- This is a really remarkable graph. I just didn't realize how thoroughly it was over for SO. It stuns me as much as when Encyclopædia Britannica stopped selling print versions a mere 9 years after the publication of Wikipedia, but at an even faster timescale.
- I disagree with most comments that the brusque moderation is the cause of SO's problems, though it certainly didn't help. SO has had poor moderation from the beginning. The fundamental value proposition of SO is getting an answer to a question; if you can the same answer faster, you don't need SO. I suspect that the gradual decline, beginning around 2016, is due to growth in a number of other sources of answers. Reddit is kind of a dark horse here, as I began seeing answers on Google to more modern technical questions link to a Reddit thread frequently along with SO from 2016 onwards. I also suspect Discord played a part, though this is harder to gauge; I certainly got a number of answers to questions for, e.g., Bun, by asking around in the Bun Discord, etc. The final nail in the coffin is of course LLMs, which can offer a SO-level answer to a decent percentage of questions instantly. (The fact that the LLM doesn't insult you is just the cherry on top.)
- I know I'm beating a dead horse here, but what happens now? Despite stratification I mentioned above, SO was by far the leading source of high quality answers to technical questions. What do LLMs train off of now? I wonder if, 10 years from now, LLMs will still be answering questions that were answered in the halcyon 2014-2020 days of SO better than anything that came after? Or will we find new, better ways to find answers to technical questions?
> I disagree with most comments that the brusque moderation is the cause of SO's problems, though it certainly didn't help. SO has had poor moderation from the beginning.
I was an early SO user and I don’t agree with this.
The moderation was always there, but from my perspective it wasn’t until the site really pushed into branching out and expanding Stack Exchange across many topics to become a Quora style competitor that the moderation started taking on a life of its own. Stack Overflow moderator drama felt constant in the later 2010s with endless weird drama spilling across Twitter, Reddit, and the moderator’s personal blogs. That’s about the same time period where it felt like the moderation team was more interested in finding reasons to exercise their moderation power than in maintaining an interesting website.
Since about 2020 every time I click a Stack Overflow link I estimate there’s a 50/50 chance that the question I clicked on would be marked as off topic or closed or something before anyone could answer it. Between the moderator drama and the constant bait-and-switch feeling of clicking on SO links that didn’t go anywhere the site just felt more exhausting than helpful.
There was definitely a bit of Pournelle's Iron Law of Bureaucracy [0] at work. I worked there during a lot of the drama you allude to, and... It sucked, for everyone. But also...
For most of SO's history, the vast majority of visitors (and this questions, answers) came in via Google. Not "search engines"; Google. This was pretty much baked in right at the start, and it effectively served as the site's primary user interface for years. And it worked. It worked pretty well! Until it didn't.
At some point, Google started surfacing fewer "tried and true" Q&A examples and more unanswered, poorly-answered or moderated examples. This broke the fundamental assumption that sat behind SO's moderation - that curating a smaller set of posts was preferable to encouraging more, and newer. Suddenly, Google wasn't a very good UI for SO anymore.
...and SO didn't really have a fallback. Heck, for a while during this period they actually stopped showing questions on their homepage unless you were already logged in; the core assumption remained that there was a never-ending spring of new people feeding the thing. LLMs weren't the start of the problem, they were the end - the final wake-up call.
I don't know that a site like SO can exist without the old Google, the old Internet; it is a product of all that, in the same way that mass-market TV shows were a product of 20th-century broadcast technology, or trade paperbacks of a particular intersection of printing tech and reading habits.
I know the feeling of being happy not being the only one with that same problem (and that somebody bothered to actually ask on SO) and the crushing feeling that the question was closed as off topic (so no reason for me to ask) or marked as duplicate (referencing that is clearly not a duplicate and just showing that the mod took no effort to understand the question)
The moderation definitely got kind of nasty in the last 5 years or so. To the point where you would feel unwelcome for asking a question you had already researched, and felt was perfectly sound to ask. However, that didn't stop millions of people from asking questions every day, it just felt kinda shitty to those of us who spent more time answering, when we actually needed to ask one on a topic we were lacking in. (Speaking as someone who never moderated).
My feeling was always that the super mods were people who had too much time on their hands... and the site would've been better without them (speaking in the past tense, now). But I don't think that's what killed it. LLMs scraping all its content and recycling it into bite-sized Gemini or GPT answers - that's what killed it.
I asked a question for the first time mid last year. It was a question about "default" sizes in HTML layout calculations, with lots of research and links to relevant parts of the spec.
It was immediately closed as off topic, and there were a bunch of extremely vitriolic comments offended that I'd ask such a question on SO. It was briefly reopened weeks (?) later and then I guess closed again and now is deleted, so you can't even view the question any more.
I'd long heard of abusive moderation but... experiencing it first hand is something else. Anecdote of one, but I know I'm never going to ask there again.
In case anyone's wondering, I ended up asking on the WhatWG or W3C or something github project (via an issue?). The TLDR was rather eye opening, that basically the spec only codifies points of contention for browsers and old behaviors are generally undocumented. With some pointers I figured out the default size behavior through code diving, and it was complex (as in, hard to use) and very unintuitive.
Friend in my group was in the public beta back in '08. We all ended up signing up by the end of '09. I used it off-and-on over the years (have some questions and replies with hundreds of upvotes). Though SO had a rap for having what might seem like harsh replies or moderation, it was often imho just blunt/curt, to the point, and often objectively defensible. I also agree with your timeframe that, in the later 2010s, the site became infected with drama, and moderation suddenly started reaching its tendrils into non-technical areas, when it should not have. And on an ostensibly technical site, no less!
I found myself contributing less and less (same with Wikipedia), because I merely wanted to continue honing my craft through learning and contributing technical data with others who shared this same passion... I did not want to have politics shoved in my face, or have every post of mine have to be filtered through an increasingly extreme ideology which had nothing to do with the technical nature of the site. When I had my SO suspended with no warning or recourse for writing "master" in a reply, I knew it was time to leave for good. Most of the admins on the site transformed from technical (yet sometimes brash!) geeks, into political flag-waving and ideology-pushing avatars (including pushing their sexual agendas front and center), and not of the FSF/FLOSS kind, either.
These types of dramas have infected nearly everything online, especially since 2020. Even Linus has lost his mind with pushing politics into what should be purely technical areas https://news.ycombinator.com/item?id=41936049
LLMs were a final blow for many reasons, though I think that a huge part of it is that LLMs won't chide you and suspend/ban you for wanting to stick to strictly technical matters. I don't have to pledge allegiance to a particular ideology and pass a purity test before asking technical questions to an LLM.
Quite frankly you are wrong. Jeff and Joel spoke about their goals for very harsh moderation in their podcast while they were still building SO. The moderation from the very beginning was a direct result of the culture they created and it was completely intentional.
> The fundamental value proposition of SO is getting an answer to a question
I read an interview once with one of the founders of SO. They said the main value stackoverflow provided wasn't to the person who asked the question. It was for the person who googled it later and found the answer. This is why all the moderation pushes toward deleting duplicates of questions, and having a single accepted answer. They were primarily trying to make google searches more effective for the broader internet. Not provide a service for the question-asker or answerer.
> This is why all the moderation pushes toward deleting duplicates of questions, and having a single accepted answer.
My personal single biggest source of frustration with SO has been outdated answers that are locking out more modern and correct answers. There are so many things for which there is no permanently right answer over time. It feels like SO started solidifying and failed to do the moderation cleaning and maintenance needed to keep it current and thriving. The over-moderation you described helps people for a short time but then doesn’t help the person who googles much later. I’ve also constantly wished that bad answers would get hidden or cleaned out, and that accepted answers that weren’t very good would get more actively changed to better ones that showed up, it’s pretty common to see newer+better answers than the accepted one.
> This is why all the moderation pushes toward deleting duplicates of questions, and having a single accepted answer.
Having duplicates of the question is precisely why people use LLMs instead of StackOverflow. The majority of all users lack the vocabulary to properly articulate their problems using the jargon of mathematicians and programmers. Prior to LLMs, my use case for StackOverflow was something like this:
30 minutes trying (and failing) to use the right search terms to articulate the problem (remember, there was no contextual understanding, so if you used a word with two meanings and one of those meanings was more popular, you’d have to omit it using the exclusion operator).
30 minutes reading through the threads I found (half of which will have been closed or answered by users who ignored some condition presented by the OP).
5 minutes on implementation.
2 minutes pounding my head on my desk because it shouldn’t have been that hard.
With an LLM, if the problem has been documented at any point in the last 20 years, I can probably solve it using my initial prompt even as a layman. When you’d actually find an answer on StackOverflow, it was often only because you finally found a different way of phrasing your search so that a relevant result came up. Half the time the OP would describe the exact problem you were having only for the thread to be closed by moderators as a duplicate of another question that lacked one of your conditions.
I think that's a great policy. I don't think anyone wants duplicate questions. The problem is moderation marking unrelated questions as duplicates.
I'd say 9/10 times I find a direct match for my question on SO it's been closed as offtopic with links to one or more questions that are only superficially similar.
There are other problems that they don't even try to address. If 10 people ask the same question, why does only the first person to ask it get to choose the answer? Then lots of "XY" questions where the original asker didn't actually have problem X so selects an answer for Y, leaving the original X unsolved, and now all the duplicates only have an answer for Y too.
Sad? No. A good LLM is vastly better than SO ever was. An LLM won't close your question for being off-topic in the opinion of some people but not others. It won't flame you for failing to phrase your question optimally, or argue about exactly which site it should have been posted on. It won't "close as duplicate" because a vaguely-similar question was asked 10 years ago in a completely-different context (and never really got a great answer back then).
Moreover, the LLM has access to all instances of similar problems, while a human can only read one SO page at a time.
The question of what will replace SO in future models, though, is a valid one. People don't realize what a massive advantage Google has over everyone else in that regard. So many site owners go out of their way to try to block OpenAI's crawlers, while simultaneously trying to attract Google's.
Quite often, when my search returned a 'closed as duplicate' reply, I found the allegedly duplicate question did not accurately describe my problem, and the answers to it were often inferior, for my purposes, than those which had been given to my original question before the gate was closed.
The disconnect here is that they built it this way, but still call it a question and answer site and give a lot of power over to the person who created the question. They get to mark an answer as the solution for themselves, even if the people coming from Google have another answer as the solution.
If they were to recreate the site and frame it as a symptom and issue site, which is what the interview described, that would yield many different choices on how to navigate the site, and it would do a lot better. In particular, what happens when two different issues have the same symptom. Right now, that question is closed as a duplicate. Under a symptom and issue site, it's obvious that both should stay as distinct issues.
> They were primarily trying to make google searches more effective for the broader internet
This is mostly how I engaged with SO for a long, long time. I think it’s a testament to SO’s curation of answers that I didn’t ask almost any questions for like 5+ years after starting programming
If this were true, then treating any question as an X-Y problem shouldn't be allowed at all. I.e. answers should at least address the question as posed before/instead of proposing an alternative approach.
In reality the opposite is encouraged. For countless times, I've landed on questions with promising titles/search extracts, only to find irrelevant answers because people grabbed onto some detail in the question irrelevant to my case and provided X-Y answers.
This often also causes subsequent useful questions to be marked as dups even though they no longer contain that irrelevant detail. The appeal process is so unfriendly that most would not bother.
I agree with that and I think it was the right decision. There was grousing about overmoderation but I think a lot of people got unreasonably annoyed when their question was closed. And the result was a pretty well-curated and really useful knowledge base.
Thinking from first principles, a large part of the content on stack overflow comes from the practical experience and battle scars worn by developers sharing them with others and cross-curating approaches.
Privacy concerns notwithstanding, one could argue having LLMs with us every step of the way - coding agents, debugging, devops tools etc. It will be this shared interlocutor with vast swaths of experiential knowledge collected and redistributed at an even larger scale than SO and forum-style platforms allow for.
It does remove the human touch so it's quite a different dynamic and the amount of data to collect is staggering and challenging from a legal point of view, but I suspect a lot of the knowledge used to train LLMs in the next ten years will come from large-scale telemetry and millions of hours in RL self-play where LLMs learn to scale and debug code from fizzbuzz to facebook and twitter-like distributed system.
> Privacy concerns notwithstanding, one could argue having LLMs with us every step of the way - coding agents, debugging, devops tools etc.
That might work until an LLM encounters a question it's programmed to regard as suspicious for whatever reason. I recently wanted to exercise an SMTP server I've been configuring, and wanted to do it by an expect script, which I don't do regularly. Instead of digging through the docs, I asked Google's Gemini (whatever's the current free version) to write a bare bones script for an SMTP conversation.
It flatly refused.
The explanation was along the lines "it could be used for spamming, so I can't do that, Dave." I understand the motivation, and can even sympathize a bit, but what are the options for someone who has a legitimate need for an answer? I know how to get one by other means; what's the end game when it's LLMs all the way down? I certainly don't wish to live in such a world.
I don't know how others use LLMs, but once I find the answer to something I'm stuck on I do not tell the LLM that it's fixed. This was a problem in forums as well but I think even fewer people are going to give that feedback to a chatbot
No longer interacting with your peers but an LLM instead? The knowledge centralized via telemetry and spying on every user’s every interaction and only available thru a enshitified subscription to a model that’s been trained on this stolen data?
As long as software is properly documented, and documentation is published in LLM-friendly formats, LLMs may be able to answer most of the beyond basic questions even when docs don't explicitly cover a particular scenario.
Take an API for searching products, one for getting product details, and then an API for deleting a product.
The documentation does not need to cover the detailed scenario of "How to delete a product" where the first step is to search, the second step is to get the details (get the ID), and the third step is to delete.
The LLM is capable of answering the question "how to delete the product 'product name'".
To some degree, many of the questions on SO were beyond basic, but still possible for a human to answer if only they read documentation. LLMs just happen to be capable of reading A LOT of documentation a LOT faster, and then coming up with an answer A LOT faster.
If the LLM is also writing the documentation, because the developers surely don’t want to, I’m not sure how well this will work out.
I have some co-workers who have tried to use Copilot for their documentation (because they never write any and I’m constantly asking them questions as a result), and the results were so bad they actually spent the time to write proper documentation. It failed successfully, I suppose.
> I disagree with most comments that the brusque moderation is the cause of SO's problems
The moderation was precisely the reason I stopped using stackoverflow and started looking for answers and asking questions elsewhere. It was nearly impossible to ask anything without someone replying "Why would you even want to do that, do <something completely different that does not solve my problem> instead!". Or someone claiming it's a duplicate and you should use that ancient answer from another question that 1) barely fits and doesnt solve my problem and 2) is so outdated, it's no longer useful.
Whenever I had to ask something, I had to add a justification as to why I have to do it that way and why previous posts do not solve the issue, and that took more space than the question itself.
That's not moderation. That's a lack of moderation. Comments like this weren't posted by mods but by people looking for stuff to answer. A mod would definitely delete such a comment if they were notified about it. But not enough users flagged stuff and preferred to complain about being hurt instead. There's also a problem that not enough mods were present at all times to handle the volume of comments posted.
I will miss it but you are right about moderation. I don't know what the issue is on some platforms, reddit and SO come to mind. Moderators on many other platforms or forums seem to be alright and keep a clear head, even when they have to deal with a lot of vitriol and they get little thanks for their work.
There are probably negative examples as well but some platforms seem to be especially vulnerable. If I had to run reddit or SO, I would limit moderation to one subreddit/subdomain. No idea if that would help, but the problem isn't exactly invisible.
