  3. The hard part: the business must find some way to construct an alternative form of community #1 that is actively willing and interested in providing the experience that community #2 is looking for. Continuing to try to pound the square peg of the Q&A community into the round hole of a hypothetical mutual assistance site is going to continue to fail. And, perhaps to state the obvious, the experience of reading posts like this question, juxtaposed with the sporadic pounding, reads as gaslighting.

All this happens in a context of a series of truly epic faux pas over the years, in which the business actively attacked community #1, and/or adopted policies or practices which were seen as inappropriate or hostile by significant swaths of it. I'll be specific: for me, Monica still looms over the landscape. For others, it's LLMs or licensing questions.

So enough of the long posts from SO staff. Let's see some small, concrete, progress, in any good direction, and let's see it sustained and iterated.

6
  • 1
    Community #3: experienced questioners who ask hard questions. I used to think Community #1 included both questioners and answerers (it was in 2008-2010), but there is definitely a divide between the "curators" and the rest of people who engage in Q&A. I lost my motivation years ago for trying to answer SO questions because of this divide, so I pop up now and then to ask my questions and accept that some of them will get closed/downvoted by the curators, because there is some chance that I might receive useful feedback. It's not an encouraging community anymore. Commented Jan 28 at 16:49
  • Yes, I have observed this effect, but my answer is long enough already. Commented Jan 28 at 17:07
  • 1
    Agreed, but I think it's significant. What I remember from 2008-2010 is that there were lots of people asking good questions, not because they were good questions to curate, but just because people needed help, and kind souls like Jon Skeet answered them, and the rest of us followed by example. Commented Jan 28 at 18:15
  • 1
    I completely agree. Commented Jan 29 at 0:17
  • The second community is the soil from which the first community grows. Commented yesterday
  • 2
    A noble theory that has never been true in practice, as per Jason S's remarks and my experience. It is very rare for the sort of people who are non-experts to join community #1. SO started as a sort of mutual-aid club of experts (I might be an expert on some thing, but need help on others) and the bad press started when the complete non-experts tried to join in. Commented yesterday
23

Ónen i-Estel Edain, ��-chebin estel anim

I do believe that the network can and is worth saving but it's worth considering not just that this and other graphs are used to celebrate the downfall of SO - and frankly often by toxic folks but very few people have the energy to counter that narrative.

We have a community that's exhausted, beaten down, and constantly told 'there's a plan, trust us' - often going opposite to what we ask for, or need.

The network is ailing, and while it has strong bones - we've had a fairly long period where we've been in survival mode, and many decisions have alienated the community.

We're never going to win over the folks for whom "SO mods close everything as dupes" is both a meme, and an article of faith. But losing people who see a problem and go "OOH SHINY" or understand what sharks means hurts us.

The platform and changes to the platform, both cultural and mechanical need to take into account the needs of the folk using it as much as the folks who might use it.

And while we have people who'd still fight for this place - a lot of us are tired and could use some good news, and some respite. I'd reflect on that, and the why we're where we are - if we're to consider how we are going to get where we need to be.

6
  • 8
    As an aside, I'd say to an extent, telling folks AI is inevitable and we need to change our ways to it, and talking about how the human factor is.. a little tonal whiplashy. Commented Jan 22 at 16:14
  • 5
    What i find so frustrating is every time the conversation about how stack is using/pushing ai, any criticism of it is thrown aside as "You just hate AI!" and... it's frustrating. I hate how businesses abuse AI and use it in ways that aren't better than "the old ways". I like when i see it used in a way that is actually better than what it's replacing. Stack has done nothing but spin their wheels with solutions that don't do anything useful or do it in a low value way. Commented Jan 22 at 16:54
  • 7
    tbf, that's basically what every AI fanboy says. "Its the future! It is inevitable!" . It may not be the hammer we need to turn our bolts Commented Jan 23 at 3:09
  • 4
    I was going to upvote this. Until I saw the "frankly often by toxic folks". I disagree. Those aren't "toxic" trolls hating the site "for the lol". Those are exasperated people that face the despair of seeing that they helped build destroyed and distorted little by little. I doubt the network is so popular to be seen as trolling material by random internet hater that never cared or interacted with it. Whatever "toxic hater" we have were created by the actions the company took over the last years. Commented Jan 28 at 9:43
  • 2
    You'd be surprised - quite a lot of the hate outside is aimed at 'moderators' and curation actions, and very little is about the company and actions here unless they're current or former SE users. Quite a lot of the hate is from people who heard from people who heard from people. I've mentioned this specifically a little further on. Commented Jan 28 at 9:55
  • Also spam posts with "downfall graph" always looked noticeably sloppish to me. I.e. it wouldn't be surprising if it will be discovered that medium.com was regularly ordering posts in Twitter and Reddit about SO fall from 2020 to 2024 as part of their marketing strategy. They were the main competitor of SO in Google if to exclude AI. Commented Jan 28 at 13:27
21

Stack Overflow, the community, has been progressively killed off by Stack Overflow, the company, for years now. It should be of no surprise that yes... it's done.

We, the community, have all but screamed about it here in meta for years. We have protested everything from bad UX tweaks to improper moderator interventions. Most importantly, we have tried desperately to refocus the attention on what really matters.

You, the corporation, should have spent more time seeing yourselves as the facilitators. You took all of our human-to-human efforts in raising up our co-workers, peers, friends, in this industry... and you tried to package it up as some product. Instead of realizing your small-but-critical role in this community, you failed to understand, took out a ton of debt, and tried to squeeze it out of us. You killed it.

I've spent about 17 years here, along with the thousands of others, learning and teaching. Stack Overflow has had the most significant impact on my career and beyond. I'm afraid it's too late now to pivot to a new model. We've proposed on several occasions to make this a truly community-ran non-profit, but have failed to get anything serious off the ground.

I don't know where we go from here, but I'm afraid we've come to a dead-end on this path.

2
  • So my take on this only slightly less pessimistic. I think it is possible to carve out a shadow of what SO was in 2008-2010, by removing it from the "curation" mandate and focusing on community-building. IDK how you differentiate those kinds of questions, though. Commented Jan 28 at 16:52
  • @JasonS How would we attract a critical mass of folks to use such a site? Would people go there if it had 1/1000th of the usage of Stack Overflow, or less? Commented yesterday
20

I know things aren’t perfect right now. The once-unrelenting tidal wave of knowledge sharing has slowed to a trickle, and it feels like the world is moving faster than ever, leaving places like this behind.

I agree; falling site traffic will not kill Stack Overflow. Stack Overflow has a huge archive of high-quality answers; people could stop posting right now and I would still be coming here for answers for the next decade. Stack Overflow will not die unless it is killed.

We have better than something new: We have something old. All problems eventually come back to first principles, and Stack Overflow's huge archive of knowledge addresses an extremely basic problem. LLMs cannot replace Stack Overflow, but only pipe information from it; if they somehow choke it off, they will have pinched off their own airline of knowledge.

4
  • 3
    All problems eventually come back to first principles - that's a bingo! That's something that was even mentioned on one of the podcases back in '08 and why there was discussing capturing the "long tail" of programming problems. Commented Jan 21 at 19:28
  • 4
    How do you propose to finance the servers and make the site a meaningful business proposition with this mass of knowledge? The most willing buyer would have been the AI companies, but they already have the data unfortunately... Commented Jan 22 at 15:14
  • @AlexGeorg SO makes enough now to keep the servers up, right? Justifying the two billion dollars is harder. Maybe sue the AI companies for failing to attribute cited code as required by the Creative Commons license? Commented Jan 22 at 18:39
  • In my opinion, the maximal profit comes from cutting losses, handing management to the mods, and allowing SE to operate hands-off. If Prosus spends money on changes for growth, I don't think they'll get it back; SE already has everything its users want. A small positive figure beats a negative one. Commented Jan 22 at 18:52
20

A community is a group of people who share the same goals, interests, or mission. It's a place where individuals come together to contribute to the whole.

Why would I answer people's questions here instead of spending my time helping people elsewhere? That's a sincere question, not rhetorical negativity. Sell me on the company's vision for the community keeping in mind that I'm an engineer and literally have trouble understanding fluffy, aspirational language.

The company should communicate an explicit mission to the community so that we can decide if we want to be part of it or not. I don't believe the company when it says the core mission hasn't changed, because its actions say otherwise.

What I do believe is that the company wants free (as in beer) human created and curated knowledge, which is like printing money in the AI age. On top of that, only the company and people who pay them get to use this human-curated knowledge in it's most accessible form. The community can't even get a full data dump for personal use. Even if we jump through all the hoops, the data is tainted with fabricated data

As I have said before, Stack Overflow has plenty of meat left on its bones, so you will be able to dine off its corpse for probably many years to come. That doesn't mean it's not dead. On the other hand, it may be possible to keep enough of an audience to make enough money to stay in business, but I doubt that business is going to be creating a library of knowledge for the world or look anything like what I signed up for 11 years ago. Which is fine, things change, but it's infuriating to get strung along because the company wants to have its cake and eat it too.

I am not anti-AI by the way. I am simply opposed to companies making money from my expertise and effort without compensating me in some way. Soft compensation is fine. I have no regrets about what I have already contributed to Stack Exchange. I traded my time for enjoyment, connecting with people and attempting to build a lasting knowledge resource. I don't contribute any more because I no longer believe the mission is to create a lasting public source of knowledge.

No, I will not answer low-effort questions just because chatbots now tell barely literate people everything they type is awesome. I will not feel bad about closing those questions no matter how hard someone cries about it or how much the company scolds me for not being more welcoming of help vampires. I will donate my time somewhere where it is more appreciated and it feels less like shoveling water.

9
  • 1
    Why would you go elsewhere, either? I can see just two reasons: auto-generated activity and closed community. Alternatives with auto-generated activity are not interesting to even bother with: they are exactly like SO, but with N fake upvotes and N auto-generated support comments on each post. Closed communities do have the advantage of not directly or indirectly sharing your knowledge for training databases, but is this enough or viable? Commented Jan 28 at 13:37
  • 6
    @halt9k I think you are making an assumption that I wouldn't contribute to training AI for any reason. I am pro-AI. I think it is going to change the world for the better. Anyone who wants my time needs to convince me why I should give it to them. My employer gets it because they pay for it, I love the work and feel it is important. Individuals get my time when it's not to onerous for me, because I like to help. I'm not going to donate my time to a company unless I have a stake or I agree with their mission. There are many more ways to spend my time than the hours I have to spend. Commented Jan 28 at 14:17
  • Fair elaboration, but in the end, it's still a question like "sharing my solution where: SO or Discord group about JavaScript or <10 other options> or none at all". And in this context, you make a point that there must be a mission stated and followed - again, fair. But, for example, "Discord group about JavaScript" doesn't even have a concept mission, does it? As soon as they realise they have valuable answers, who knows what will happen. What I mean, it's good to also mention context of classical (rollback comments UI) or state-of-the-art solution (pay for the best answers) of your answer. Commented Jan 28 at 15:22
  • 2
    @halt9k I don't understand what you're saying. There are many ways someone can convince me something is worth my time. Those things may or may not work for other people. My expectation is that the company will never again be something I want to donate my time to, which makes me a little sad, but that's life. I've provided a ton of feedback to the company both publicly on Meta and privately in the moderator-only internal stack over the last decade and it's fairly clear there's no bridging the gap between my interests and the company's. Commented Jan 28 at 16:48
  • 4
    I don't have a need to answer questions for strangers online. I could spend that time doing lyrics for MusixMatch, or learning how to use AI effectively, or sewing an art quilt, or volunteering at a local animal shelter. It's not a choice between answering a question here or on Reddit. Commented Jan 28 at 17:05
  • 1
    "spending my time helping people elsewhere" sounded like there could be better QA alternatives right now. You elaborated both unclear parts with edit and comment, thanks. Commented 2 days ago
  • 5
    I don't think many are for/against AI as such, but rather against the AI bubble hype. AI is a tool like any other. If we need a tool for something like spam detection, and someone suggests 'screwdriver', then I will be against that, because it is a nonsense suggestion. That doesn't mean that I hate screwdrivers. Similarly, if someone asks what tool to use to give quality technical answers about programming I would say "How about the technical Q&A site SO explicitly designed for such". But I will not support some strange person who rather suggests "AI!" or "screwdriver!" for that job. Commented 2 days ago
  • 2
    @Lundin The only reason I mentioned it was because I didn’t want people to confuse my objections to being expected to donate my time to the company’s dreams of making millions with AI with me objecting to AI in general. Commented 2 days ago
  • 1
    @halt9k Ah I see now. When I said "helping people" I probably should've have been clearer I was thinking about volunteering my time in general. Commented 2 days ago
12

I've been on Stack Exchange for a while now and have tried my best to ask good questions, and I think at that I've been better than the average question-asking account (yes, accounts, since this is often gamed), which is to say I'm no newbie expecting to do minimal research and have basic questions answered like we're on Reddit. The effort needed to ask a "good" question and then be prepared to defend it while the "community" downvotes it and throws every close reason it can until one sticks (let's not get into how often that community actually upvotes good questions) used to be worth it just to get a high-quality answer, but since the advent of LLMs that's no longer true, and I find myself asking most questions to them nowadays, and saving myself a lot of time and headache as a result.

SO's value was never its "community", which is arguably the worst thing about it. This is why the company over the years has tried its best to reform it into the more "empathetic" community you mention in your question, precisely because it has long had the reputation of being the complete opposite - cold, clinical and only sometimes logical. See my last Meta question for an example of how selectively SO's rules are applied and how much they hinge on the biases and whims of the most active power users in the "community"... and that's assuming a user even gets to the point of caring enough about a closed question to open a Meta question for it.

SO's only real value for the vast majority of programmers in the world was its curated, high-quality code, which the LLM companies ran away with while cutting SO out of the equation, and now programmers are fleeing from it in droves because they no longer need to put themselves through the same old masochistic, bureaucratic rituals to get the same code. I don't know whether SO has a future as LLMs only get better and the window of questions that can only be answered on SO gets smaller and smaller, but at least in its current form, I see it as highly unlikely.