If we're going to diagnose pre-AI Stack Overflow problems I see two obvious ones:
1. The attempt to cut back on the harshness of moderation meant letting through more low-quality questions.
2. More importantly, a lot of the content is just stale. Like you go to some question and the accepted answer with the most votes is for a ten-year-old version of the technology.
> Like you go to some question and the accepted answer with the most votes is for a ten-year-old version of the technology.
This is still a problem with LLMs as a result. The bigger problem is that now the LLM doesn’t show you it was a 10 year old solution, you have to try it, watch it fail, then find out it’s old, and ask for a more up to date example, then watch it flounder around. I’ve experienced this more times than I can count.
> The fundamental value proposition of SO is getting an answer to a question
But the horrible moderation was in part a reason why many SO questions had no answers.
I am not saying poor moderation caused all of this, but it contributed negatively and many people were pissed at that and stopped using SO. It is not the only reason SO declined, but there are many reasons for SO failure after its peak days.
To the extent that moderation ever prevented questions from getting answers, that was by closing them.
When a question gets closed before an answer comes in, the OP has nine days to fix it before it gets deleted automatically by the system.
The value proposition is getting an answer to a question that is useful to a reasonably broad audience. That very often means a question that someone else asked, the answer to which is useful to you. It is not getting an "answer" to a "question" where an individual dumps some code trying to figure out what's wrong.
There was, obviously, only one main reason: LLMs. Anything else makes no sense. Even if the moderation was "horrible" (which sounds to me like a horrible exaggeration), there was nothing which came close to being as good as SO. There was no replacement. People will use the best available platform, even if you insist in describing it as "horrible". It's was not horrible compared to the alternatives, web forums like Reddit and HN, which are poorly optimized for answering questions.
That "Dead Internet" phrase keeps becoming more likely, and this graph shows that. Human-to-human interactions, LLMs using those interactions, less human-to-human interactions because of that, LLMs using... ?
> I disagree with most comments that the brusque moderation is the cause of SO's problems, though it certainly didn't help.
By the time my generation was ready to start using SO, the gatekeeping was so severe that we never began asking questions. Look at the graph. The number of questions was in decline before 2020. It was already doomed because it lost the plot and killed any valuable culture. LLMs were a welcome replacement for something that was not fun to use. LLMs are an unwelcome replacement for many other things that are a joy to engage with.
> I disagree with most comments that the brusque moderation is the cause of SO's problems, though it certainly didn't help. SO has had poor moderation from the beginning.
Overwhelmingly, people consider the moderation poor because they expect to be able to come to the site and ask things that are well outside of the site's mission. (It's also common to attribute community actions to "moderators" who in reality have historically done hardly any of it; the site simply didn't scale like that. There have been tens of millions of questions, versus a couple dozen moderators.)
The kinds of questions that people are getting quick, accurate answers for from an LLM are, overwhelmingly, the sort of thing that SO never wanted. Generally because they are specific to the person asking: either that person's issue won't be relevant to other people, or the work hasn't been done to make it recognizable by others.
And then of course you have the duplicates. You would not believe the logic some people put forward to insist that their questions are not duplicate; that they wouldn't be able, in other words, to get a suitable answer (note: the purpose is to answer a question, not solve a problem) from the existing Q&A. It is as though people think they are being insulted when they are immediately given a link to where they can get the necessary answer, by volunteers.
I agree that Reddit played a big role in this. But not just by answering questions; by forming a place where people who objected to the SO content model could congregate.
Insulting other users is and always has been against Stack Overflow Code of Conduct. The large majority of insults, in my experience, come from new users who are upset at being politely asked to follow procedures or told that they aren't actually allowed to use the site the way they're trying to. There have been many duplicate threads on the meta site about why community members (with enough reputation) are permitted to cast close votes on questions without commenting on what is wrong. The consensus: close reasons are usually fairly obvious; there is an established process for people to come to the meta site to ask for more detailed reasoning; and comments aren't anonymous, so it makes oneself a target.
It seems you deny each problem that everyone sees in SO. The fact is SO repulsed people, so there is a gap between your interpretation and reality.
> It is as though people think they are being insulted when they are immediately given a link to where they can get the necessary answer, by volunteers.
This, for example. Question can be marked as duplicate without an answer. In this case yes, it feels insulting because the other is asked in such a weird way, that no-one will find the old when they search for the new (for example after a library change) and marking it as duplicate of an unanswered answer if a guarantee that the next SEO user won’t see it.
> It is as though people think they are being insulted when they are immediately given a link to where they can get the necessary answer, by volunteers.
Multiple times my questions closed as duplicates of question that was answering a different question.
Even when I explicitly linked that QA in my question and described how it differs from mine.
I don't think "good moderation or not" really touches what was happening with SO.
I joined SO early and it had a "gamified" interface that I actually found fun. Putting in effort and such I able to slowly gain karma.
The problem was as the site scaled, the competition to answer a given question became more and more intense and that made it miserable. I left at that point but I think a lot people stayed with dynamic that was extremely unhealthy. (and the quality of accepted questions declined also).
With all this, the moderation criteria didn't have to directly change, it just had to fail to deal with the effects that were happening.
This doesn't mean that it's over for SO. It just means we'll probably trend towards more quality over quantity. Measuring SO's success by measuring number of questions asked is like measuring code quality by lines of code. Eventually SO would trend down simply by advancements of search technology helping users find existing answers rather than asking new ones. It just so happened that AI advanced made it even better (in terms of not having to need to ask redundant questions).
> I wonder if, 10 years from now, LLMs will still be answering questions that were answered in the halcyon 2014-2020 days of SO better than anything that came after?
I've wondered this too and I wonder if the existing corpus plus new GitHub/doc site scrapes will be enough to keep things current.
> I disagree with most comments that the brusque moderation is the cause of SO's problems
Just to add another personal data point: i started posting in on StackOverflow well before llms were a thing and moderation instantly turned ne off and i immediately stopped posting.
Moderators used to edit my posts and reword what i wrote, which is unacceptable. My posts were absolutely peaceful and not inflammatory.
Moderation was an incredible problem for stack overflow.
> Moderators used to edit my posts and reword what i wrote, which is unacceptable. My posts were absolutely peaceful and not inflammatory.
99.9% probability the people who made those edits a) were not moderators; b) were acting completely in accordance with established policy (please read: "Why do clear, accurate, appropriately detailed posts still get edited?" https://meta.stackexchange.com/questions/403176)
Why do you think you should be the one who gets to decide whether that's "acceptable"? The site existed before you came to it, and it has goals, purposes and cultural norms established beforehand. It's your responsibility, before using any site on the Internet that accepts user-generated content, to try to understand the site's and community's expectations for that content.
On Stack Overflow, the expectations are:
1. You license the content to the site and to the community, and everyone is allowed to edit it. (This is also explicitly laid out in the TOS.)
2. You are contributing to a collaborative effort to build a useful resource for the programming community: a catalog of questions whose answers can be useful to many people, not just to yourself.
3. Content is intended to be matter-of-fact and right to the point, and explicitly not conversational. You are emphatically not participating in a discussion forum.
"I suspect that the gradual decline, beginning around 2016, is due to growth in a number of other sources of answers."
I think at least one other reason is that a lot of the questions were already posted. There are only so many questions of interest, until a popular new technology comes along. And if you look at mathoverflow (which wouldnt have the constant shocks from new technologies) the trend is pretty stable...until right around 2022. And even since then, the dropoff isn't nearly so dramatic.
https://data.stackexchange.com/mathoverflow/query/edit/19272...
I'll tell you what happens now: LLMs continue to regurgitate and iterate and hallucinate on the questions and answers they ingested from S.O. - 90% of which are incorrect. LLM output continues to poison itself as more and more websites spring up recycling outdated or incorrect answers, and no new answers are given since no one wants to waste the time to ask a human a question and wait for the response.
The overall intellectual capacity sinks to the point where everything collaboratively built falls apart.
The machines don't need AGI to take over, they just need to wait for us to disintegrate out of sheer laziness, sloth and self-righteous.... /okay.
there was always a needy component to Stack Overflow. "I have to pass an exam, what is the best way to write this algorithm?" and shit like that. A lazy component. But to be honest, it was the giving of information which forced you to think, and research, and answer correctly, which made systems like S.O. worthwhile, even if the questioners were lazy idiots sometimes. And now, the apocalypse. Babel. The total confusion of all language. No answer which can be trusted, no human in the loop, not even a smart AI, just a babbling set of LLMs repeating Stack Overflow answers from 10 years ago. That's the fucking future.
Things are gonna slide / in all directions / won't be nothin you can measure anymore. The blizzard of the world has crossed the threshold and it's overturned the order of the soul.[0]
Labs are spending billions on data set curation and RL from human experts to fill in the areas where they're currently weak. It's higher quality data than SO, the only issue is that it's not public.
There's another significant forum: GitHub, the rise of which coincided with the start of SO's decline. I bet most niche questions went over to GH repos' issue/discussion forums, and SO was left with more general questions that bored contributors.
The moderation was a lot of the problem, but not the whole problem. Honestly these days a larger part is that old, low-quality answers are usually stuck as the top answer on old questions, despite the fact that the situation has changed massively in the past decade and there are newer answers further down that give the new answer. Or better yet: when the top answer is a decade old and says "that doesn't even make sense, why would you want to do that, it's impossible, but you can look at literally the entire Handbook to see what you can do" (with a link to the frontpage of the FreeBSD Handbook) and you have to scroll down nearly to the bottom to find the one answer that actually answers the question (how to add an on-link route on FreeBSD) (and that it's not actually impossible like the arrogant jerk on top claimed)...
Too bad stack overflow didn't high-quality-LLM itself early. I assume it had the computer-related brainpower.
with respect to the "moderation is the cause" thing... Although I also don't buy moderation as the cause, I wonder if any sort of friction from the "primary source of data" can cause acceleration.
for example, when I'm doing an interenet search for the definition of a word like buggywhip, some search results from the "primary source" show:
> buggy whip, n. meanings, etymology and more | Oxford English Dictionary
> Factsheet What does the noun buggy whip mean? There is one meaning in OED's entry for the noun buggy whip. See 'Meaning & use' for definition, usage, and quotation evidence.
which are non-answer to keep their traffic.
but the AI answer is... the answer.
If SO early on had had some clear AI answer + references, I think that would have kept people on their site.
If my questions become more open ended and challenging as I become more senior, and I am not allowed to ask them on SO, why would I stay on the website?
So you get junior questions only and no seniors to answers.
The decline started in 2016, not 2026, they killed themselves
The newer questions that LLMs can't answer will be answered in forums - either SO, reddit, or elsewhere. There will be a much higher percentage of relevant content with far fewer new pages regurgitating questions about solved problems. So the LLMs will be able to keep up.
> - I know I'm beating a dead horse here, but what happens now? Despite stratification I mentioned above, SO was by far the leading source of high quality answers to technical questions. What do LLMs train off of now? I wonder if, 10 years from now, LLMs will still be answering questions that were answered in the halcyon 2014-2020 days of SO better than anything that came after? Or will we find new, better ways to find answers to technical questions?
To me this shows just how limited LLMs are. Hopefully more people realize that LLMs aren't as useful as they seem, and in 10 years they're relegated to sending spam and generating marketting websites.
I think the interesting thing here for those of us who use open source frameworks is that we can ask the LLM to look at the source to find the answer (eg. Pytorch or Phoenix in my case). For closed source libraries I do not know.
> What do LLMs train off of now? I wonder if, 10 years from now, LLMs will still be answering questions that were answered in the halcyon 2014-2020 days of SO better than anything that came after? Or will we find new, better ways to find answers to technical questions?
That's a great question. I have no idea how things will play out now - do models become generalized enough to handle "out of distrubition" problems or not ? If they don't then I suppose a few years from now we'll get an uptick in Stackoverflow questions; the website will still exist it's not going anywhere.
Specialized research AI agents are coming at which point we'll have numerous LLMs running and verifying experiments and creating a higher quality text corpus than the 2014-2020 halcyon, which is then used for other LLMs to be trained on.
It will be the reverse I suspect. Eventually we will see that LLM quality is lower when it is training data from 2014-2020 and will chalk it up to human limitations and the data not being written with a laser-focused goal of training better AI.
Instead of having chat-interfaces target single developers, moving towards multiplayer interfaces may bring back some of what has been lost--looping in experts or third-party knowledge when a problem is too though to tackle via agentic means.
Now all our interactions are neatly kept in personalised ledgers, bounded and isolated from one another. Whether by design or by technical infeasability, the issue remains that knowledge becomes increasingly bounded too instead of collaborative.
> SO was by far the leading source of high quality answers to technical questions
We will arrive on most answers by talking to an LLM. Many of us have an idea about we want. We relied on SO for some details/quirks/gotchas.
Example of a common SO question: how to do x in a library or language or platform? Maybe post on the Github for that lib. Or forums.. there are quirky systems like Salesforce or Workday which have robust forums. Where the forums are still much more effective than LLMs.
> The fundamental value proposition of SO is getting an answer to a question; if you can the same answer faster, you don't need SO.
Plus they might find the answer on SO without asking a new question - You probably would expect the # of new questions to peak or plateau even if the site wasn't dying, due to the accumulation of already-answered questions.
We'll get to the point where we can mass moderate core knowledge eventually. We may need to hand out extra weight for verified experts and some kind of most-votes-win type logic (perhaps even comments?), but live training data updates will be a massive evolution for language models.
Google also played a part. After a while, I noticed that for my programming related questions, almost no SO discussions showed up. When they did appear on the first page, they were usually abysmal and unusable for me.
When it started all kinds of very clever people were present and helped even with very deep and complex questions and problems. A few years later these people disappeared. The moderation was ok in the beginning, then they started wooing away a lot of talented people. And then the mods started acting like nazis, killing discussions, proper questions on a whim.
And then bots (?) or karma obsessed/farming people started to upvote batshit crazy, ridiculous answers, while the proper solution had like 5 upvotes and no green marker next to it.
It was already a cesspool before AI took over and they sold all their data. Initial purpose achieved.
As an early user of SO [1], I feel reasonably qualified to discuss this issue. Note that I barely posted after 2011 or so so I can't really speak to the current state.
But what I can say is that even back in 2010 it was obvious to me that moderation was a problem, specifically a cultural problem. I'm really talking about the rise of the administrative/bureaucratic class that, if left unchecked, can become absolute poison.
I'm constantly reminded of the Leonard Nimoy voiced line from Civ4: "the bureaucracy is expanding to meet the needs of the expanding bureaucracy". That sums it up exactly. There is a certain type of person who doesn't become a creator of content but rather a moderator of content. These are people who end up as Reddit mods, for example.
Rules and standards are good up to a point but some people forget that those rules and standards serve a purpose and should never become a goal unto themselves. So if the moderators run wild, they'll start creating work for themselves and having debates about what's a repeated question, how questions and answers should be structured, etc.
This manifested as the war of "closed, non-constructive" on SO. Some really good questions were killed this way because the moderators decided on their own that a question had to have a provable answer to avoid flame wars. And this goes back to the rules and standards being a tool not a goal. My stance was (and is) that shouldn't we solve flame wars when they happen rather than going around and "solving" imaginary problems?
I lost that battle. You can argue taht questions like "should I use Javascript or Typescript?" don't belong on SO (as the moderators did). My position was that even though there's no definite answer, somebody can give you a list of strengths and weaknesses and things to consider.