6
  • 16
    It's not for nothing that this process is a close mirror of academia. It is easy to ask questions, but it doesn't take long to learn the importance of asking "good" questions. I think understanding will still have a place in the jungle of LLMs. Commented Jan 20 at 22:50
  • 9
    This is probably spot on. Stubbornly refusing to accept that there isn't a culture problem and that the voting models don't work at all will be the death of the site and very soon - it might already be dead. However, we should also note that every single programmer forum that ever existed had lots of culture problems - they were always full of asocial types with serious lack of empathy issues. Commented Jan 21 at 7:31
  • 2
    The rules are being applied around a consistent basis, with some freedom to fluctuate around it due to the behaviors of individual users. Some questions are handled using a more strict criterion and got deleted, some others got away with it. You don't see the deleted one, you see the one that are left and it's a survivorship bias. Asking about your question on meta helps push the curation criterion close to the baseline that it should be. Using similar questions (that involves a survivorship bias) to justify your own question never works on meta. Commented Jan 22 at 4:56
  • It is infeasible to have a tight bind around the standard criterion, but it's possible to have a way to challenge it when the criterion is applied in a way that drifts too far away from the baseline, and that's the mechanism your last meta post resorted to. People looked at it and decided that according to the baseline it's not worth reopening. Note that I'm not claiming there is not a cultural problem with SO. Commented Jan 22 at 4:59
  • "let's not get into how often that community actually upvotes good questions" - out of curiosity I checked my voting stats: 1813 upvotes, 332 downvotes (not sure how to filter by questions/answers, and not sure if downvotes count after question was deleted), so to address your doubts, I would say I upvote fairly often. I also found this post with voting stats, and it seems that most of the community actually upvotes rather than downvotes. Commented Jan 28 at 12:17
  • "SO's value was never its "community", which is arguably the worst thing about it." --- Hashim, it looks like you joined in 2012. Unfortunately you missed the few golden years at the beginning, when there was a community and the curation mandate was less important than the value to the person who asked the question. Commented Jan 28 at 17:00
11

This is one of those graphs that's used because it's easy to generate, not because it's a good representation of reality. There's a lot going on here that isn't obvious, and that makes it easy to jump to the wrong conclusions.

First of all, that's exactly what I would expect the graph to look like for a Q&A site. Every question that gets answered is a question that doesn't need to be asked again. A mathematician might describe it as the set of as-yet-unanswered questions is shrinking. This naturally leads to fewer questions that need to be asked, and thus fewer answers required. That's not a sign of decline, that's the whole point. This isn't a social media site where value is derived from new content. A 10-year-old tweet is practically worthless, but a 10-year-old answer on why you shouldn't parse XML with regular expressions is as valuable now as it was when it was written. There are answers I wrote 15+ years ago that still get upvotes and comments to this day. So many people over-focus on "engagement" and value as connected with new content creation. Those things are easy to measure, whereas the value associated with existing content is harder to measure. Reality doesn't always have easy metrics. It would be silly to criticize the Louvre museum as dead because they don't add new content often enough. There is tremendous value in what already exists.

New technologies will bring with them new question opportunities, which should "recharge the pool" so to speak. This was generally true at the beginning, when Stack Overflow was the site. However, we have grown over time to a network of over 100 sites. Many new technologies spawn their own dedicated sub-site. The chart only looks at "StackOverflow posts", so all of this other content is missed.

A high-level query like this is also incapable of measuring the most important things. Primarily, a database can't tell you anything about quality. If the numbers drop and the delta is mostly low-quality content, then that's a sign that the site is thriving. As moderation tools improved over time, we got better at identifying and blocking spammers, sockpuppets, and other malicious users, at discouraging duplicate questions, and at migrating content to other sites where it's more on topic. Now, it's impossible to judge the quality of content that was never posted, but our moderators and bots have done a tremendous amount of work over the years and I can't fathom how that work wouldn't have a measurable impact on the amount of low-quality content being posted. If the new content rate drops to 10% of the peak value but that content is 10x more likely to be high-quality, then the overall value of the site has increased tremendously. Nobody wants to dig through a dumpster to find the one morsel of knowledge that they need. It's the difference between Merriam-Webster's Dictionary and Urban Dictionary. Our value is in being the former, whereas over-focusing on new content creation is how you become the latter.

Why not focus on more meaningful things? For instance, what is the trend in the percentage of questions that receive an answer (or multiple answers)? That should be one of the most important metrics. Similarly, how many new users have a question answered, then go on to contribute more questions/answers? What is the average time it takes for a question to receive an answer that gets accepted? If someone can come here, post a question, and get an answer to it in a reasonable time, then the site is still very much alive and well.

6
  • 2
    Agreed. Even without the effects of GenAI, I expect the rate of good new questions to decline over time. Similarly, the number of new pages created per month on Wikipedia isn't as high these days as it was a decade ago. We (the company & the community) should be focusing on ways to improve the existing content. Make it easier to find the pearls. Identify obsolete questions & answers. Get rid of, or at least de-emphasise, unhelpful low-quality content. Etc. Commented Jan 24 at 16:47
  • You are focusing on a wrong graph. Look at ad revenue. Commented Jan 24 at 17:07
  • 5
    Technology is still evolving faster than you can blink, speed rather picking up than slowing down, and innovations generates questions. If a resource for finding help for questios is going down, that is not explained by need for answers going down in the world at the current time. Commented Jan 25 at 11:39
  • 1
    "that's exactly what I would expect the graph to look like for a Q&A site" Sure, if it was about constructing bridges or crafting bows and arrows, or anything else like that which humanity has been doing for thousands of years. But programming is a very young technology still taking its first baby steps in comparison. Like the slightly older but still young technology called electronics, it is very rapidly evolving in all manner of directions. I would not expect questions to stagnate but rather increase exponentially, just like the evolution of the technology that the questions are about. Commented 2 days ago
  • Also, there aren't really that many sub-sites about specific programming branches at all to explain that activity has moved elsewhere. And those sites which are such sites are also in decline. Commented 2 days ago
  • Lol. The 'group by' questions for javascript alone used to account for multiple questions a day a few years ago and even getting them closed as duplicate was impossible — "...but I'm grouping color names not animals". Commented 7 hours ago
10

Well I have been using SO for now more that 10 years, and IMHO there are at least 3 ways to use it to solve problems:

  1. use Google to make a research, and browse through articles from SO. It has always been my highest use, and is still the best way I know to solve a problem, because of the high quality of upvoted answers, and the number or questions already asked and solved
  2. after doing that research ask a precise question with some context and a summary or the current research: it is the best way to add new nice Q&A topics
  3. ask a question without any research and little effort

The problem is that 1 and 2 are fine for people educated to those researches, but are not what beginners do. And I do believe that there are many more beginners seeking for help, than old fashion educated ones. And question from 3rd category are not really well received on SO, simply because the users that accept to spend their time to answer questions on SO want to contribute to a nice quality Q&A site. So beginners just wanting a little help for solving a basic question quickly learn that AI are much more friendly.

This is not really a problem for users: educated ones can still use SO, and beginners can use AI which can make use of the Q&A database. But it is a real problem for the company owning the site, because the generated revenue directly depends on the number of direct accesses. So if the majority of users access the database through an AI the revenue cannot reach the expectations. I am sorry, but I cannot imagine a simple and clean solution. The best way would be that AI engines accept to pay something when they use information from a third party site but I do not believe that they want to go that way, and I am not sure that it would simply be possible, beyond a flat rate voluntary contribution.

8

This seems to be pure company rhetoric, gaslighting or bullshit bingo. A general's bold speech before sending the troops into certain oblivion.

How about:

  • Specific diagnosis (why did the usage/contribution fall)
  • Clear strategy (product/community changes, positioning vs. AI)
  • Concrete initiatives (which concrete changes are planned)
  • Metrics and targets (how many questions are needed to keep the watermark above 0; when will it be known whether there was success)
  • Timeline, ownership, resourcing (who's doing what till when)
7

"Seeing is believing."

Looks like SO Co. has its eyes closed. You don't "believe" this is "the end" because you don't want to believe. Wake up. If you don't want this to be the end, then actually pay attention and do work to revive the community you've been mistreating for the last decade. Stop whetting your tongues claiming you care and try caring.

New contributor
Bentley Carpenter is a new contributor to this site. Take care in asking for clarification, commenting, and answering. Check out our Code of Conduct.
5

In the end, just remember:

"I am the cause and effect of my problems." ~Anoop S. Rana

This is as applicable to SO as humans. The mess(or whatever we can call it) created by SO will be sorted by SO through the efforts of its members(community, mods, staff etc).

14
  • 4
    Specifically - every single person who posted/commented here should ask themselves: what quality input did I contribute with to the main Q&A of SO this week? This month? The past year? And an awkward silence falls. If you sit here on meta and have strong opinions about everything, yet do not actually contribute to the main Q&A, you are part of the problem, not part of the solution. I think step one to recover SO is to block everyone from access to meta if they had no participation on the main site during the past month. Commented Jan 21 at 7:56
  • @Lundin "block everyone from access to meta if they had no participation on the main site during the past month..." is way to harsh, imo. I think allowing access to meta regardless of what they've contributed to meta is fine. Sometimes new people(who haven't posted on SO ever) or people who haven't posted in last year or so(because they didn't need to) might suggest some things to improve upon. Also, I think making who upvoted/downvoted a post visible to all can be done to make the site more transparent. There's still no 2FA on SO(which should take priority above all imo). Commented Jan 21 at 8:13
  • 2
    @Richard It might be a one time thing. We need to get rid of all the meta dead weight that is just actively causing harm, close-voting everything etc but not actively and constructively contributing to the Q&A. Commented Jan 21 at 9:06
  • @Lundin Yeah, then that is ok but I still think the time limit should be more than 1 month(like 3-4 month and one time thing). Commented Jan 21 at 9:09
  • @Richard Why? Anyone can come up with some manner of question to ask. If someone can't even do that in 1 months time but post on meta at daily basis, they need to be suspended from meta. Commented Jan 21 at 9:11
  • @Lundin Anyone can but doesn't need to. If we fix a 1 month time frame then that will force people to ask bad questions(even worse than they currently ask), but that will just be a formality. Commented Jan 21 at 9:16
  • 3
    @Richard It will also force people to realize - hey look, this is the actual purpose of the site - technical Q&A. Not meta, not close votes, not flag handling, not editing, not review queues - those are all peripheral things of less importance than the actual site content. A lot of people seem to think that the purpose of SO is to cast close votes. Those people need to go away. Commented Jan 21 at 9:44
  • @Lundin My point is we need to mantain a balance between the two. We should not be too forceful. 1 month time limit for meta access and as a one time thing seems too limited time to me. I think, atleast 3 months should be ok. Note that I completely agree that some type of participation must be done, I just disagree on the time limit of 1 month. Commented Jan 21 at 13:54
  • 3
    @Lundin we need to get rid of all the SO dead weight that is just actively causing harm, posting bad questions and answers, upvoting or interacting with off topic posts instead of closing and downvoting them, etc. A lot of people seem to think that the purpose of SO is to be a free service where they can dump a vague task description and other people will do their work for them. Those people need to go away. See how this can go both ways? I'm not saying you're wrong that the content is core, but good content is what is needed, bad content is doing harm instead of helping. Commented Jan 21 at 15:18
  • 3
    @Lundin also, you still have a "Reinstate Monica" avatar. I don't need to explain to you the myriad reasons people can have to boycott the site and decide to not provide free work for a for-profit company with questionable decision making. You're saying the input of all of these people shoud be ignored? That sounds extremely short sighted and won't help to re-attract experts who left after being dissatisfied with the platform. Commented Jan 21 at 15:20
  • 3
    @l4mpi I have seen many high rep users answering question that have canonical dupe targets. I agree with you that there should be some action(small) against such things like we have for other things. Commented Jan 21 at 15:28
  • 1
    eh, i completely disagree there. The only problem that arrises due to a high rep user answering a dupe is when they don't also close it as a dupe. More answers existing isn't a real problem. Storage is cheap. Search (and AI) is improved by more recent answers existing if they are tuned to prefer more recent content. Commented Jan 22 at 18:17
  • @Lundin pushing people off meta won't change their close-voting behavior. I think that there is been a major shift in the site in the last 10 years of narrowing the acceptable range of questions -- and I don't mean 'boat programming.' I mean that we used to cheerfully answer substantive, objective, questions that did not lend themselves to 'minimal code examples.' I'd be happy to go back to that, but I have no idea how. Commented Jan 27 at 17:22
  • @Lundin and, by the way, I watch the front page. I rescue some questions from being pecked to death by ducks on staging ground. Once in a blue moon, I see something that I can answer. I have given up on asking. And, yes, I TvC slop. Commented Jan 27 at 17:24
5

Me neither I don't believe it's "the end", unless the company will deliberately "end" it. You mention

(...) graph that is making the rounds. The question posed.

but do we also have the graph of "useful questions posed"? I don't have the access to statistics, but my feeling is that (under the tags I observed)

  • a few years ago there were maybe 50-100 new questions per day, most of them downvoted and closed, among them most were downvoted and closed because of absolutely no research effort (no mcve, no expected input/output, sometimes even no explanation, just "I clicked something and I got error"); 5-10 questions that remained with positive score were good enough to get straightforward answers, upvotes, and stay on site
  • yesterday I believe saw ~10 new C++ questions (not sure how many might have gotten deleted), and today I only see 2 of them with negative score; around half of them got meaningful answers, and those that didn't at least got some meaningful suggestions in the comments

For me today's situation looks cleaner and without much difference in added value to the site.

The decline in number of useless questions is most likely caused by availability of LLM tools (e.g. ChatGPT) that can answer them straight away without forcing the user to do tremendous effort of googling the answer and processing available SO knowledge by himself. But I fail to see why would it be considered "bad" for SO.

1
  • 6
    For Fortran,where wo used to get, on average, a few questions per day, and many, probably most, remained, we now get almost zero per month. It can be pronounced officially dead now. OTOH the Fortran community Discourse is quite alive and well. Commented Jan 22 at 22:37
3

This isn’t just a website. It’s a community.

I think this might be why many people are going to AIs rather than sticking here.

AIs are cheerful and helpful even when you make a mistake.

Make a mistake here and you seem to be immediately surrounded by people who make you feel bad.