Even something that does have a definite answer like "how do I efficiently code a factorial function?" has multiple but different defensible answers. Even in one language you can have multiple implementations that might, say, be compile-time or runtime.
Another commenter here talked about finding the nearest point on an ellipse and came up with a method they're proud of where there are other methods that would also do the job.
Anyway, I'd occasionally login and see a constant churn on my answers from moderators doing pointless busywork as this month they'd decided something needed to be capitalized or not capitalized.
A perfect example of this kind of thing is Bryan Henderson's war on "comprised of" on Wikipedia [2].
Anyway, I think the core issue of SO was that there was a lot of low-hanging fruit and I got a lot of accepted answers on questions that could never be asked today. You'll also read many anecdotes about people having a negative experience asking questions on SO in later years where their question was immediately closed as, say, a duplicate when the question wasn't a duplicate. The moderator just didn't understand the difference. That sort of thing.
But any mature site ultimately ends with an impossible barrier to entry as newcomers don't know all the cultural rules that have been put in place and they tend to have a negative experience as they get yelled at for not knowing that Rule 11.6.2.7 forbids the kind of question they asked.
> This manifested as the war of "closed, non-constructive" on SO. Some really good questions were killed this way because the moderators decided on their own that a question had to have a provable answer to avoid flame wars.
It's literally a Q&A site. Questions need actual answers, not just opinions or "this worked for me".
> This manifested as the war of "closed, non-constructive" on SO. Some really good questions were killed this way because the moderators decided on their own that a question had to have a provable answer to avoid flame wars.
Please point at some of these "really good" questions, if you saved any links. (I have privileges to see deleted questions; deletion is normally soft unless there's a legal requirement or something.) I'll be happy to explain why they are not actually what the site wanted and not compatible with the site's goals.
> I lost that battle. You can argue taht questions like "should I use Javascript or Typescript?" don't belong on SO (as the moderators did). My position was that even though there's no definite answer, somebody can give you a list of strengths and weaknesses and things to consider.
Dunno why you are being downvoted - there is a certain type of person who contributes virtually nothing on Wikipedia except peripheral things like categories. BrownHairedGirl was the most toxic person in Wikipedia but she was lauded by her minions - and yet she did virtually no content creation whatsoever. Yet made millions of edits!
Perhaps they’ll rely on what was used by people who answered SO questions. So: official docs and maybe source code. Maybe even from experience too, i.e. from human feedback and human written code during agentic coding sessions.
> The fact that the LLM doesn't insult you is just the cherry on top.
Arguably it does insult even more, just by existing alone.
I spent the last 14 days chasing an issue with a Spark transform. Gemini and Claude were exceptionally good at giving me answers that looked perfectly reasonable: none of them worked, they were almost always completely off-road.
Eventually I tried with something else, and found a question on stackoverflow, luckily with an answer. That was the game changer and eventually I was able to find the right doc in the Spark (actually Iceberg) website that gave me the final fix.
This is to say that LLMs might be more friendly. But losing SO means that we're getting an idiot friendly guy with a lot of credible but wrong answers in place of a grumpy and possibly toxic guy which, however, actually answered our questions.
Not sure why someone is thinking this is a good thing.
What I always appreciate about SO is the dialogue between commenters. LLMs give one answer, or bullet points around a theme, or just dump a load of code in your IDE. SO gives a debate, in which the finer points of an issue are thrashed out, with the best answers (by and large) floating to the top.
SO, at its best, is numerous highly-experienced and intelligent humans trying to demonstrate how clever they are. A bit like HN, you learn from watching the back and forth. I don't think this is something that LLMs can ever replicate. They don't have the egos and they certainly don't have the experience.
Whatever people's gripes about the site, I learned a hell of a lot from it. I still find solutions there, and think a world without it would be worse.
The fundamental difference between asking on SO and asking an LLM is that SO is a public forum, and an LLM will be communicated with in private. This has a lot of implications, most of which surround the ability for people to review and correct bad information.
> What I always appreciate about SO is the dialogue between commenters.
Stack Overflow is explicitly not for "dialogue", recent experiments (which are generally not well received by the regulars on the meta site) notwithstanding. The purpose of the comments on questions is to help refine the question and ensure it meets standards, and in some cases serve other meta purposes like pointing at different-but-related questions to help future readers find what they're looking for. Comments are generally subject to deletion at any time and were originally designed to be visually minimal. They are not part of the core experience.
Of course, the new ownership is undoing all of that, because of engagement metrics and such.
I don't know if this is still the case but back in the day people would often redirect comments to some stackoverflow chat feature, the links to which would always return 404 not found errors.
You can ask an LLM to provide multiple approaches to solutions and explore the pros and cons of each, then you can drill down and elaborate on particular ones. It works very well.
It's flat wrong to suggest SO had the right answer all the time, and in fact in my experience for trickier work it was often wrong or missing entirely.
The example wasn't even finding a right answer so I don't see where you got that..
Searching questions/answers on SO can surface correct paths on situations where the LLMs will keep giving you variants of a few wrong solutions, kind of like the toxic duplicate closers.. Ironically, if SO pruned the history to remove all failures to match its community standards then it would have the same problem.
It entirely depends on the language you were using. The quality of both questions and answers between e.g. Go and JavaScript is incredible. Even as a relative beginner in JS I could not believe the amount of garbage that I came across, something that rarely happened for Go.
Because what you’re describing is the exception. Almost always with LLM’s I get a better solution, or helpful pointer in the direction of a solution, and I get it much faster. I honestly don’t understand anyone could prefer Google/SO, and in fact that the numbers show that they don’t. You’re in an extreme minority.
> But losing SO means that we're getting an idiot friendly guy with a lot of credible but wrong answers in place of a grumpy and possibly toxic guy which, however, actually answered our questions.
Which by the way is incredibly ironic to read on the internet after like fifteen years of annoying people left and right about toxic this and toxic that.
Extreme example: Linus Torvalds used to be notoriously toxic.
Would you still defend your position if the “grumpy” guy answered in Linus’ style?
> Would you still defend your position if the “grumpy” guy answered in Linus’ style?
If they answered correctly, yes.
My point is that providing _actual knowledge_ is by itself so much more valuable compared to _simulated knowledge_, in particular when that simulated knowledge is hyper realistic and wrong.
Sadly, an accountable individual representing an organization is different from a community of semi-anonymous users with a bunch of bureaucracy that can't or doesn't care about every semis anonymous user
That grumpy guy is using an LLM and debugging with it. Solves the problem. AI provider fine tunes their model with this. You now have his input baked into it's response.
How you think these things work? It's either a human direct input it's remembering or a RL enviroment made by a human to solve the problem you are working on.
Nothing in it is "made up" it's just a resolution problem which will only get better over time.
I'm hoping increasing we'll see agents helping with this sort of issue. I would like an agent that would do things like pull the spark repo into the working area and consult the source code/cross reference against what you're trying to do.
Once technique I've used successfully is to do this 'manually' to ensure codex/Claude code can grep around the libraries I'm using
Not a big surprise once LLMs came along: stack overflow developed some pretty unpleasant traits over time. Everything from legitimate questions being closed for no good reason (or being labeled a duplicate even though they often weren’t), out of date answers that never get updated as tech changes, to a generally toxic and condescending culture amongst the top answerers. For all their flaws, LLMs are so much better.
People in this thread are missing another key component in the decline of StackOverflow - the more experienced you become, the less useful it is.
The harder the problem, the less engagement it gets. People who spend hours working on your issue are rewarded with a single upvote. Meanwhile, "how do I concat a string" gets dozens or hundreds of upvotes.
The incentive/reward structure punished experienced folks with challenging/novel questions.
Pair that with the toxic moderation and trigger-happy close-votes, you get a zombie community with little new useful content.
Same here. I just didn't want to expend energy racing trigger happy mods. It was so odd, to this day remember vividly how they cleanup their arguments once proven wrong on the closing vote. Literally minutes before it would the close threshold.
And you can't delete your post when you realize how awful it was years later! That anti-information sticks around for ages. Even worse when there are bad answers attached to it, too.
The dumbest part of SO is how the accepted answer would often be bad, and sometimes someone had posted a better answer after the fact, and if the all-powerful moderators had the power to update it, they sure never did. Likewise, there were often better insights in comments. Apparently if you have the right mod powers, you can just edit an answer (such as the accepted one) to make it correct, but that always struck me as a bizarre feature, to put words in other people’s mouths.
I think overall SO took the gamification, and the “internet points” idea, way too far. As a professional, I don’t care about Reddit Karma or the SO score or my HN karma. I just wanted answers that are correct, and a place to discuss anything that’s actually interesting.
I did value SO once as part of the tedious process of attempting to get some technical problem solved, as it was the best option we had, but I definitely haven’t been there since 2023. RIP.
Seemed like for every other question, I received unsolicited advice telling me how I shouldn't be doing it this way, only for me to have to explain why I wanted to do it this way (with silence from them).
This is called the XY problem https://meta.stackexchange.com/a/66378 . You ask for X, I tell you that what you really want is Y, I bully you, and I become more convinced that you and people that ask for X want Y.
Stack Overflow would still have a vibrant community if it weren't for the toxic community.
Imagine a non-toxic Stack Overflow replacement that operated as an LLM + Wiki (CC-licensed) with a community to curate it. That seems like the sublime optimal solution that combines both AI and expertise. Use LLMs to get public-facing answers, and the community can fix things up.
No over-moderation for "duplicates" or other SO heavy-handed moderation memes.
Someone could ask a question, an LLM could take a first stab at an answer. The author could correct it or ask further questions, and then the community could fill in when it goes off the rails or can't answer.
You would be able to see which questions were too long-tail or difficult for the AI to answer, and humans could jump in to patch things up. This could be gamified with points.
This would serve as fantastic LLM training material for local LLMs. The authors of the site could put in a clause saying that "training is allowed as long as you publish your weights + model".
Someone please build this.
Edit: Removed "LLMs did not kill Stack Overflow." first sentence as suggested. Perhaps that wasn't entirely accurate, and the rest of the argument stands better on its own legs.
It doesn't have anything to do with LLMs. It has to do with shifting one's focus from doing good things to making money. Joel did that, and SO failed because of it.
Joel promised the answering community he wouldn't sell SO out from under them, but then he did.
And so the toxicity at the top trickled down into the community.
Those with integrity left the community and only toxic, selfcentered people remained to destroy what was left in effort to salvage what little there was left for themselves.
Mods didn't dupe questions to help the community. They did it to keep their own answers at the top on the rankings.
How did Joel sell out? Curious as I’m not aware of any monetary changes. I watched Joel several times support completely brain dead policies in the meta discussions which really set the rules and tone. So my respect there is low.
I also wonder if GitHub Discussions was also a (minor) contributing factor to the decline. I recall myself using GitHub Discussions more and more when it came to repo specific issues.
My memory is there were a spate of SO scraping sites that google would surface above SO and google just would not zap.
It would have been super trivial to fix but google didn’t.
My pet theory was that google were getting doubleclick revenue from the scrapers so had incentives to let them scrape and to promote them in search results.
> legitimate questions being closed for no good reason
They are closed for good reasons. People just have their own ideas about what the reasons should be. Those reasons make sense according to others' ideas about what they'd like Stack Overflow to be, but they are completely wrong for the site's actual goals and purposes. The close reasons are well documented (https://meta.stackoverflow.com/questions/417476) and well considered, having been exhaustively discussed over many years.
> or being labeled a duplicate even though they often weren’t
I have seen so many people complain about this. It is vanishingly rare that I actually agree with them. In the large majority of cases it is comically obvious to me that the closure was correct. For example, there have been many complaints in the Python tag that were on the level of "why did you close my question as a duplicate of how to do X with a list? I clearly asked how to do it with a tuple!" (for values of X where you do it the same way.)
> a generally toxic and condescending culture amongst the top answerers.
On the contrary, the top answerers are the ones who will be happy to copy and paste answers to your question and ignore site policy, to the constant vexation of curators like myself trying to keep the site clean and useful (as a searchable resource) for everyone.
> For all their flaws, LLMs are so much better.
I actually completely agree that people who prefer to ask LLMs should ask LLMs. The experience of directly asking (an LLM) and getting personalized help is explicitly the exact thing that Stack Overflow was created to get away from (i.e., the traditional discussion forum experience, where experts eventually get tired of seeing the same common issues all the time and all the same failures to describe a problem clearly, and where third parties struggle to find a useful answer in the middle of along discussion).
You seem to have filled this thread with a huge number of posts that try to justify SO's actions. Over and over, these justifications are along the lines of "this is our mission", "read our policy", "understand us".
Often, doing what your users want leads to success. Stamping authority over your users, and giving out a constant air of "we know better than all of you", drives them away. And when it's continually emphasized publicly (rather than just inside a marketing department) that the "mission" and the "policy" are infinitely more important than what your users are asking for, that's a pretty quick route to failure.
When you're completely embedded in a culture, you don't have the ability to see it through the eyes of the majority on the outside. I would suggest that some of your replies here - trying to deny the toxicity and condescension - are clearly showing this.
> "why did you close my question as a duplicate of how to do X with a list? I clearly asked how to do it with a tuple!" (for values of X where you do it the same way.)
This is a great example of a question that should not be closed as a duplicate. Lists are not tuples in Python, regardless of how similar potential answers may be.
I think it has more to do with the fact that when you offer zero salary for moderators, you have to take what you can get, and it ain't good. I don't really see a connection to the voting mechanic.
But LLMs get their answers from StackOverflow and similar places being used as the source material. As those start getting outdated because of lack of activity, LLMs won't have the source material to answer questions properly.
I regularly use Claude and friends where I ask it to use the web to look at specific GitHub repos or documentation to ask about current versions of things. The “LLMs just get their info from stack overflow” trope from the GPT-3 days is long dead - they’re pretty good at getting info that is very up to date by using tools to access the web. In some cases I just upload bits and pieces from a library along with my question if it’s particularly obscure or something home grown, and they do quite well with that too. Yes, they do get it wrong sometimes - just like stack overflow did too.
StackOverflow answers are outdated. Every time I end up on that site these days, I find myself reading answers from 12 years ago that are no longer relevant.
That depends on what you mean by "came along". If you mean "once that everyone got around to the idea that LLMs were going to be good at this thing" then sure, but it was not long ago that the majority of people around here were very skeptical of the idea that LLMs would ever be any good at coding.
What you're arguing about is the field completely changing over 3 years; it's nothing, as a time for everyone to change their minds.
LLMs were not productified in a meaningful way before ChatGPT in 2022 (companies had sufficiently strong LLMs, but RLHF didn't exist to make them "PR-safe"). Then we basically just had to wait for LLM companies to copy Perplexity and add search engines everywhere (RAG already existed, but I guess it was not realistic to RAG the whole internet), and they became useful enough to replace StackOverflow.
I dont think this is true. People were skeptical of agi / better than human coding which is not the case. As a matter of fact i think searching docs was one of the first manor uses of llms before code.
That's because there has been rapid improvement by LLMs.
Their tendency to bullshit is still an issue, but if one maintains a healthy skepticism and uses a bit of logic it can be managed. The problematic uses are where they are used without any real supervision.
Enabling human learning is a natural strength for LLMs and works fine since learning tends to be multifaceted and the information received tends to be put to a test as a part of the process.
We can't. I don't think the LLMs themselves can recognize when an answer is stale. They could if contradicting data was available, but their very existence suppresses the contradictory data.
They will, but model updates and competition help solve the problem. If people find that Claude consistently gives better/more relevant answers over GPT, for example, people will choose the better model.
The worst thing with Q/A sites isn't they don't work. It's that they there are no alternatives to stackoverflow. Some of the most upvoted answers on stackoverflow prove that it can work well in many cases, but too bad most other times it doesn't.
They still use the official documentation/examples, public Github Repos, and your own code which are all more likely to be evergreen. SO was definitely a massive training advantage before LLMs matured though.
Yep, LLMs are perfect for the "quick buy annoying to answer 500 times" questions about writing a short script, or configuring something, or using the right combination of command line parameters.
Quicker than searching the entirety of Google results and none of the attitude.
Indeed. StackOverflow was by far the most unpleasant website that I have regularly interacted with. Sometimes, just seeing how users were treated there (even in Q&A threads that I wasn’t involved in at all) disturbed me so much it was actually interfering with my work. I’m so, so glad that I can now just ask an AI to get the same (or better) answers, without having to wade through the barely restrained hate on that site.
not only stackoverflow, but also reddit.com/r/aws reddit.com/r/docker reddit.com/r/postgresql all 3 of them have extremely toxic communities. ask a question and get downvoted instantly! Noo!! your job is to actually upvote the question to maximize exposure for the algorithm unless it is a really really stupid question that a google search could fix
They will no doubt blame this on AI, somehow (ChatGPT release: late 2022, decline start: mid 2020), instead of the toxicity of the community and the site's goals of being a knowledgebase instead of a QA site despite the design.
Right. I often end up on Stack Exchange when researching various engineering-related topics, and I'm always blown away by how incredibly toxic the threads are. We get small glimpses of that on HN, but it was absolutely out of control on Stack Exchange.
At the same time, I think there was another factor: at some point, the corpus of answered questions has grown to a point where you no longer needed to ask, because by default, Google would get you to the answer page. LLMs were just a cherry on top.
I agree there was some natural slow down as the corpus grew - the obvious questions were answered. But if the community was healthy, that should not have caused growth to stop. New technologies get created all the time, each starting with zero SO questions. (Or Google releases v2.0 which invalidates all answers written about v1.)
SO just stopped being fun for me. I wish more systems would use their point systems though.
People overestimate the impact of toxicity on number of monthly questions. The initial growth was due to missing answers. After some time there is a saturation point where all basic questions are already answered and can be found via Google. If you ask them again they are marked as dups.
The downward trend seems to start ~2017, and was interrupted by a spike during the early months of COVID-19. I'd be interested to know what drove that jump, perhaps people were less hesitant to post when they were working from home?
More people spent lot more time learning new tech skills (at every experience level).
The excess time available (less commute or career pause etc) and more interest (much more new opportunities) were probably leading reasons why they spent more time I would imagine.
A huge amount of people were just starting to learn programming, because they were stuck at home and had the time to pick something up.
If you look at the trends tag by tag, you can see that the languages, libraries, technologies etc. that appeal to beginners and recreational coders grew disproportionately.
Other tech support forums are terrible in other ways. AI is a godsend.
Typical response:
I am RJ, an Independent Advisor and Microsoft Gold Certified Support Specialist Enthusiast.
I know how your system is not functioning as desired! Rest assured, I am here to help you resolve this today.
Please follow these steps in order. Do not skip any steps.
Step 1: Reboot your computer
Step 2: Reinstall windows
Step 3: Contact Microsoft support
Did this resolve your issue? [ Yes ] [ No ]
If this helped, please mark this as the Answer and give me a 5-star rating so I can continue providing high-quality, scripted responses to other users!
Standard Disclaimer: I do not work for Microsoft. I am an independent volunteer who enjoys copying and pasting from a manual written in 2014.
> the site's goals of being a knowledgebase instead of a QA site despite the design.
A Q&A site is a knowledge base. That's just how the information is presented.
If you want a forum — a place where you ask the question to get answered one-on-one — you have countless options for that.
Stack Overflow pages have a different design from that explicitly to encourage building a knowledge base. That's why there's a question at the top and answers underneath it, and why there are not follow-up questions, "me too" posts, discussion of annoyances related to the question, tangential rants, generic socialization etc.
Jeff Atwood was quite clear about this from the beginning.
I wonder what is the role of moderating duplicate questions. More time passes - more existing data there is and less need for new questions. If you moderate duplicate questions, will they disappear from these charts? Is this decline actually logical?
Many people are pointing out the toxicity, but the biggest thing that drove me away, especially for specific quantitative questions, was that SO was flat out wrong (and confidently so) on many issues.
It was bad enough that I got back in the habit of buying and building a library of serious reference books because they were the only reliable way to answer detailed technical questions.
There is an obvious acceleration of the downwards trend at the time ChatGPT got popular. AI is clearly a part of this, but not the only thing that affects SO activity.
Ironically they could probably do some really useful deduplication/normalization/search across questions and answers using AI/embeddings today, if only they’d actually allowed people to ask the same questions infinite different ways, and treated the result of that as a giant knowledge graph.
I was into StackOverflow in the early 2010s but ultimately stopped being an active contributor because of the stupid moderation.
Toxic community is mostly a meme myth. I have like 30k points and whatever admins were doing was well deserved as 90% of the questions were utterly impossible to help with. Most people wanted free help and couldn't even bother to put in 5 minutes of work.
Actual analysts here that have looked at this graph like... a lot, so let me contextualize certain themes that tend to crop up from these:
- The reduction of questions over time is asymptomatic of SO. When you have a library of every question asked, at some point, you asked most of the easy questions. Have a novel question becomes hard.
- This graph is using the Posts table, not PostsWithDeleted. So, it only tells you of the questions that survived at this point in time, this [0] is the actual graph which while describes a curve that shows the same behavior, it's more "accurate" of the actual post creation.
- This is actually a Good Thing™. For years most of the questions went unanswered, non-voted, non-commented, just because there was too many questions happening all the time. So the general trend is not something that the SO community needs to do anything about. Almost 20% of every question asked is marked as duplicate. If people searched... better™ they wouldn't ask as many questions, and so everyone else had more bandwidth to deal with the rest.
- There has been a shift in help desk style of request, where people starting to prefer discord and such to get answers. This is actually a bad thing because that means that the knowledge isn't public nor indexed by the world. So, information becomes harder to find, and you need to break it free from silos.
- The site, or more accurately, the library will never die. All the information is published in complete archives that anyone can replicate and restart if the company goes under or goes evil. So, yeah, such concerns, while appreciated, are easily addressed. At worst, you would be losing a month or two of data.
> When you have a library of every question asked, at some point, you asked most of the easy questions. Have a novel question becomes hard
This would be true if programming were a static field, but given that new programming languages/frameworks/technologies/techniques/etc. are constantly coming out and evolving, that argument doesn't make sense.
Programming is not a static field in the answers side, but it's in the question side. "How to print characters on a terminal with python?" is the same problem today as it was 25 years ago. The answer changed but the problem remained. That's what people saying that programming isn't static is missing: the problem space grows significantly slower than the solution space.
OP here: I had the same thought, but noticed a very similar trend in both [0]; I think this graph is more interesting because you'd expect the number of new users to be growing [1], but this seems to have very little effect on deleted questions or even answers
The second graph here ([1]) is especially interesting because the total montly number of new users seems completely unrelated to number of posts, until you filter for a rep > 1 which has a close to identical trend
I guess I'm the only one that was a fan of SO's moderation. I never got too deep into it (answered some TypeScript questions). But the intention to reduce duped questions made a lot of sense to me. I like the idea of a "living document" where energy is focused on updating and improving answers to old versions of the same question. As a user looking for answers it means I can worry less about finding some other variation of the same question that has a more useful answer
I understand some eggs got cracked along the way to making this omelette but overall I'd say about 90% of the time I clicked on a SO link I was rewarded with the answer I was looking for.
Those saying that StackOverflow became toxic are absolutely correct. But we should not let that be it's legacy. It is IMO still today one of the greatest achievements in terms of open data on the internet. And it's impact on making programming accessible to a large audience cannot be understated.
Many users left because they had had overly strict moderation for posting your questions. I have 6k reputation, multiple gold badges and I will remember StackOverflow as a hostile place to ask a questions, honestly. There were multiple occasions when they actually prevented me from asking, and it was hard to understand what exactly went wrong. To my understanding, I asked totally legit questions, but their asking policy is so strict, it's super hard to follow.
So "I'm not happy he's dead, but I'm happy he's gone" [x]
I have around 2k points, not something to brag about, but probably more than most stackoverflow users. And I know what I am talking about given over a decade of experience in various tech stacks.
But it requires 3,000 points to be able to cast a vote to reopen a question, many of which incorrectly marked as duplicate.
I was an early adopter. Have over 30k reputation because stack overflow and my internship started at the same time. I left because of the toxic culture, and that it's less useful the more advanced you get
Please feel free to cite examples. I'll be happy to explain why I think they're duplicates, assuming I do (in my experience, well over 90% of the time I see this complaint, it's quite clear to me that the question is in fact a duplicate).
But more importantly, use the meta site if you think something has been done poorly. It's there for a reason.
It's also was a bit frustrating for me to answer. There was time when I wanted to contribute, but questions that I could answer were very primitive and there were so many people eager to post their answer that it demotivated me and I quickly stopped doing that. Honestly there are too many users and most of them know enough to answer these questions. So participating as "answerer" wasn't fun for me.
Once StackOverflow profiles, brief as they were, became a metric they ceased to be worth a helluva lot. Back in the early 2010s I used to include a link to my profile. I had a low 5-figure score and I had more than one interviewer impressed with my questions and answers on the site. Then came point farmers.
I remember one infamous user who would farm points by running your questions against some grammar / formatting script. He would make sure to clean up an errant comma or a lingering space character at the end of your post to get credit for editing your question, thereby “contributing.”
To their early credit, I once ran for and nearly won a moderator slot. They sent a nice swag package to thank me for my contributions to the community.
I spent a lot of time answering rather primitive questions, but since it was on a narrow topic (Logstash, part of the ELK stack), there wasn't many other people eager to post answers. Though it often ended up with the same type of issues, not necessarily duplicates, but similar enough that I got bored with it.
25k here, stopped posting cause you'd spend 10m on a reply to a question just to have the question closed on you by some mod trying to make everything neat.
Maybe it was a culture clash but I came from newsgroups where the issue was is that someone needed help. However SO had the idea that the person who needed help wasn't as important as the normalisation of the dataset.
I sometimes wonder how much time I could have saved for those whose questions got closed before I could answer them. But yeah, that and the swearing culture clash were issues I struggled with, and ultimately meant I stopped contributing.
> Maybe it was a culture clash but I came from newsgroups where the issue was is that someone needed help. However SO had the idea that the person who needed help wasn't as important as the normalisation of the dataset.
Yes, because doing things that way was explicitly part of the goal, from the beginning. There are countless other places where you can directly respond to people who need help (and if you like doing that, you should stick to those places). Doing things that way has negative consequences in terms of making something that's useful for on-lookers, and causing a lot of experts to burn out or get frustrated. This is stuff that Jeff Atwood was pointing out when explaining the reason for creating SO in the first place.
> I sometimes wonder how much time I could have saved for those whose questions got closed before I could answer them.
I consider it the most beautiful piece of code I've ever written and perhaps my one minor contribution to human knowledge. It uses a method I invented, is just a few lines, and converges in very few iterations.
People used to reach out to me all the time with uses they had found for it, it was cited in a PhD and apparently lives in some collision plugin for unity. Haven't heard from anyone in a long time.
It's also my test question for LLMs, and I've yet to see my solution regurgitated. Instead they generate some variant of Newtons method, ChatGPT 5.2 gave me an LM implementation and acknowledged that Newtons method is unstable (it is, which is why I went down the rabbit hole in the first place.)
Today I don't know where I would publish such a gem. It's not something I'd bother writing up in a paper, and SO was the obvious place were people who wanted an answer to this question would look. Now there is no central repository, instead everyone individually summons the ghosts of those passed in loneliness.
The various admonitions to publish to a personal blog, while encouraging, don't really get at the 0xfaded's request which I'd summarize as follows:
With no one asking questions these technical questions publicly, where, how and on what public platform will technical people find the problems that need solving so they can exercise their creativity for the benefit of all?
Clearly we need something in between the fauxpen-access of journals and the wilde west of the blogosphere, probably. Why wouldn't the faded ox publish in a paper? Idk, but I guess we need things similar to those circulars that British royal society members used to send to each other...except not reserved for a club. The web should be a natural at this. But it's either centralized -> monetized -> corrupted, or decentralized -> unindexed/niche -> forgotten fringe. What can come between?
You can (and always were encouraged to) ask your own questions, too.
And there are more sites like this (see e.g. https://codidact.com — fd: moderator of the Software section). Just because something loses popularity isn't a reason to stop doing it.
Seriously where will we get this info anymore? I’ve depended on it for decades. No matter how obscure, I could always find a community that was talking about something I needed solved. I feel like that’s getting harder and harder every year. The balkanization of the Internet + garbage AI slop blogs overwhelming the clearly declining Google is a huge problem.
> where, how and on what public platform will technical people find the problems that need solving so they can exercise their creativity for the benefit of all?
The same place people have always discovered problems to work on, for the entire history of human civilization. Industry, trades, academia, public service, newspapers, community organizations. The world is filled with unsolved problems, and places to go to work on them.
This is a perfect example of an element of Q&A forums that is being lost. Another thing that I don't think we'll see as much of anymore is interaction from developers that have extensive internal knowledge on products.
An example I can think of was when Eric Lippert, a developer on the C# compiler at the time, responded to a question about a "gotcha" in the language: https://stackoverflow.com/a/8899347/10470363
Developer interaction like that is going to be completely lost.
I agree that there will be some degradation here, but I also think that the developers inclined to do this kind of outreach will still find ways to do it.
I believe the community has seen the benefit of forums like SO and we won’t let the idea go stale. I also believe the current state of SO is not sustainable with the old guard flagging any question and response you post there. The idea can/should/might be re-invented in an LLM context and we’re one good interface away from getting there. That’s at least my hope.
I used to look at all TensorFlow questions when I was on the TensorFlow team (https://stackoverflow.com/tags/tensorflow/info). Unclear where people go to interact with their users now....Reddit? But the tone on Reddit is kind of negative/complainy
I had a similar beautiful experience where an experienced programmer answered one of my elementary JavaScript typing questions when I was just starting to learn programming.
He didn't need to, but he gave the most comprehensive answer possible attacking the question from various angles.
He taught me the value of deeply understanding theoretical and historical aspects of computing to understand why some parts of programming exist the way they are. I'm still thankful.
If this was repeated today, an LLM would have given a surface level answer, or worse yet would've done the thinking for me obliviating the question in the first place.
Had a similar experience. Asked a question about a new language feature in java 8 (parallell streams), and one of the language designers (Goetz) answered my question about the intention of how to use it.
An LLM couldn't have done the same. Someone would have to ask the question and someone answer it for indexing by the LLM. If we all just ask questions in closed chats, lots of new questions will go unanswered as those with the knowledge have simply not been asked to write the answers down anywhere.
You can prompt the LLM to not just give you the answer. Possibly even ask it to consider the problem from different angles but that may not be helpful when you don't know what you don't know.
You can write a paper, submit the arxiv, and you can also make a blog post.
At any rate, I agree - SO was (is?) a wonderful place for this kind of thing.