3
  • 1
    Who exactly feels bad? Maybe people that get offended and yell in the comments "that's not what I asked!" if you politely point to not using namespace std. Otherwise (at least on tags I observe) even the most noobie question, with answer easily found in 1st google result, is answered within minutes. Perhaps downvoted and closed, but answered nonetheless. Commented Jan 22 at 7:16
  • 7
    @pptaszni you are both right though. Yes, objectively the people "should" not feel bad. But they do. And thus not use SO anymore because they have an alternative now... Commented Jan 22 at 15:24
  • 3
    @pptaszni I think your response perfectly sums up the feeling. Commented Jan 24 at 6:48
3

Once upon a time a cute little shark sang a song as she left her current persona behind her.
Like you she told us that "this is not end". So we waited for her. We searched for her. And now we are with her again. Laughing. Playing. Being happy together.
Because we trusted her. We loved her.

You lost this long ago.
Most of us don't trust you anymore.
Most of us don't love the AI-centric network you are attempting to build.
Most of us are still here just because inertia keeps us here.

"Is he friend or is he foe?" the pony wonders. I can assure you... I am no friend." (MLP S4E25)

This is not the end. The end started long ago when the site mission changed to "make tons on $$$ by whatever way possible, people will provide us with content to sell no matter what".
This is just another spin in a long tantrum spiral, and you keep rowing towards the center.

Let's dance together my friends, Walpurgisnacht is comming

If you really want to embark in a desperate struggle to escape the maelstrom you willing entered, empty words about how you love "the community" won't cut it. Because there is no "community" anymore: there is only the "you" and the "us" you created by alienating your own users. We are not in the same boat anymore, it just happens that our old, rusty dinghy is being towed by your big ship as you sail towards the mouth of the kraken, thinking there is gold to be found there.

Well... let me tell you something.

Hic sunt dracones.

5
  • The hololive and MLP references had my mouse hovering over the downvote arrow, the Madoka image moved it back to neutral. The vibe-to-content ratio of this answer is not that great, but then again the same can be said for the OP... But to add to the references: "This isn't your community - you threw your community over the bridge!", just that in case of SE there's no HomuHomu running after it to get it back. Commented Jan 28 at 10:31
  • 1
    @l4mpi vote as you wish, I won't take it as an offense. As for the references, the first is just to show the difference between following someone VS being forced to share the same site. The second is just because I really like the "I am no friend" and the duality in how you can read that (who would be speaking there? the user or the company?). Commented Jan 28 at 10:47
  • As for the third one... I have my personal belief that Walpurgis itself is the essence of "going round and round to save the wrong thing" - Homura whole quest is basically "Mu" (en.wikipedia.org/wiki/Mu_(negative)) - she lost sight of how to really reach an "happy end" by pursuing the saving of just one person, something that she will never be able to do without saving everyone else. Just like the company thinks that small posts like this one will have any effect Commented Jan 28 at 10:50
  • 1
    I'm not entirely on board with that Walpurgis analogy; for starters Walpurgis is an inevitable external factor while SE's problems with the community are mostly selfmade (also, Homura did nothing wrong). Thus SE as Madoka throwing away the community's soul gem and then hugging its lifeless husk not understanding what went wrong seems like a more fitting metaphor. Not sure how QB fits into that analogy but I'll spend a bit more time procrastinating at work to think about it :D Commented Jan 28 at 11:29
  • 2
    I thought for sure that would be a much better known shark song Commented Jan 28 at 13:07
-1

A few (probably well known) issues that could be solved already.

  1. Outdated answers on often searched topics, outdated answers make SO look like a relic from the past.
    • Let curators vote to pin an answer and let curators vote to make questions subject to these votings.
  2. Join forces. Splitting programming related topics seems to be a bad idea.
    • Let's have it all on the same site, implement tabs for it. Most people never heard of other SE sites, besides SO. There are people loving these topics but the implementation of them is somewhat hiding what SE has to offer.
  3. New quality standards for closing and opening question has to be implemented.
    • The issues and opinions of the community are well known.
13
  • 5
    All users on Stack Overflow can now participate in chat Commented Jan 22 at 15:33
  • @samcarter_is_at_topanswers.xyz thanks I've missed that. Commented Jan 22 at 15:40
  • 5
    we should be carfull about how to do 2 but i think a reddit-esque approach (with sub-reddits/stacks you can early switch between and cross post) might be a viable route Commented Jan 22 at 15:43
  • 1
    @A-Tech I haven't put details in it, because people tend to downvote if they don't like one thing about it. For pinning I would recommend a community wiki for example. for joining I would suggest to show a sum of SE points on the profile and so on. These things can be discussed separate. The company just has to go from general to specific with these topic. We argue over specifics, while no general approach has been accepted. Then SE drops something on their on and the community doesn't like it. It's somewhat the same story everytime. Commented Jan 22 at 15:49
  • 1
    Having all of SE under SO means... we get submerged under the bigger sites. I think the solution to lack of visibility is... more visibility. We used to have active promotion efforts for smaller sites - like graduation swag and contests. That the smaller SE sites is entirely a result of SO being promoted to the exclusion of other sites, and prolonged neglect. Commented Jan 23 at 10:32
  • @JourneymanGeek you may know more about the history. However, having it all at the same place is just convenient. Stores in a mall make usually more money than an isolated one. Commented Jan 23 at 11:21
  • 2
    BTW: I don't think "having ALL of SE under SO".. Just the programming ones, CodeReview/SoftwareEngeneering and so on. These are more or less the same communities but split across sites. Commented Jan 23 at 11:47
  • 2
    @Thingamabobs a lot of the smaller dev centric sites split off cause they were a poor fit for SO. If memory serves Software Engineering failed once before they got it to work even, and was a way to try to mitigate 'on a boat' questions. Things are the way they are for a reason. Commented Jan 23 at 12:40
  • @JourneymanGeek sure they are not just happened. My point is just the circumstances have changed and we might want to adopt to that. It's no big change I'm advocating here. I just want to see everything in one filtered question feed where I can see everything available I'm interested in participating with the option of selection of course. No-one needs to be bothered by topics they don't like. I don't mind keeping the reputation system as is also. It's rather a flow change than building a new structure. Commented Jan 23 at 12:58
  • Ah alright, that was quick, good... But then hum, "opening question" is still not completely clear to me, do you you mean the workflow through the 'Ask Wizard' / 'Staging Ground' + Reviewing/Approving or the "new concept" with 'Workshop' + 'Archive' to open/ask questions on the Site, or rather "reopening questions" after they've been closed...? // (And "opening question" is also grammatically not correct or "Russian Penglish", also unclear if you mean "opening a question", or "opening questions"...) Commented Jan 23 at 13:51
  • @chivracq we have had discussions about renew opening and closing questions and that's my point on this topic of "it's not the end, it's a beginning". Commented Jan 23 at 14:07
  • 1
    The "outdated answers" one is the final nail in the coffin (there were oh so many before it). But to my mind solutions to that might have actually helped if they had come sooner. SO never expanded on the "Wiki" concept. It's own attempt of "community answers" was so half baked. Now if it had a way to copy a question to a wiki section where both question and answers became fully owned by the community rather than the OP... That might have helped ease the tension between being a QA site here to help and a knowledge base. Commented 2 days ago
  • 1
    With regard to point 2: I stopped using the gamedev SE years ago when a new moderator decided that acceptable questions there were a strict subset of general programming questions that would be acceptable on SO, where they would get more exposure. A huge amount of interesting and valuable discussion about what makes a good game was closed and deleted. Commented yesterday
-3

Traffic started falling off when moderators started treating Joel Spolsky's impractical ideals like scripture. I have always maintained that this site was more valuable and "welcoming" during its first five years when it was a de facto help desk. People asked and answered more questions when they weren't afraid of aggressive moderation. The larger user base produced a lower percentage but larger quantity of exactly the kind of content that the idealists consider righteous and pure Q&A. Frankly, it's astounding that the idealists can't seem to comprehend this.