I once had a professor mention that they knew me from SO because I posted a few underhanded tricks to prevent an EKF from "going singular" in production. That kind of community is going to be hard to replace, but SO isnt going anywhere, you can still ask a question and answer your own question for permanent, searchable archive.
Has anyone tried building a modern Stack Overflow that's actually designed for AI-first developers?
The core idea: question gets asked → immediately shows answers from 3 different AI models. Users get instant value. Then humans show up to verify, break it down, or add production context.
But flip the reputation system: instead of reputation for answers, you get it for catching what's wrong or verifying what works. "This breaks with X" or "verified in production" becomes the valuable contribution.
Keep federation in mind from day one (did:web, did:plc) so it's not another closed platform.
Stack Overflow's magic was making experts feel needed. They still do—just differently now.
Oh, so it wasn't bad enough to spot bad human answers as an expert on Stack Overflow... now humans should spend their time spotting bad AI answers? How about a model where you ask a human and no AI input is allowed, to make sure that everyone has everyone else's full attention?
AI is generally setup to return the "best" answer as defined as the most common answer, not the rightest, or most efficient or effective answer, unless the underlying data leans that way.
It's why AI based web search isn't behaving like google based search. People clicking on the best results really was a signal for google on what solution was being sought. Generally, I don't know that LLMs are covering this type of feedback loop.
That seems like a horrible core idea. How is that different from data labeling or model evaluation?
Human beings want to help out other human beings, spread knowledge and might want to get recognition for it. Manually correcting (3 different) automation efforts seems like incredible monotone, unrewarding labour for a race to the bottom. Nobody should spend their time correcting AI models without compensation.
I think this could be really cool, but the tricky thing would be knowing when to use it instead of just asking the question directly to whichever AI. It’s hard to know that you’ll benefit from the extra context and some human input unless you already have a pretty good idea about the topic.
thanks for sharing that, it was simple, neat, elegant.
this sent me down a rabbit hole -- I asked a few models to solve that same problem, then followed up with a request to optimize it so it runs more efficiently.
chatgpt & gemini's solutions were buggy, but claude solved it, and actually found a solution that is even more efficient. It only needs to compute sqrt once per iteration. It's more complex however.
yours claude
------------------------------
Time (ns/call) 40.5 38.3
sqrt per iter 3 1
Accuracy 4.8e-7 4.8e-7
Claude's trick: instead of calling sin/cos each iteration, it rotates the existing (cos,sin) pair by the small Newton step and renormalizes:
// Rotate (c,s) by angle dt, then renormalize to unit circle
float nc = c + dt*s, ns = s - dt*c;
float len = sqrt(nc*nc + ns*ns);
c = nc/len; s = ns/len;
Thanks for pushing this, I've never gone beyond "zero" shotting the prompt (is it still called zero shot with search?)
As a curiosity, it looks like r and q are only ever used as r/q, and therefore a sqrt could be saved by computing rq = sqrt((rxrx + ryry) / (qxqx + qyqy)). The if q < 1e-10 is also perhaps not necessary, since this would imply that the ellipse is degenerate. My method won't work in that case anyway.
For the other sqrt, maybe try std::hypot
Finally, for your test set, could you had some highly eccentric cases such as a=1 and b=100
Thanks for the investigation:)
Edit: BTW, the sin/cos renormalize trick is the same as what tx,ty are doing. It was pointed out to me by another SO member. My original implementation used trig functions
I can relate. I used to have a decent SO profile (10k+ reputation, I know this isnt crazy but it was mostly on non low hanging fruit answers...it was a grind getting there). I used to be proud of my profile and even put it in my resume like people put their Github. Now - who cares? It would make look like a dinosaur sharing that profile, and I never go to SO anymore.
I too, around 2012 was too much active on so, in fact, it had that counter thing continuously xyz days most of my one liners, or snippets for php are still the highest voted answers. Even now when sometimes I google something, and an answer comes up, I realize its me who asked the same question and answered it too.
I don't disagree completely by any means, it's an interesting point, but in your SO answer you already point to your blog post explaining it in more detail, so isn't that the answer, you'd just blog about it and not bother with SO?
Then AI finding it (as opposed to already trained well enough on it, I suppose) will still point to it as did your SO answer.
Looks like solid code. My only gripe is the shadowing of x. I would prefer to see `for _ in range`. You do redefine it immediately so it's not the most confusing, but it could trip people up especially as it's x and not i or something.
I once wrote this humdinger, that's still on my mostly dead personal website from 2010... one of my proudest bits of code besides my poker hand evaluator ;)
The question was, how do you generate a unique number for any two positive integers, where x!=y, such that f(x,y) = f(y,x) but the resulting combined id would not be generated by any other pair of integers. What I came up with was a way to generate a unique key from any set of positive integers which is valid no matter the order, but which doesn't key to any other set.
My idea was to take the radius of a circle that intersected the integer pair in cartesian space. That alone doesn't guarantee the circle won't intersect any other integer pairs... so I had to add to it the phase multiple of sine and cosine which is the same at those two points on the arc. That works out to:
(x^2+y^2)+(sin(atan(x/y))*cos(atan(x/y)))
And means that it doesn't matter which order you feed x and y in, it will generate a unique float for the pair. It reduces to:
x^2+y^2+( (x/y) / (x^2+y^2) )
To add another dimension, just add it to the process and key it to one of the first...
It looks like you have typos?
(x^2+y^2)+(sin(atan(x/y))*cos(atan(x/y)))
reduces to
x^2+y^2+( (x/y) / (x^2/y^2 + 1) ) - not the equation given? Tho it's easier to see that this would be symmetrical if you rearrange it to:
x^2+y^2+( (xy) / (x^2+y^2) )
Also, if f(x,y) = x^2+y^2+( (x/y) / (x^2+y^2) )
then f(2,1) is 5.2 and f(1,2) is 5.1? - this is how I noticed the mistake. (the other reduction gives the same answer, 5.4, for both, by symmetry, as you suggest)
There's a simpler solution which produces integer ids (though they are large): 2^x & 2^y. Another solution is to multiply the xth and yth primes.
I only looked because I was curious how you proved it unique!
SO in 2013 was a different world from the SO of the 2020's. In the latter world your post would have been moderator classified as 'duplicate' of some basic textbook copy/pasted method posted by a karma grinding CS student and closed.
The question boils down to: can you simulate the bulk outcome of a sequence of priority queue operations (insert and delete-minimum) in linear time, or is O(n log n) necessary. Surprisingly, linear time is possible.
Reddit is my current go-to for human-sourced info. Search for "reddit your question here". Where on reddit? Not sure. I don't post, tbh, but I do search.
Has the added benefit of NOT returning stackoverflow answers, since StackOverflow seems to have rotted out these days, and been taken over by the "rejection police".
Even if LLMs were trained on the answer, that doesn't mean they'll ever recommend it. Regardless of how accurate it may be. LLMs are black box next token predictors and that's part of the issue.
Why did SO decide to do that to us? to not invest in ai and then, iirc, claim our contributions their ownership. i sometimes go back to answers i gave, even when answered my own questions.
AFAICT all they did is stop providing dumps. That doesn't change the license.
I was very active, In fact I'm actually upset at myself for spending so much time there. That said, I always thought I was getting fair value. They provided free hosting, I got answers and got to contribute answers for others.
While AI might have amplified the end, the drop-off preceded significant AI usage for coding.
So some possible reasons:
- Success: all the basic questions were answered, and the complex questions are hard to ask.
- Ownership: In its heyday, projects used SoF for their support channel because it meant they don't have to answer twice. Now projects prefer to isolate dependencies to github and not lose control over messaging to over-eager users.
- Incentives: Good SoF karma was a distinguishing feature in employment searches. Now it wouldn't make a difference, and is viewed as being too easy to scam
- Demand: Fewer new projects. We're past the days of Javascript and devops churn.
- Community: tight job markets make people less community-oriented
Some non-reasons:
- Competition (aside from AI at the end): SoF pretty much killed the competition in that niche (kind of like craigslist).
> - Success: all the basic questions were answered, and the complex questions are hard to ask.
I think this is one major factor that is not getting enough consideration in this comment thread. By 2018-2020, it felt like the number of times that someone else had already asked the question had increased to the point that there was no reason to bother asking it. Google also continued to do a better and better job of surfacing the right StackOverflow thread, even if the SO search didn't.
In 2012 you might search Google, not find what you needed, go to StackOverflow, search and have no better luck, then make a post (and get flamed for it being a frequently-asked question but you were phrasing yours in a different / incorrect way and didn't find the "real" answer).
In 2017, you would search Google and the relevant StackOverflow thread would be in the top few results, so you wouldn't need to post and ask.
In 2020, Google's "rich snippets" were showing you the quick answers in the screen real estate that is now used by the AI Overview answers, and those often times had surfaced some info taken from StackOverflow.
And then, at the very end of 2022, ChatGPT came along and effectively acted as the StackOverflow search that you always wanted - you could phrase your question as poorly as you want, no one would flame you, and you'd get some semblance of the correct answer (at least for simple questions).
I think StackOverflow was ultimately a victim of it's own success. Most of the questions that would be asked by your normal "question asker" type of user were eventually "solved" and it was just a matter of how easy it was to find them. Google, ChatGPT, "AI Overviews", Claude Code, etc have simply made finding those long-answered questions much easier, as well as answering all of the "new" questions that could be posed - and without all of the drama and hassle of dealing with a human-moderated site.
Not sure. As software becomes a commodity I can see the "old school" like tech slowing down (e.g. programming languages, frameworks frontend and backend, etc). The need for a better programming language is less now since LLM's are the ones writing code anyway more so these days - the pain isn't felt necessarily by the writer of the code to be more concise/expressive. The ones that do come out will probably have more specific communities for them (e.g. AI)
The graph is scary, but I think it's conflating two things:
1. Newbies asking badly written basic questions, barely allowed to stay, and answered by hungry users trying to farm points, never to be re-read again. This used to be the vast majority of SO questions by number.
2. Experiencied users facing a novel problem, asking questions that will be the primary search result for years to come.
It's #1 that's being canibalized by LLM's, and I think that's good for users. But #2 really has nowhere else to go; ChatGPT won't help you when all you have is a confusing error message caused by the confluence of three different bugs between your code, the platform, and an outdated dependency. And LLMs will need training data for the new tools and bugs that are coming out.
The newbies vastly outnumber the experienced people (in every discipline), and have more to ask per-capita, and are worse at asking it. Category 2 is much smaller. The volume of Stack Overflow was never going to be sustainable and was not reasonably reflective of its goals.
We are talking about a site that has accumulated more than three times as many questions as there are articles on Wikipedia. Even though the scope is "programming languages" as compared to "literally anything that is notable".
I’m going to argue the opposite. LLMs are fantastic at answering well posed questions. They are like chess machines evaluating a tonne of scenarios. But they aren’t that good at guessing what you actually have on your mind. So if you are a novice, you have to be very careful about framing your questions. Sometimes, it’s just easier to ask a human to point you in the right direction. But SO, despite being human, has always been awful to novices.
On the other hand, if you are experienced, it’s really not that difficult to get what you need from an LLM, and unlike on SO, you don’t need to worry about offending an overly sensitive user or a moderator. LLMs never get angry at you, they never complain about incorrect formatting or being too lax in your wording. They have infinite patience for you. This is why SO is destined to be reduced to a database of well structured questions and answers that are gradually going to become more and more irrelevant as time goes by.
Yes, LLMs are great at answering questions, but providing reasonable answers is another matter.
Can you really not think of anything that hasn't already been asked and isn't in any documentation anywhere? I can only assume you haven't been doing this very long. Fairly recently I was confronted with a Postgres problem, LLMs had no idea, it wasn't in the manual, it needed someone with years of experience. I took them IRC and someone actually helped me figure it out.
Until "AI" gets to the point it has run software for years and gained experience, or it can figure out everything just by reading the source code of something like Postgres, it won't be useful for stuff that hasn't been asked before.
And that is exactly why so many people gripe about SO being "toxic". They didn't present a well posed question. They thought it was for private tutoring, or socializing like on reddit.
All I can say to these is: Ma'am, this is a Wendy's.
I think you overestimate 2 by a longshot most problems only appear novel because they couched in a special field, framework or terminology, otherwise it would be years of incremental work. Some are, they are more appropriately put in a recreational journal or BB.
The reason the "experts" hung around SO was to smooth over the little things. This create a somewhat virtuous cycle, but required too much moderation and as other have pointed out, ultimately unsustainable even before the release of LLMs.
The first actually insightful comment under the OP. I agree all of it.
If SO manages to stay online, it'll still be there for #2 people to present their problems. Don't underestimate the number of bored people still scouring the site for puzzles to solve.
SE Inc, the company, are trying all kinds of things to revitalize the site, in the service of ad revenue. They even introduced types of questions that are entirely exempt from moderation. Those posts feel literally like reddit or any other forum. Threaded discussions, no negative scores, ...
If SE Inc decides to call it quits and shut the place down and freeze it into a dataset, or sell it to some SEO company, that would be a loss.
Another plausible explanation is that the new owners didn’t develop the community in a good way. Instead of fixing the myriad of issues that were obvious to almost all contributors they instead basically let it die?
The new owners (well, not really new any more) are focused on adding AI to SO because it's the current hotness, and making other changes to try to extract more money that they're completely ignoring the community's issues and objections to their changes, which tend to be half-assed and full of bugs.
Somewhere out there, there's an alternate universe in which the Stackoverflow community was so friendly, welcoming, helpful, and knowledgeable that this seems like a tragedy and motivates people to try to save it.
But in this universe, most people's reaction is just "lol".
Lots of the comments here are attributing the decline to a toxic community or overly-strict moderation, but I don't think that that is the main reason. The TeX site [0] is very friendly and has somewhat looser moderation, yet it shows the exact same decline [1].
A similar but not as strong decline. Taking the one but last datapoint for both (stackoverflow/tex respectively): 4436 and 394. If you compare this to how it looked like between 2015-2020 you get (my guess from scanning): 160,000 and 1700. So Stackoverflow as a whole went from 160K -> ~4.4K. That's like a 35x drop, compared to tex, where it's a 1700 -> 394, 4x drop.
Hate to argue with people on the internet, but your graph doesn't actually show what you claim. The TeX data was stable until late 2021, whereas the SO decline started in 2017. I also would expect some correlation so that SO was a drag on the TeX site.
I would ascribe that to these communities evolving differently. There is no reason to assume that the popularity of LaTeX tracks the popularity of programming languages. It's a type setting system. And that doesn't even take into account communities that exist parallel to SO/SE. Surely there exist communities today for LaTeX that have been around since before SO began its life.
The obvious culprit here are the LLMs, but I do wonder whether Github's social features, despite its flaws, have given developers fewer reasons to ask questions on SO?
Speaking from experience, every time I hit a wall with my projects, I would instinctively visit the project's repo first, and check on the issues / discussions page. More often than not, I was able to find someone with an adjacent problem and get close enough to a solution just by looking at the resolution. If it all failed, I would fall back to asking questions on the discussion forum first before even considering to visit SO.
I used to contribute a ton to Stack Overflow at the beginning in 2009 and 2010 and then stopped cold turkey. One of the senior product execs emailed me to see what turned me off.
What killed it for me was community moderation. People who cannot contribute with quality content will attempt to contribute by improperly and excessively applying their opinion of what is allowed.
Unfortunately it happens to every online technical community once they become popular enough. I even see it happening on HN.
Fellow OG! And it's been happening on HN since the mid-2010s, too. Moderation went out of control everywhere, but at least this site isn't branded as some strictly technical site. Can't believe I'm even saying this, but moderation on a site that encompasses a cornucopia of topics is their prerogative. The mind-boggling thing to me about SO was that the moderation used non-technical criteria (such as failing to recognize why certain problems were asked and were not dupes, and later to shoehorn in political and sexual ideologies) to shape a technical site.
HN has full time staff that provides moderation and does an excellent job. Nonetheless there are numerous users who take it upon themselves to determine what content should be available to the rest of us, as if they were heroes in their own mind.
It’s a form of narcissism. While they think of themselves as community saviors everyone else thinks they are censoring assholes. Just let the moderators do their job. Unwanted content will naturally fall off either by downvoting or it will be ignored.
All the rest of ask for is just don’t be an asshole.
I feel like I'm taking crazy pills, reading some of these comments.
It looks like a pretty clear divide between the people that wanted to ask questions to get solutions for their own specific problems; and those who were aware of what the site wanted to be and how it actually operated, and were willing to put in the time and answer questions, etc.
The sheer amount of garbage that used to get posted every day required some pretty heavy moderation. Most of it was not by actual moderators, it was by high-reputation users.
(I have 25K reputation on StackOverflow, and was most active between 2011 and 2018.)
And half the garbage is from people "moderating"! You are literally rewarded points for doing moderating activities, so of course every post is flooded with BS edits, votes to close, etc.. Cobra effect and whatnot.
As someone that spent a fair bit of time answering questions on StackOverflow, what stood out years ago was how much the same thing would be asked every day. Countless duplicates. That has all but ceased with LLMs taking all that volume. Honestly, I don't think that's a huge loss for the knowledge base.
The other thing I've noticed lately is a strong push to get non-programming questions off StackOverflow, and on to other sites like SuperUser, ServerFault, DevOps, etc.
Unfortunately, what's left is so small I don't think there's enough to sustain a community. Without questions to answer, contributors providing the answers disappear, leaving the few questions there often unanswered.
I do use Claude a lot, but I still regularly ask questions on https://bioinformatics.stackexchange.com/. It's often just too niche, LLMs hallucinate stuff like an entire non-existent benchmarking feature in Snakemake, or can't explain how I should get transcriptome aligners to give me correct quantifications for a transcript. I guess it's too niche. And as a lonely Bioinformatician it can be nice to get confirmation from other bioinformaticians.
Looking back at my Stack Exchange/Stack Overflow (never really got the difference) history, my earlier, more general programming questions from when I just started are all no-brainers for any LLM.
The decline is not surprising. I am sure AI is replacing Stackoverflow for a lot of people. And my experience with asking questions was pretty bad. I asked a few very specific questions about some deep detail in Windows and every time I got only some smug comments about my stupid question or the question got rejected outright. That while a ton of beginner questions were approved. Definitely not a very inviting club. I found i got better responses on Reddit.
I joined Stackoverflow early on since it had a prevalence towards .NET and I’ve been working with Microsoft web technologies since the mid 90’s.
My SO account is coming up to 17 years old and I have nearly 15,000 points, 15 gold badges, including 11 famous questions and similar famous answer badges, also 100 silver and 150 bronze. I spent far much time on that site in the early days, but through it, I also thoroughly enjoyed helping others. I also started to publish articles on CodeProject and it kicked off my long tech blogging “career”, and I still enjoy writing and sharing knowledge with others.
I have visited the site maybe once a year since 2017. It got to the point that trying to post questions was intolerable, since they always got closed. At this point I have given up on it as a resource, even though it helped me tremendously to both learn (to answer questions) and solve challenging problems, and get help for edge cases, especially on niche topics. For me it is a part of my legacy as a developer for over 30 years.
I find it deeply saddening to see what it has become. However I think Joel and his team can be proud of what they built and what they gave to the developer community for so many years.
As a side note it used to state that was in the top 2% of users on SO, but this metric seems to have been removed. Maybe it’s just because I’m on mobile that I can’t see it any more.
LLM’s can easily solve those easy problems that have high commonality across many codebases, but I am dubious that they will be able to solve the niche challenging problems that have not been solved before nor written about. I do wonder how those problems get solved in the future.
Apparently it was removed to reduce the load on the database see [0], [1].
The top voted response points out that SO are [2]:
> destroying a valuable feature for users.
Kinda wild they allowed it. As that answer also suggests, perhaps rather than remove it entirely, they could just compute those stats at a lesser frequency to reduce load.
i've been using SO for 17 years as well but ultimately gave up out of frustration, and a lot of comments here are correctly pointing at the toxicity but the real-time chats were on a next level, it was absolutely maddening how toxic and aggressive these moderators were.
A lesson can be learned here. If you don't introduce some form of accountability for everyone that influences the product, it eventually falls apart. The problem, as we all know now, is that the moderators screwed things up, and there were no guardrails in place to stop them from killing the site. A small number of very unqualified moderators vandalized the place and nobody with common sense stepped in to put an end to it.
Interestingly, stagnation started around 2014 (in the number of questions asked no longer rising,) and a visible decline started in 2020 [1]: two years before ChatGPT launched!
It’s an interesting question if the decline would have happened regardless of LLMs, just slower?
Some commenters suggest it's not the moderation. I think it is the key problem, and the alternative communities were the accumulated effect. Bad questions and tough answer competition is part of it, but moderation was more important, I think. Because in the end what kept SO relevant was that people made their own questions on up to date topics.
Up until mid-2010s you could make a seriously vague question, and it would be answered, satisfactory or not. (2018 was when I made the last such question. YMMV) After that, almost everything, that hadn't snap-on code answer, was labelled as offtopic or duplicate, and closed, no matter what. (Couple of times I got very rude moderators' comments on the tickets.)
I think this lead some communities to avoid this moderator hell and start their own forums, where you could afford civilized discussion. Discourse is actually very handy for this (Ironically, it was made by the same devs that created SO). Forums of the earlier generation, have too many bells and whistles, and outdated UI. Discourse has much less friction.
Then, as more quality material was accumulated elsewhere, newbies stopped seeing SO on top of search, and gradually language/library communities churned off one by one. (AI and other summaries, probably did contribute, but I don't think they were the primary cause.)
One might even say that Joel and Jeff created their very own puts on shades coding horror yeaaahhhh with SO, and it is indeed ironic that them building Discourse has created far better communities than SO ever was.
Given the fact that when I need a question answered I usually refer to S.O., but more recently have taken suggestions from LLM models that were obviously trained on S.O. data...
And given the fact that all other web results for "how do you change the scroll behavior on..." or "SCSS for media query on..." all lead to a hundred fake websites with pages generated by LLMs based on old answers.
Destroying S.O. as a question/answer source leaves only the LLMs to answer questions. That's why it's horrific.
SO has lost against LLMs because it has insistently positioned itself as a knowledge base rather than a community. The harsh moderation, strict content policing, forbidden socialization, lack of follow mechanics etc have all collectively contributed to it.
They basically made a bet because they wanted to be the full anti-thesis of ad-ridden garbage-looking forums. Pure information, zero tolerance for humanity, sterile looking design.
They achieved that goal, but in the end, they dug their own grave too.
LLMs didn’t admonish us to write our questions better, or simply because we asked for an opinion. They didn’t flag, remove our post with no advance notice. They didn’t forbid to say hello or thanks, they welcomed it. They didn’t complain when we asked something that was asked many times. They didn’t prevent us from deleting our own content.
Oh yeah, no wonder nobody bothers with SO anymore.
Reddit is a forum morphed into social media. I usually use "question + reddit" on Google to confirm my suspicions about a subject. It is a place to discuss things rather than find answers. It is extremely politicized (leftist/liberal), but that's a whole other story.
Even if somehow that is "deserved", that's not a healthy ecosystem.
All that is left of SI are clueless questioners and bitter, jaded responders.
SO worked when "everyone" was new to it, and they felt energized to ask questions ( even "basic" questions, because they hadn't been asked before ), and felt energized to answer them.
SO solved a real problem - knowledge being locked into forum posts with no follow-up, or behind paywalls.
It is not "karma". It is not to be taken personally. It represents the objective usefulness of the question, not the personal worth of the person asking it.
It's both. I stopped asking questions because the mods were so toxic, and I stopped answering questions because I wasn't going to train the AI for free.
It's not zero but it's very low. You can glance at the site now for confirmation.
I was using the site recently (middle of a US workday) and the "live stats" widget showed 10s of questions asked per hour and ~15K current users. I have not done the work to compare these values to historical ones but they're _low_.
The last data point is from January 2026, which has just begun. If you extrapolate the 321 questions by multiplying by 10 to account for the remaining 90 % of the month, you get to within the same order of magnitude as December 2025 (3862). The small difference is probably due to the turn of the year.
AI didn't necessarily kill SO because it was strictly better at giving technical answers (and it certainly wasn't better when GPTs initially burst onto the mass-appeal scene several years ago), but that it provided an alternative (even if subpar) where users could actually get responses to their questions (and furthermore not be ridiculed by psychopaths while doing so was the cherry on top).
In the past people asked questions of real people who gave answers rooted in real use. And all this was documented and available for future learning. There was also a beautiful human element to think that some other human cared about the problem.
Now people ask questions of LLMs. They churn out answers from the void, sometimes correct but not rooted in real life use and thought. The answers are then lost to the world. The learning is not shared.
LLMs have been feeding on all this human interaction and simultaneously destroying it.
StackOverflow was a pub where programmers had fun while learning programming. The product of that fun was valuable.
Instead of cultivating the pub, the owners demanded that the visitors be safe, boring and obedient witers of value. This killed the pub and with it the business.
The most visible aspect was the duplicate close. Duplicate closes scare away fresh patrons, blocking precisely the path that old timers took when they joined. And duplicates allow anyone with a grudge to take revenge. After all, there are no new questions, and you will always find a duplicate if you want to.
To create a new Stack Overlflow, create a pub where programmers enjoy drinking a virtual beer, and the value will appear by itself.
Another exemple being "Comments are not for extended discussion ! if you want to actively bring value by adding information, later updates, history, or just fun that cultivates a community, please leave and go do that somewhere else like our chat that doesn't follow at all the async functionnality of this platform and is limited to the regular userbase while scaring the newcomers."
"comments are not for extended discussion" is one of the biggest own goals of SO product development. Like, they had a feature that people were engaging with actively, and the discussions were adding value and additional context to posts, and they decided "yeah, let's kill this".
The people who run SO have some sort of control-freak complex. If there's anything I've learned from the SO saga, it is that oftentimes just letting a community do what it wants (within reasonable boundaries, of course) leads to a better and more successful product than actively trying to steer things in a certain direction.
hahaha, I almost forgot about that! "stop talking about edge cases and other things pertinent to this topic in comments about this topic!! reeeeeeeee!!!!"
>StackOverflow was a pub where programmers had fun while learning programming. The product of that fun was valuable.
I really like this description. I and others here who are talking about negative experiences there seem to decry how we enjoy programming (you see words like "fun" and "passion" used in these posts), and how SO decided to take this good faith and cheer and bludgeon users for often opaque reasons, just so they could power trip. As much as I have many reservations about LLMs, I can ask LLMs to be as emotionless (or even emotional but chipper/happy) as I want. On SO, you needed to prostrate yourself and self-criticize to even have the opportunity to be bludgeoned further by the moderators. Who tf would want to spend their time contributing there? Even if you contributed a decent or even great amount to the site, you would still get whacked over the head if you dared to ask a question of your own.
This is why people jumped to LLMs, even when they were far less capable than they are now. Most people (SO moderators don't view others as "people", as is apparent in this thread) would rather receive mid-tier answers from an LLM (though LLMs have now exceeded this level of quality) while still having fun, than get castigated and "closed as duped" on SO.
SO was built to disrupt the marriage of Google and Experts Exchange. EE was using dark patterns to sucker unsuspecting users into paying for access to a crappy Q&A service. SO wildly succeeded, but almost 20 years later the world is very different.
I recall when they disabled the data export a few years ago [0], March 2023. Almost certainly did this in response to the metrics they were seeing, but it accelerated the decline [1].
As one of my good friends pointed out back in 2012, most people don't know how to ask questions[0].
I'm feeling a bit sorry for zahlman in the comment section here, they're doing a good job of defending SO to a comment section that seems to want SO to bend to their own whims, no matter what the stated aims and goals of SO really were. There does seem to be a lot of people in the comments here who wanted SO to be a discussion site, rather than the Q&A site that it was set out to be.
I do think it's very unfair of many of you who are claiming SO was hostile or that they unfairly closed questions without bringing the citations required. I'm not saying at all that SO was without it's flaws in leadership, moderators, community or anything else that made the site what it was. But if you're going to complain, at least bring examples, especially when you have someone here you could hold somewhat accountable.
The problem is, you still see a lot of it today, whether it's in IRC channels, Discord chats, StackOverflow or GitHub issues. People still don't know how to ask questions:
One factor I haven't seen mentioned is the catastrophic decline in quality of Google search. That started pre-llm and now the site is almost unusable to search web. You can access something you know exists and you know where it exists, but to actually search..?
Most SO users are passive readers who land there using search, but these readers are also the feed of new active users. Cut off the influx, and the existing ones will be in decline (the moderation just accelerates it).
As everyone is saying, it was already down-trending before AI, and probably experts exchange traffic and whatever came before looks similar
Also not sure exactly when they added the huge popup[0] that covers the answer (maybe only in Europe as it's about cookies?) but that's definitely one of the things that made me default reach for other links instead of SO.
Those popups were a big contributor for me to stop using SO. I stopped updating my uBlock origin rules when LLMs became good enough. I am now using the free Kimi K2 model via Groq over CLI, which is much faster.
Here’s how SO could still be useful in the LLM era:
User asks a question, llm provides an immediate answer/reply on the forum. But real people can still jump in to the conversation to add additional insights and correct mistakes.
If you’re a user that asks a duplicate question, it’ll just direct you to the good conversation that already happened.
A symbiosis of immediate usually-good-enough llm answers PLUS human generated content that dives deeper and provides reassurances in correctness
Users could upvote whether Claude, Gemini or ChatGPT provided the best answer. The best of three is surfaced, the others are hidden behind a "show alternatives."
However, I can see how this would be labelled "shoving AI into everything" and "I'm not on SO for AI."
Stackoverflow is like online gaming--lots of toxic people, but I still get value out of it. Ignore the toxic people, get your questions answered and go home to your family with your paycheck.
It's surprisingly tame still given it interests tens (hundreds?) of millions of people at varying age and background and mostly when the mind is occupied by a problem. I always found it surprising there's not more defacing and toxicity.
I was tasked to add OpenOffice's hyphenation lib to our software at work back in 2010 when I was a junior dev. I had to read the paper and the C code/documentation to understand how it works but got stuck in one particular function.
It was such an obscure thing (compare to web dev stuffs) that I couldn't find anything on Google.
Had no choice but to ask on Stackoverflow and expected no answers. To my surprise, I got a legit answer from someone knowledgable, and it absolutely solve my problem at the time. (The function has to do with the German language, which was why I didn't understand the documentation)
AI is a vampire. Coming to your corner of the world, to suck your economic blood, eventually. It’s hard to ignore the accelerated decline that started in late 2022/early 2023.
LLMs absolutely body-slammed SO, but anyone who was an active contributor knows the company was screwing over existing moderators for years before this. Writing was on the walls
If by "body-slammed" you mean "trained on SO user data while violating the terms of the CC BY-SA license", then sure.