I'm looking forward to the Archive feature. I hope you will consider my suggestion of allowing every user to curate their own archive of "good" content and, if they choose, to view the site through the lens of other users' curation. Since traffic is lower than it was in SO's first months of existence, now is the perfect opportunity to completely rethink the back end if necessary.

Meanwhile, the fastest and cheapest way to bring users back would be to ditch the "advice" feature with its terrible UX and spin up a new SE for experience-based Q&A. The haters can simply not look at it.

6
  • What "impractical ideals" more specifically? Spolsky seemed overall AFK. Commented yesterday
  • 3
    As for all the new horrible experiments, I don't think it is enough to just cancel them all. That's a good start, but someone ought to take responsibility for how very poorly every single experiment was implemented. We can't improve anything on the site unless someone gets to the bottom with how every single idea launched, good or bad, turns into a fiasco during the implementation. There's a very clear pattern of this happening over and over in the last couple of years, starting somewhere around Discussions and the new editor. There's also a pattern of silently abandoning all experiments. Commented yesterday
  • 2
    @Lundin Specifically, that it was ever remotely realistic to only allow rigidly defined "quality" Q&A and expect enough people to be interested in the project to create the desired library of quality content. Such a library was created, but this happened because the standards for Q&A were lower. As noted in other answers, the library that was created before the site peaked is still valuable but rapidly getting dated. As for the implementation of the experiments, suffice it to say that my theory is extremely politically incorrect. Commented yesterday
  • 3
    If anything, I think the era around 2014 marked a switch from quality to quantity. That is: go from programmer helping programmer towards programmer helping student. And that's later one of the main reasons for the major loss of users - the students are gone, since GenAI can answer day 1 beginner level questions without tripping. (As can books, it turns out.) Before 2014 we did have "user must demonstrate a minimum of knowledge about the topic". Commented yesterday
  • 1
    Anyway I think you are correct in that one of the major problems with the site was always zealous moderation, but how you came to the conclusion that Spolsky was somehow the mastermind behind that, I have no idea. Commented yesterday
  • @Lundin The zealous moderators often cite Spolsky as the reason the site operates how it does, and why the users have no "moral right" to question it. Commented yesterday
-7

I agree, and a comment from Ewan on a Software Engineering meta post sums it up best:

It's easy to get AI to fix your code for you now, or answer any objective question based on published documentation. The questions that are left are the more opinionated, human questions which are naturally not 100% settled. AI will avoid giving a direct answer to "which is best scrum or kanban" but you can bet humans have some strong opinions and experiences on those options

(emphasis, mine)

I think Q&A's days are numbered. I see sooooo many long-time members willing to die on the hill of Q&A, but the rest of humanity has simply moved on to a different hill; in fact, they were moving on to a different hill before AI. In 2008, we had nothing but opinions and very few facts on the Internet. Wikipedia and Stack Overflow largely filled that gap for the masses, but Stack Overflow comes with a toxic community; a community hyper-focused on quality content at the expense of the people contributing that content.

People who ask questions are people too, and too often I see them get treated as if they were the pre-AI versions of slop mills — churning out low quality content in an endless river of sewage expecting us experts to filter it into drinkable water. Oh boy, did we not know what was to come...

Artificial intelligence has replaced the knowledgebase. It has fulfilled the Stack Overflow prophesy of "Ask questions. Get answers." Human Q&A is obsolete, I'm afraid. The bar is too high compared to LLMs. You can ask AI a fantastically stupid question, and it will not berate you, and it provides a detailed and (simulated) thoughtful answer.

I've written a number of questions over the years on SO. Many of those questions were well-received, but it sometimes took me days of painful debugging and research to be able to communicate what the problem is, and half a workday to write those thoughts into a question that was definitively answerable by total strangers. That's a lot of labor by querents for a comparatively small payback; and so, the questions have stopped coming in.

I feel like we are standing on the Q&A hill to defend it, but nobody else seems to want it. We defend a hill surrounding by empty countryside with nary a soul in site willing to threaten it. The masses have congregated in the cities orbiting the tech giants who produce LLMs. Every once in a while, we get a stray soul who isn't satisfied with what AI produces, but they come up to our hill and are met with rattling sabers and torches; a nice warning sign to stay away. So they do.

As Ewan said, "The questions that are left are the more opinionated, human questions which are naturally not 100% settled." The messy ones. The ones with no right answer. The vague ones written by people who are plenty smart enough but lack the knowledge to formulate a Q&A post.

Let's not forget that many in the younger generation are dealing with AI and remnants of the pandemic where everyone worked from home. Those random situational conversations with colleagues didn't happen during the formative years of their careers. These are the holes left behind by AI and the pandemic. There's no dearth of questions out there; they just don't fit the Stack Overflow Q&A model.

So, what will we do atop our lonely hill? Dig in? Entrench ourselves against an enemy that doesn't exist, damning ourselves to chasing windmills for eternity until Stack Overflow goes broke?

11
  • 9
    It may be easy to get genAI to correctly answer basic programming questions, but from what I have seen it is not easy to get them to correctly answers that have even a moderate level of complexity (read: a level of competency that I would expect from someone with 6+ months working in a language/technology, or perhaps someone with 5+ years of experience as a programmer, but not with the language/technology in question). Commented Jan 21 at 17:12
  • 6
    So, naturally, I disagree with the premise of this argument and the quote it begins with. Commented Jan 21 at 17:12
  • 8
    I believe there is some truth to this, but at the end of the day, LLM's need sources. We can be a source or we can fold and just settle on being irrelevant. I don't believe the community is inherently toxic, we're just using the tools we were given. We have no better way of ensuring low quality content gets deleted and or isn't used as a source than preventing it from getting an answer currently. That isn't a toxic community problem. Commented Jan 21 at 17:22
  • 2
    @tylerh It's equally difficult to get them answered here as they would probably be closed as too broad. Commented Jan 21 at 20:15
  • 9
    "churning out low quality content in an endless river of sewage expecting us experts to filter it into drinkable water" That is certainly how it felt. So many users thought they could just post whatever and get a solution. No regard for our quality standards. It was like wadding through sewage. Commented Jan 21 at 20:17
  • LLMs are extremely good at asking beginner-level questions as asked. The problem is that beginners are not good at asking the question they should ask. Commented Jan 21 at 23:43
  • 8
    You've lost me at your first line. "I think Q&A's days are numbered. I see sooooo many long-time members willing to die on the hill of Q&A, but the rest of humanity has simply moved on to a different hill". This is AI-bro rhetoric. It follows the whole vibe that AI is the magical bullet that solves everything. The users here who are against the use of AI simply see AI's limitations, and are arguing to apply AI where it HELPS, not just everywhere. Commented Jan 22 at 8:27
  • 5
    Yeah AI is great. I just asked the latest and greatest ChatGPT 5 some very basic beginner-level electronics questions: demonstrate Ohm's Law, parallel resistance and voltage dividers - something that most students probably can answer. i.sstatic.net/65x9zG9B.png. Except... that's serial resistance, not parallel... Where is the evidence that it can even answer beginner-level questions? Commented Jan 23 at 10:16
  • @Lundin the LLM was clearly going for comedy. Battery with 2 plus poles, check. Nonsensical "V" circle (voltmeter?) in the top left diagram with only one connected pole and a "current" arrow not going through a cable, check. Two separate instances of R1 in one diagram, check. "V oui", check (what's "check" in french?). Nonsensical formulas, check - highlights: "I=" and "R=" lower left sharing a right side; "I=V/R=R" so I=R and R=V/R - seems legit; and that Vout term bot center that's several symbols short of being an expression. Confusing parallel and serial might be the least wrong part :D Commented Jan 26 at 16:12
  • 2
    Also, for an example closer to home: I tried the SO LLM for a small code review (~20 lines Java) last week for fun while waiting for a CI pipeline. I already had unit tests covering all cases so I knew the code was correct. It found three things "wrong" with the code; two were trivial but didn't apply (relating to empty input which is impossible in my case). The third "error" was a hallucination by the LLM, complaining about an unhandled case which was in fact handled. After pointing out it got that part wrong, the LLM concurred and generated an "apology". Surpremely useful! Commented Jan 26 at 16:26
  • 1
    BTW, have you ever tried to publish content on Wikipedia? IMHO both community work the same: they build a nice repository of high quality information (that AI companies do use at no cost...) that anybody can use. Most questions that are badly received here should simply not have been asked if the poster had done a minimal research... Simply you do not have to ask a question on Widipeidia (and anyway cannot...) Commented 2 days ago
-8

We all have seen the graph that is making the rounds.

We all have seen who immensely contributed to that willingly.

-12

This isn’t just a website. It’s a community.

It's not a social network (Can StackExchange network site(s) be considered "Social Media"?).

The purpose of this site is to build a library of detailed questions and answers. It should not matter who or what asked them or answered them, only that the quality is high.

And for that, AI is the future. People use AI fall all kinds of things, it will become more rather than less. AI can write questions and answers both, but needs human curation for quality. AI can not just improve the grammar and tone of content, also the structure and even the content. Even knowing LLMs hallucinate, it would be plain dumb to try to create any textual content in this day and age without some AI support.

Humans online will in the future not have the patience to write down questions and answers themselves. AI is much more efficient at that.

If we could check how many questions were typed into computers (mobile phones, etc.) in 2023 and 2025, I am sure we would see much more questions were typed in 2025. And some got great answer by AI directly, some got great answers via RAG, but some got bad or no answers. And those latter questions should go into StackExchange, as a global library of great questions (and answers).

The human incentive to ask questions by typing is greater than ever, but you wont get them if you keep AI out. The days of humans typing a lot to create online content are over.

People already write the following things with AI:

  • chat messages
  • newsletters
  • news articles
  • online dating messaging
  • job resumes
  • letters to the landlord, city council
  • scientific articles
  • books
  • software
  • obituaries
  • ...

Just on SE, when my CPU is overheating and I need some help, somehow asking by using an AI to write out the question is not allowed.

AIs of the future will be trained on AI output anyways, because generators gonna generate, and crawlers gonna crawl. And as it has been said, human question and answer quality is not pristine either.

The main issues for SE are changing the gamification system of SE sites which just don't work anymore once AI is allowed, and avoid drowning curation volunteers with duplicate or spammy content (and other stuff, like copyright, attribution...). It's not as simple as saying "AI content is allowed now", but "We will never allow AI" is not going to bring back traffic.

4
  • 1
    If you train a new AI model on AI generated text, then it will perform much worse. I think this is the company's reason behind forbidding it. About the community, I couldn't imagine worse thing nowadays than trying to find the small bug in the AI generated code. Commented Jan 22 at 16:02
  • 8
    @zerocukor287 No they wanted to spew as much AI-generated crap as possible all over the place, while at the same time sell out the posts as AI training data. Yes those are mutually exclusive - nobody accused them for being rational. The reason why we didn't get as much AI trash as possible was that the moderators of SO went on strike. meta.stackexchange.com/questions/389811/… Commented Jan 22 at 16:24
  • My bad. This is even funnier, thanks for bringing it up Commented Jan 22 at 16:27
  • 5
    "it would be plain dumb to try to create any textual content in this day and age without some AI support" - that's certainly an order in which words could be placed. Commented Jan 26 at 9:04
-15

Discussion: We must talk on how we 'd fix this. Reality: We are dead but we 're so stubborn we can't see it.

Discussion: There's a plan, trust us, we can make SO great again (c.f. Trump). Reality: There is no plan whatsoever, we 're just begging users not to go.

Discussion: SO is still useful for developers. Reality: We are still trying to find users to apply our superiority moderation abilities.

Discussion: We don't like your comments/answers/posts/ -> you are closed, banned, harassed, etc and so you will lose the benefits in our site and the privelege of asking questions. Reality: We do not give ^%^%$ anymore.

Reality: AI killed SO but not for the bad reason it kills other good-effort companies, communities, stores etc. If that was true, all of us good-faith developers would try to stay and support it. It killed SO for the good reason: SO became a bin conducted by people that thought they had superior power over the world. All of us wanted to leave, and now we just can - and did.

Rest in pieces.

P.S. just go to ChatGPT and ask it why SO is dead.

  • gatekeeping club
  • power trip mods
  • “duplicate”, “off-topic”, “needs more details”
  • downvotes instead of helping
  • punish the newbie

Instead of knowledge base, it turns out into a priesthood.

LOL

FINAL REALITY: We are just a bunch of 200+ rep users desperately waiting for one like me to downvote.

5
  • 4
    Downvotes on Meta mean something different than the main site. On the main site they mean that a post is off-topic, difficult to understand, or of poor quality. Here, it is used as disagreement. I have downvoted here, because I disagree (the whole premise is faulty). Commented yesterday
  • 2
    I put it to Franck (elsewhere on this page) that the underlying cause of complaints is that a subset of users do not like on-topic rules. They believed, erroneously, that the rules should be more loose, or that they should not exist at all. I hold it to be true that, had those people gotten their way, the site would have collapsed many years ago. Commented yesterday
  • 2
    In fact, if there are a million universes, and Stack Overflow collapsed in half of them, the complainers would have found fault with the other half a million. We do not know if "this is the end", but our quality standards have kept us going for this long. Commented yesterday
  • 3
    Sure, there a bunch of people on the site who could have been kinder. If there is a way they could have been coerced to be kinder, I don't know what that is. Would the complainants have had the company moderate the contributing userbase more aggressively in favour of newcomers? I suppose they might have that view, but I wonder if chasing established users away would have also caused the site's collapse much earlier. Commented yesterday
  • 1
    So, what is the solution? The complainants on Reddit and do not say. Moreover, those who would give up so easily have no interest in community-building; I wonder if they want to ask a question, demand an answer, and not contribute in return. I acknowledge those folks aren't going to change their mind, but hopefully I have explained why the network was right not to cater to them. Commented yesterday
-15

This isn’t just a website. It’s a community. A place where people come together not just to solve problems, but to share what they’ve learned, to connect, to help each other grow.