In the best case scenario, LLMs might give you the same content you were able to find on SO. In the common scenario, they'll hallucinate an answer and waste your time.
What should worry everyone is what system will come after LLMs. Data is being centralized and hoarded by giant corporations, and not shared publicly. And the data that is shared is generated by LLMs. We're poisoning the well of information with no fallback mechanism.
> If by "body-slammed" you mean "trained on SO user data while violating the terms of the CC BY-SA license", then sure.
You know that's not what they meant, but why bring up the license here? If they were over the top compliant, attributing every SO answer under every chat, and licensing the LLM output as CC BY-SA, I think we'd still have seen the same shift.
> In the best case scenario, LLMs might give you the same content you were able to find on SO. In the common scenario, they'll hallucinate an answer and waste your time.
Best case it gives you the same level of content, but more customized, and faster.
SO being wrong and wasting your time is also common.
So the question for me is how important was SO to training LLMs? Because now that the SO is basically no longer being updated, we've lost the new material to train on? Instead, we need to train on documentation and other LLM output. I'm no expert on this subject but it seems like the quality of LLMs will degrade over time.
Why publish anything for free on the internet if it's going to be scanned into some corporation's machine for their free use? I know artists who have stopped putting anything online. I imagine some programmers are questioning whether or not to continue with open source work too.
It has often been claimed, and even shown, that training LLMs on their own outputs will degrade the quality over time. I myself find it likely that on well-measurable domains, RLVR improvements will dominate "slop" decreases in capability when training new models.
Are we in the age of all CS problems being solved and everything being invented? Even if so, do LLM incorporate all that knowledge?
A lot of my knowledge in CS come from books and lectures, LLMs can shine in that area by scraping all those sources.
However SO was less about academic knowledge but more about experience sharing. You won't find recipes for complex problems in books, e.g. how to catch what part of my program corrupts memory for variable 'a' in gdb.
LLMs know correct answer to this question because someone shared their experience, including SO.
Are we Ok with stopping this process of sharing from one human to another?
it is indeed a shame. if you are doing anything remotely new and novel, which is essential if you want to make a difference in an increasingly competitive field, LLMs confidently leave you with non-working solutions, or sometimes worse they set you on the wrong path.
I had similar worries in the past about indexable forums being replaced by discord servers. the current situation is even worse.
LLMs caused this decline. Stop denying that. You don't have to defend LLMs from any perceived blame. This is not a bad thing.
The steep decline in the early months of 2023 actually started with the release of ChatGPT, which is 2022-11-30, and its gradually widening availability to (and awareness of) the public from that date. The plot clearly shows that cliff.
The gentle decline since 2016 does not invalidate this. Were it not for LLMs, the site's post rate would now probably be at around 5000 posts/day, not 300.
LLMs are to "blame" for eating all the trivial questions that would have gotten some nearly copy-pasted answer by some eager reputation points collector, or closed as a duplicate, which nets nobody any rep.
Stack Overflow is not a site for socializing. Do not mistake it for reddit. The "karma" does not mean "I hate you", it means "you haven't put the absolute minimum conceivable amount of effort into your question". This includes at least googling the question before you ask. If you haven't done that, you can't expect to impose on the free time of others.
SO has a learning curve. The site expects more from you than just to show up and start yapping. That is its nature. It is "different" because it must be. All other places don't have this expectation of quality. That is its value proposition.
For those who miss SO, check out Stack Overflow Simulator: A functional museum for developers to relive the good ol' days of asking innocent questions and being told to "RTFM"
Game over. I didn’t notice all the toxicity mentioned in the other comments, although I did stop using it around 2016 maybe. It had its days, it was fundamentally a verb at some point. Its name is part of web history, and there’s no denying that.
They're desperately trying to save it e.g. by introducing "discussions" which are just questions that would normally have been closed. The first one I saw, the first reply was "this should have been a question instead of a discussion".
Let's never forget that Stackoverflow was killed by its mods. Sure, it needed AI as an alternative so people could actually leave, but the thing that actually pushed them away was the mods.
Maybe it's a mix of me using the site less, or questions I previously answered not being as relevant anymore, however as it stands, it's just not fun to visit the site any more.
I have about ~750 answers and 24K rep after almost 12 years of being a member. The site was a great way to spend some free cycles and help people. My favorite bounty answer lead to me finding a bug in the Java compiler! I even got recruited into my current role from the old Stack Overflow Jobs board.
With AI, not only did the quality and frequency of posts go down, but the activity on my existing posts are basically zero now. I used to have a few notifications a week with either comments on my past answers/questions or a few upvotes (for those fun little serotonin boosts). Looking at my past stats.. in 2023 I had ~170 notifications, in 2024 that dropped to ~100, and in 2025 it went down to ~50 (with only 5 notifications since September).
I don't feel engaged with the community, and even finding new questions to answer is a struggle now with (the unanswerable) "open-ended questions" being mixed into the normal questions feed.
This is concerning on two fronts. The questions are no longer open (SO is CC-BY-SA) and if Q&A content dies then this herds even more people towards LLM use.
It's basically draining the commons.
Yup. This, to me, provides another explanation for why the social contract is being used as toilet paper by the owner class. They literally see the writing on the wall.
For this occasion, I just logged in to my SO profile; I've been a member for 9 years now.
To me, back when I started out learning web dev, as a junior with no experience and barely knowing anything, SO seemed like a paradise for programmers. I could go on there and get unblocked for the complex (but trivial for experts) issues I was facing. Most of the questions I initially posted, which were either closed as duplicates or "not good enough," really did me a lot of discouragement. I wasn't learning anything by being told, "You did it wrong, but we're also not telling you how you could do it better." I agree with the first part; I probably sucked at writing good questions and searching properly. I think it's just a part of the process to make mistakes but SO did not make it better for juniors, at least on the part of giving proper guidance to those who "sucked".
Was already dying a decade ago, but AI pretty much guarantees we'll never see a public forum that useful ever again.
AI may be fine for people asking the basic stuff or who don't care about maintenance, but for a time SO was the place to find extraordinary and creative solutions that only a human can come up with.
When you were in a jam and found a gem on there it not only solved your problem, but brought clarity and deep knowledge to your entire situation in ways that I've never seen an LLM do. It inspired you to refactor the code that got you into that mess to begin with and you grew as a developer.
This timeline shows the death of a lot more than just the site itself.
I just looked it up, and "Note: This tag is currently blacklisted and can no longer be used." lmfao, what a braindead site, so glad I left years ago, after many years of greatly reduced activity. Source (at the top):
https://stackoverflow.com/questions/tagged/code-golf?tab=New...
You do realize that they made an entire site for CodeGolf that is reasonably active (and has its own culture... and lead to the creation of specialized languages for it... and even pushed the bounds of OEIS a few times - https://codegolf.meta.stackexchange.com/q/5318)?
I ended up having a high reputation on SO. Not sure why, but it’s over 7000.
I also experienced many of the issues I see described here. The most egregious was when I asked a completely valid question for R: How to fit a curve through a set of points, with each point having an error associated.
This is something completely normal in a physics experiment. Each measurement had its own error interval. But, for people using R, this seemed like something completely new. So, they just downvoted the question and told me I was wrong.
I ended up answering my own question… but was also told that was wrong and that all points must have the same error interval.
Instead of answering a programming question, people just went around denying experimental physics.
I think that was the beginning of the end of SO for me.
This entire thread is fantastic. I felt nostalgic, angry and then concerned all at once.
I love LLMs. But I miss SO. I miss being able to have that community. How do we bring it back?
If anyone from the Stack Overflow team is reading this (I assume you are): what’s the plan?
My take: stop optimizing for raw question volume and start optimizing for producing and maintaining “known good” public knowledge. The thing SO still has that Discord and LLMs don’t is durable, linkable, reviewable answers with accountable humans behind them. But the workflow needs to match how devs work now.
A concrete idea: make “asking” a guided flow that’s more like opening a good GitHub issue. Let me paste my error output, environment, minimal repro, what I tried, and what I think is happening. Then use tooling (including an LLM if you want) to pre check duplicates, suggest missing details, and auto format. Crucially: don’t punish me for being imperfect. Route borderline questions into a sandbox or draft mode where they can be improved instead of just slammed shut.
Second idea: invest hard in keeping answers current. A ton of SO is correct but stale. Add obvious “this is old” signaling and make it rewarding to post updates, not just brand new answers.
Last thing that I don’t see an easy answer to: LLMs are feasting on old SO content today. But LLMs still need fresh, high quality, real world edge cases tomorrow. They need the complexity and problem solving that humans provide. A lot of the answers I get are recycled. No net new thinking. If fewer people ask publicly, where does that new ground truth come from? What’s the mechanism that keeps the commons replenished?
So… TLDR…my question to this group of incredibly intelligent people: how does SO save itself?
One thing you won’t get with in an LLM is genuine research. I once answered a 550 point question by researching the source code of vim to see how the poster’s question could be resolved. [0]
People are mentioning the politicization of moderation. But also don’t forget when Joel broke the rules to use the site to push his personal political agenda.
Are there any publicly available options to actually interact with real people about software development anymore? There doesn't seem to be anywhere that's accessible with something like a google search... Sure there are derelict IRC/Discord/$language forums, but of the handful I've been part of they aren't active or in the case of discord, weirdly disjointed.
AI is great and all, but somewhere with a little bit of an opinion and push back to carelessly thrown out questions would be nice (as a thrower of careless questions).
SO obviously went off the toxic deep end, but has that culture of shared problem solving just died completely online?
Good riddance. There were some ok answers there, but also many bad or obsolete answers (leading to scrolling down find to find the low-ranked answer that sort of worked), and the moderator toxicity was just another showcase of human failure on top of that. It selected for assholes because they thought they had a captive, eternally renewing audience that did not have any alternative.
And that resulted in the chilling effect of people not asking questions because they didn't want to run the moderation gauntlet, so the site's usefulness went even further down. Its still much less useful for recent tech, than it is for ancient questions about parsing HTML with regex and that sort of thing.
LLMs are simply better in every way, provided they are trained on decent documents. And if I want them to insult me too, just for that SO nostalgia, I can just ask them to do that and they will oblige.
Looking forward to forgetting that site ever existed, my brain's health will improve.
GitHub Issues and Disscussions + searching the code base, fetching the docs and some reasoning on top. Maybe even firing up a sandbox VM and testing some solutions.
"firing up a sandbox VM and testing some solutions"
If the LLM can start up a VM and test a solution, to identify a new unique problem, and find it's own solution. That would be pretty impressive. I'm not sure they are really to that point. But some AI's are winning the Math Olympiad, so maybe it is happening. I'm sure this is the overall goal.
Everyone agrees their community and moderators turned toxic. But why? Was it inevitable that people would turn bitter / jaded after answering questions for years? Was it wrong incentives from StackOverflow itself? The outside tech environment becoming worse?
The precipitous decline was already happening long before LLM's dealt the final blow.
Yes, it was intended by SO itself. Basically moderate mercilessly. See posts by Jeff Atwood:
> Avoid asking questions that are subjective, argumentative, or require extended discussion. This is not a discussion board, this is a place for questions that can be answered!
> Stack Overflow – like most online communities I’ve studied – naturally trends toward increased strictness over time. It’s primarily a defense mechanism, an immune system of the sort a child develops after first entering school or daycare and being exposed to the wide, wide world of everyday sneezes and coughs with the occasional meningitis outbreak. It isn’t always a pleasant process, but it is, unfortunately, a necessary one if you want to survive.
> All the content on the site must exist to serve the mission of learning over entertainment – even if that means making difficult calls about removing some questions and answers that fail to meet those goals, plus or minus 10 percent.
I think the biggest issue, what lead to the toxicity, came down to the question/answer format not suiting the problem it was trying to solve — The answer could only be as good as the original question, and the platform gave little leeway to "get to the bottom" of the problem. Getting to a high-quality question/response required a back-and-forth that the platform made difficult by burying the discovery/definition work in comments and edits instead of a clear discussion mechanism.
All of this meant the learning-curve on how to participate was high, and this spurred gate-keeping and over-zealous moderation. High-quality but out-of-date information was preferred over lower-quality but more recent updates. When combined with the rapid shifts brought on with mobile development and web frameworks, the answers could easily get out-of-date months after being answered.
I remember a time when StackOverflow dominated every search query. Now we're seeing searches take you to a dedicated forum/discussion board, which feels more appropriate for the current state of the industry.
I still would like to get other humans' experiences and perspectives when it comes to solving some problems, I hope SO doesn't go away entirely.
With LLMs, at least in my experience, they'll answer your question best they can, just as you asked it. But they won't go the extra step to make assumptions based on what they think you're trying to do and make recommendations. Humans do that, and sometimes it isn't constructive at all like "just use a different OS", but other times it could be "I don't know how to solve that, but I've had better lack with this other library/tool".
Don't lose sight of one of the dreams of the early Internet: How do we most effectively make a marketplace for knowledge problems and solutions that connects human knowledge needs with AI and human responses?
It should be possible for me to put a question out there (not on any specific forum/site specific to the question), and have AI resource answer it and then have interested people weigh in from anywhere if the AI answer is unsatisfactory. Stackoverflow was the best we could do at the time, but now more general approach is possible.
StackOverflow didn't feel like a welcoming and humane place the last 10+ years, at least for me.
Actually I think it never did.
It started when I was new there and couldn't write answers, just write comments and then got blasted for writing answer-like comments as comments. What was I supposed to do? I engaged less and less and finally asked them to remove my account.
And then it seems like the power-users/moderators just took over and made it even more hostile.
I hope Wikipedia doesn't end up like this despite some similarities.
Where will LLMs be trained if no-one generates new posts and information like this? Do we sort of just stop innovating here in 2026? Probably not but it's a serious consideration.
Ideally, you'd train them on the core documentation of the language or tool itself.
Hopefully, LLMs lead to more thorough documentation at the start of a new language, framework, or tool. Perhaps to the point of the documentation being specifically tailored to read well for the LLM that will parse and internalize it.
Most of what StackOverflow was was just a regurgitation of knowledge that people could acquire from documentation or research papers. It obviously became easier to ask on SO than dig through documentation. LLMs (in theory) should be able to do that digging for you at lightning speed.
What ended up happening was people would turn to the internet and Stack Overflow to get a quick answer and string those answers together to develop a solution, never reading or internalizing documentation. I was definitely guilty of this many times. I think in the long run it's probably good that Stack Overflow dies.
Before writing the comment I had in my head I did a CTRL+F search for "toxic" in the comment section here. 42 occurences. It says everything about what's happening to SO.
Now imagine what happens when a new programming language comes along. When we have a question, we will no longer be able to Google it and find answers to it on Stack Overflow. We will ask the LLMs. They will work it out. From that moment, the LLM we used has the knowledge for solving this particular problem. Over time, this produces huge moat for the largest providers. I believe it is one of the subtler reasons why the AI race is so fierce.
I find this quite worrying: with this much decline SO might end up disappearing. This would be a very bad thing because in some answers there are important details and nuances that you only see by looking at secondary answers and comments. Also, this seems to imply that most people will just accept the solutions proposed by LLMs without checking them, or ever talking about the subject with other humans.
It's unfortunate that SO hasn't found a way to leverage LLMs. Lots of questions benefit from some initial search, which is hard enough that moderators likely felt frustrated with actual duplicates, or close enough duplicates, and LLMs seem able to assist. However I hope we don't lose the rare gem answers that SO also had, those expert responses that share not just a programming solution but deeper insight.