Stack Exchange's strength is its Q&A, not its community, because the Stack Exchange community is associated by many people with a toxic environment. There are tens of thousands of posts on Reddit (clickable posts), Quora (clickable posts), Hacker News (clickable posts) and other platforms discussing this. It is very challenging to convince people to come back to an environment they deemed negative/toxic. If Stack Exchange loses its Q&A edge (which I think it already did), then it is game over for Stack Exchange (aside from Stack Internal).

I know things aren’t perfect right now. The once-unrelenting tidal wave of knowledge sharing has slowed to a trickle

That's untrue, the wave of knowledge sharing has not slowed. On the contrary, genAI has massively boosted knowledge sharing, e.g. ChatGPT alone gets billions of messages per day, despite the fact that some websites such as Stack Exchange or Reddit try to keep human knowledge outside the reach of genAI users. The only difference is that now, knowledge sharing is human <-> AI <-> human instead of human <-> human. But since the Stack Exchange community has voted against the use of genAI, even despite the recent progress and massively increased user base of genAI tools, Stack Exchange has lost the opportunity to be part of that wave.

knowledge should be shared freely

I agree. So why is Stack Exchange trying to charge AI firms to use Stack Exchange's knowledge?

42
  • 8
    You might be right about the SO community. But isn't SO more on the other side of the AI chain--providing information to AIs rather than from AIs? Commented Jan 20 at 23:36
  • 7
    "because the Stack Exchange community is associated by many people with a toxic environment" I think it's also true for other platforms like Reddit (clickable posts) and Quora (clickable posts). The difference is that SE doesn't provide a place to openly discuss them here :) Commented Jan 21 at 1:49
  • 35
    How does a user privately messaging ChatGPT constitute "knowledge sharing"? At best, that's accessing previously shared knowledge, not sharing anything. Commented Jan 21 at 7:16
  • 3
    Crap websites or not, they're where many burnt SO users go after having a bad experience here. Quora has an awful presentation format that lists irrelevant content directly below relevant content - but from what I've seen there, contributors tend to be friendly and welcoming. Reddit is pretty hit-or-miss, depending on the sub. Reddit sentiment can be a useful meter, but these days I just ask ChatGPT to summarize the reddit sentiment regarding a product/service. SO/SE has a mostly good format, so it's sad to see when established community members or unwelcoming practices drive away newcomers. Commented Jan 21 at 7:56
  • 11
    FWIW users who complain about Stack Overflow's toxicity don't usually mean user behavior but rather that the Q&A format is toxic. And I can't understand those who complain about actual toxic user behavior because the flag system exist and we do delete even slightly unfriendly comments. It's a problem with a very easy solution. Commented Jan 21 at 12:32
  • 10
    "... the Stack Exchange community is associated by many people with a toxic environment. There are tens of thousands of posts on ..." The irony then that when I see others,or when I try myself, try to engage constructively with these communities, to understand their concerns (which often offer little substance on the why its toxic), such interactions are often met with toxic behaviour that largely goes unmoderated. These posts are rarely actually about the Stack Overflow, and more about that they are frustrated their question didn't meet Stack Overflow's quality expectations (which is mostly their fault). Commented Jan 21 at 12:36
  • 3
    @FranckDernoncourt That's not true and there are examples to prove otherwise, however this discussion is off topic for this post. I have deleted some comments here that stray away from what the main topic is. In general, posts or comments that are only inflammatory and don't contribute to the conversation are deleted. It's not censorship, it's janitorial. Commented Jan 21 at 12:36
  • 4
    I can't help but notice that all the comments supporting my answer got removed while all the comments against it are kept. Commented Jan 21 at 16:38
  • 4
    @rjzii After asking writing 10k+ posts on Stack Exchange over 15+ years I still consider Stack Exchange to be mostly a hostile place. I'm very happy AI can answer >90% of my questions nowadays. Commented Jan 21 at 19:38
  • 3
    The demand for "be even nicer" was always about a relaxing of the rules, but we have learned from history that unmoderated Q&A (Yahoo Answers, Quora, Experts Exchange) die a death eventually. This is because when the quality tanks, you get more beginners, and less experts. That very quickly becomes unsustainable. Commented Jan 21 at 21:36
  • 4
    @halfer Aside from biased moderation and inconsistent moderation, in one of my earlier comments that the moderation deleted, I was mentioning user content deletion. Beyond that there is targetted voting (such as down, votes to close, votes to delete), aggressive comments toward post author, etc Commented Jan 22 at 1:28
  • 7
    "all the comments supporting my answer got removed" For transparency, there are only four deleted comments on this post: 2 discuss the votes on the post, and so were deleted (comments on votes are almost always NLN), 1 listed similar posts and didn't add any meaning, and the last was unkind. None were in "support" of your post. Commented Jan 22 at 13:55
  • 7
    That is a comment about downvotes, @FranckDernoncourt , and was removed; from my prior comment "comments on votes are almost always NLN"; that certainly holds true for that comment. Commented Jan 22 at 14:09
  • 5
    I did, and I stand by that, @FranckDernoncourt . Denoting your post is getting downvotes isn't in support of your post, it's about the votes. Commented Jan 22 at 14:17
  • 4
    Franck, I'd like to respond to what you're saying, but the discussion is all over the shop. Your discourse in the comments has entirely veered away from the substance of your answer. It may be that your general thesis is that you can dance on the grave of Stack Overflow because it deserves it, and if so, have at it. But, as I say, your set of views do seem to be based on non-factual foundations. Commented Jan 22 at 20:14
-15

The game is over and the ideologists have won.

Not AI killed SO, it was it's own "community". Unable to see the obvious relationship between new-comers asking foolish question and point hungry experts decorating themselves with badges.

The only strong community on SO that will be left are the "curators" which won't have anything to curate anymore, after the foolish were forbidden and the experts hadn't anything to answer anymore.

Long ago, they should have made a wiki section on this site where the experts could have been immortal, the curators could have done their share to a better world and the foolish could have supplied more questions, inspirations and points.

Instead this will become a graveyard of outdated answers.

“The reasonable man adapts himself to the world: the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man.”

2
  • 3
    Did the 'ideologists' win tho ? One of the problems we've faced is losing a lot of folks who are active here, often due to choices the company made Commented 8 hours ago
  • @JourneymanGeek Sure they did, "No content is better than bad content". It's the same folks that weren't able to ignore the "bad questions" and wanted to be feed by hand, instead they just have had to cherry-pick. It's not like I miss the bad questions, don't get me wrong here, it's just you had a couple of interesting ones. Finding out useful things you are able to do, without ever planning to do them. That was the game, find the good ones, provide the best answer as fast as possible and see how it goes. Maybe SO's future will be answering Questions of AI-Bots they aren't able to answer. Commented 1 hour ago
1
2

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.