I think that SO leveraging LLMs implicitly. Like I'll always ask LLM first, that's the easiest option. And I'll only come to SO if LLM fails to answer.
Interesting timing. I just analyzed TabNews (Brazilian dev community) and ~50% of 2025 posts mention AI/LLMs. The shift is real.
The 2014 peak is telling. That's before LLMs, before the worst toxicity complaints. Feels like natural saturation, most common questions were already answered. My bet, LLMs accelerated the decline but didn't cause it. They just made finding those existing answers frictionless.
Good times. Although, I have to say, I was getting sick of SO before the LLM age. Modding felt a bit tyrannical, with a fourth of all my questions getting closed as off topic, and a lot of aggressive comments all around the site (do your homework, show proof, etc.)
Back when I was an active member (10k reputation), we had to rush to give answers to people, instead of angrily down voting questions and making snark comments.
SO has been a curse on technology. I've met teams of people who decide whether to adopt some technology based solely on if they can find SO answers for it. They refuse to read documentation or learn how the technology works; they'll only google for SO answers, and if the answer's not there, they give up. There's an entire generation like this now.
I used to joke that when SO goes under, I will move professions. The joke came from my experience of how many common issues in technology could not be solved with knowledge found via a search engine. I don’t see that niche as gone, so I wonder what is satisfying that requirement such that new questions do not show up at SO?
Why did SO traffic halve from it's maximum till the ChatGPT release date? Also, for a long time after initial release, ChatGPT was pretty much useless for coding questions. It only became more or less useful ~2 years ago. So there's about 4x decline from peak to explain with reasons that do not involve LLMs. What these could be?
There's no doubt that generally LLMs are better. In addition SO had its issues. That being said I can't help but worry about losing humans asking questions and humans answering questions. The sentimentality aside, if humans aren't posing questions and if humans aren't recommending answers, what are the models going to use?
Stack Overflow set out to be a better Q&A site but has turned into a user-unfriendly, gatekeeping platform where questions are often marked as duplicates because a similar question was answered 15 years ago. Everything and every question is banned, gatekeep, or marked as a duplicate.
For me, my usage of SO started declining as LLMs rose. Occasionally I still end up there, usually because a chat response referenced a SO thread. I was willing to put up with the toxicity as long as the site still had technical value for me.
But still, machines leave me wanting. Where do people go to ask real humans novel technical questions these days?
> Where do people go to ask real humans novel technical questions these days?
I don't think such generic place exists. I just do my own research or abandon the topic. I think that in big companies you probably could use some internal chats or just ask some smart guy directly? I don't have that kind of connections and all online communities are full of people whose skill is below mine, so it makes little sense to ask something. I still do sometimes, but rarely receive competent answer.
If you have some focused topic like a question about small program, of course you can just use github issues or email author directly. But if you have some open question, probably SO is the only generic platform out there.
To put it differently, find some experts and ask which online place to the visit to help strangers. Most likely they just don't do it.
So for me, personally, LLMs are the saviour. With enough forth and back I can research any topic that doesn't require very deep expertise. Sure, access to an actual expert willing to guide me would be better, but I just don't have that luxury.
I think one of the phenomenon that people haven't mentioned is that the question space was heavily colonized by 2016.
I was one of the top 30 or 50 answerers for the SVG tag on SO, and I found that the question flow started to degrade around 2016, because so many of the questions asked had been answered (and answered well) already.
While I generally agree with the narrative of the negative arc that stack overflow took, I found (and have as recently as a few months ago) that I could have enjoyable interactions on the math, Ux, written language, and aviation exchanges. The OS ones in the middle (always found the difference between Linux and superuser confusing).
Obviously LLMs ate StackOverflow, but perhaps developers could keep it alive for much longer if they wanted to. LLMs provide answers, but only humans provide human contact.
And that last part is where SO failed by allowing a few people power trip over the rest of us. Kind of like reddit does at times, but harder.
I am surprised at the amount of hate for Stack Overflow here. As a developer I can't think of a single website that has helped me as much over the last ten years.
It has had a huge benefit for the development community, and I for one will mourn its loss.
I do wonder where answers will come from in the future. As others have noted in this thread, documentation is often missing, or incorrect. SO collected the experiences of actual users solving real problems. Will AI share experiences in a similar way? In principle it could, and in practice I think it will need to. The shared knowledge of SO made all developers more productive. In an AI coded future there will need to be a way for new knowledge to be shared.
It makes sense to see the number of questions decline over time as people google questions and get results. It would be interesting to look at the number of comments and views of questions over time to see if that has declined as LLMs have driven declining engagement and discussion.
There are still airgapped places in the world where transferring information to offsite LLMs is expressly forbidden, but the offline LLMs available perform so terribly that they’re not worth using. An SO type application can be immensely helpful for engineering teams working in these environments.
Stack overflow was useful with a fairly sanitized search like “mysql error 1095”. Agentic LLMs do there best work when able to access your entire repository or network environment for context, which is impossible to sanitize. For a season, private environments will continue to be able to use SO. But as LLMs capture all the good questions and keep them private, public SO will become less and less relevant. It’s sad to see a resource of this class go.
I think the bigger point we should realize is LLMs offer the EXACT same thing in a better way. Many people are still sharing answers to problems but they do it through an AI which then fine tunes on it and now that problem solution is shared with EVERYONE.
Wonder if this is a good proxy for '# of Google Searches'. Or perhaps a forward indicator (sign of things to come), since LLMs are adopted by the tech-savvy first, then the general public a little later, so Stack Overflow was among the first casualties.
I'd still use SO at times if it weren't for how terribly it was managed and moderated. It offers features that LLMs can't, and I actually enjoyed answering questions enough to do it quite often at one time. These days I don't even think about it.
Seems like there are "blocked by cloudflare" number of questions per month.
Their blocking of everyone not using chrome/etc from accessing their website probably contributed quite a bit to the implied downturn I'm reading in other comments.
I imagine at least some of the leveling off could be due to question saturation. If duplicates are culled (earnestly or overzealously) then there will be a point where most of the low hanging fruit is picked.
When you see AI giving you back various coding snippets almost verbatim from SO, it really makes you wonder what will happen in the future with AI when it can't depend on actual humans doing the work first.
I certainly use it less now that I get a CloudFlare check every time I go and sometimes it fails or loads forever.
I usually just go back to search results and look elsewhere after a second or two.
This is a great example of how free content was exploited by LLMs and used against oneself to an ultimate destruction.
Every content creator should be terrified of leaving their content out for free and I think it will bring on a new age of permanent paywalls and licensing agreements to Google and others, with particular ways of forcing page clicks to the original content creators.
Man after reading some of the comments and looking at the graph I have learned a lesson. I went to SO all the time to find answers to questions, but I never participated. I mean they made it hard, but given the amount of benefit I gained I should've overcome that friction. If I and people like me had, maybe we could have diluted the moderation drama that others talk about (and that I, as a greedy user, never saw). Now it's a crap-shoot with an LLM instead of being able to peruse great answers from different perspectives to common problems and building out my own solution.
I suspect a lot of the traffic shift is from Google replacing the top search result, which used to be Stack Overflow for programming questions, with a Gemini answer.
Someone needs to archive the entirety of StackOverflow and make it available over torrent so that it can be preserved when the site shuts down. Urgently.
> As of (and including) the 2025-06-30 data dump, Stack Exchange has started including watermarking/data poisoning in the data. At the time of writing, this does not appear to apply to the 2025-09-30 data dump. The format(s), the dates for affected data dumps, and by extension how the garbage data can be filtered out, are described in this community-compiled list: https://github.com/LunarWatcher/se-data-dump-transformer/blo.... If the 2025-09-30 data dump turns out to be poisoned as well, that's where an update will be added. For obvious reasons, the torrent cannot be updated once created.
The result is not surprising! Many people are now turning to LLMs with their questions instead. This explains the decline in the number of questions asked.
Signs of over-moderation and increasing toxicity on Stack Overflow became particularly evident around 2016, as reflected by the visible plateau in activity.
Many legitimate questions were closed as duplicates or marked off-topic despite being neither. Numerous high-quality answers were heavily edited to sound more "neutral", often diluting their practical value and original intent.
Some high-profile users (with reputation scores > 10,000) were reportedly incentivized by commercial employers to systematically target and downvote or flag answers that favored competing products. As a result, answers from genuine users that recommended commercial solutions based on personal experience were frequently removed altogether.
Additionally, the platform suffers from a lack of centralized authentication: each Stack Exchange subdomain still operates with its own isolated login system, which creates unnecessary friction and discourages broader user participation.
TFMs are not a thing anymore. Most of them are merely collections of sparse random dots one might join by sheer luck only, granted no other knowledge of the system being attempted to document.
Are you sure? You can post questions even with a completely new blank account. It's comments that require some reputation, maybe you were thinking about those?
And still last month one of my questions on SO got closed because it was - "too broad".
I mean it was 2025 and how many very precise software engineering questions are there that any flagship models couldn't answer in seconds?
Although I had moderate popularity on SO I'm not gonna miss it; that community had always been too harsh for newcomers. They had the tiniest power, and couldn't handle that well.
Good riddance to bad rubbish (TLDR: Questions are now almost never being asked on Stack overflow).
The most annoying example I can think of (but can’t link to, alas) is when I Googled for an answer to a technical question, and got an annoying Stack Overflow answer which didn’t answer the question, telling the person to just Google the answer.
Not surprising. It's very often a toxic, unhelpful, stubborn community. I think maybe once or twice in years of use did I ever find it genuinely welcoming and helpful. Frequently instead I thought "Why should I even bother to post this? It'll just get either downvoted, deleted, or ignored."
I fairly recently tried to ask a question on SO because the LLMs did not work for that domain. I’m no beginner to SO, having some 13k points from many questions and answers. I made, in my opinion, a good question, referenced my previous attempts, clearly stating my problem and what I tried to do. Almost immediately after posting I got downvoted, no comments, a close- suggestions etc. A similar thing happened the last two times I tried this too. I’m not sure what is going on over there now, but whatever that site was many years ago, it isn’t any more. It’s s shame, because it was such a great thing, but now I am disincentivized to use it because I lose points each time I tip my toes back in.
LLMs are dogshit in many ways but when it comes to programming they are faster than people, respond instantaneously to further information, and can iterate until they understand the problem fully.
Bonus is that you don’t get some dipshit being snarky.
They would just use documentation. I know there is some synthesis they would lose in the training process but I’m often sending Claude through the context7 MCP to learn documentation for packages that didn’t exist, and it nearly always solves the problem for me.
The brilliance of StackOverflow was in being the place to find out how to do tricky workarounds for functionality that either wasn't documented or was buggy such that workarounds were needed to make it actually work.
Software quality is now generally a bit better than it was in 2010, but that need is ultimately still there.
StackOverflow was immediately dead for me the day they declared that AI sellout of theirs.
Pathetic thieves, they won't even allow deleting my own answers after that. Not that it would make the models unlearn my data, of course, but I wanted to do so out of principle.
This is incredible.
Anyone who claims LLMs aren't useful will need to explain how come almost every programmer can solve 95% of his problems with an LLM without needing anything else. This is real usefulness right here.
EDIT: I'm not saying I'm loving what happened and what is becoming of our roles and careers, I'm just saying things have changed forever; there's still a (shrinking) minority of people who seem to not be convinced.
Maybe we had too many programmers who weren’t capable of actually solving their own problems. Maybe only one in twenty programmers were ever actually any good at their jobs.
For those who have historically wondered about or objected to "moderation" (people usually mean curation here; as the overwhelming majority of the actions they're talking about are not performed by moderators) on Stack Overflow, here's a hand-picked list of important discussions from the meta site explaining some policy basics:
Note that IDs are in chronological order. The rate of new meta.stackoverflow.com posts fell off dramatically at some point because of the formation of a network-wide meta.stackexchange.com. The earliest entries listed here are from 2014.
Some of the comments in these links are hilariously elitist. They are actively embracing a hostile environment, especially towards newcomers, but how do they expect to grow and maintain a community when they are scaring users away?
Stackoverflow bureaucracy and rule mongering are insane. I recommend participation just to behold the natives in their biom. Its like a small european union laser focused on making asking snd answering a question the largest pain point of a site that is mainly about asking and answering questions.
Since the trend must go on, we expect StackExchange to now offer answers, and the user responses need to be questions. We could even make a quiz game show out of that! /s
Some comments:
- This is a really remarkable graph. I just didn't realize how thoroughly it was over for SO. It stuns me as much as when Encyclopædia Britannica stopped selling print versions a mere 9 years after the publication of Wikipedia, but at an even faster timescale.
- I disagree with most comments that the brusque moderation is the cause of SO's problems, though it certainly didn't help. SO has had poor moderation from the beginning. The fundamental value proposition of SO is getting an answer to a question; if you can the same answer faster, you don't need SO. I suspect that the gradual decline, beginning around 2016, is due to growth in a number of other sources of answers. Reddit is kind of a dark horse here, as I began seeing answers on Google to more modern technical questions link to a Reddit thread frequently along with SO from 2016 onwards. I also suspect Discord played a part, though this is harder to gauge; I certainly got a number of answers to questions for, e.g., Bun, by asking around in the Bun Discord, etc. The final nail in the coffin is of course LLMs, which can offer a SO-level answer to a decent percentage of questions instantly. (The fact that the LLM doesn't insult you is just the cherry on top.)
- I know I'm beating a dead horse here, but what happens now? Despite stratification I mentioned above, SO was by far the leading source of high quality answers to technical questions. What do LLMs train off of now? I wonder if, 10 years from now, LLMs will still be answering questions that were answered in the halcyon 2014-2020 days of SO better than anything that came after? Or will we find new, better ways to find answers to technical questions?
> I disagree with most comments that the brusque moderation is the cause of SO's problems, though it certainly didn't help. SO has had poor moderation from the beginning.
I was an early SO user and I don’t agree with this.
The moderation was always there, but from my perspective it wasn’t until the site really pushed into branching out and expanding Stack Exchange across many topics to become a Quora style competitor that the moderation started taking on a life of its own. Stack Overflow moderator drama felt constant in the later 2010s with endless weird drama spilling across Twitter, Reddit, and the moderator’s personal blogs. That’s about the same time period where it felt like the moderation team was more interested in finding reasons to exercise their moderation power than in maintaining an interesting website.
Since about 2020 every time I click a Stack Overflow link I estimate there’s a 50/50 chance that the question I clicked on would be marked as off topic or closed or something before anyone could answer it. Between the moderator drama and the constant bait-and-switch feeling of clicking on SO links that didn’t go anywhere the site just felt more exhausting than helpful.
There was definitely a bit of Pournelle's Iron Law of Bureaucracy [0] at work. I worked there during a lot of the drama you allude to, and... It sucked, for everyone. But also...
For most of SO's history, the vast majority of visitors (and this questions, answers) came in via Google. Not "search engines"; Google. This was pretty much baked in right at the start, and it effectively served as the site's primary user interface for years. And it worked. It worked pretty well! Until it didn't.
At some point, Google started surfacing fewer "tried and true" Q&A examples and more unanswered, poorly-answered or moderated examples. This broke the fundamental assumption that sat behind SO's moderation - that curating a smaller set of posts was preferable to encouraging more, and newer. Suddenly, Google wasn't a very good UI for SO anymore.
...and SO didn't really have a fallback. Heck, for a while during this period they actually stopped showing questions on their homepage unless you were already logged in; the core assumption remained that there was a never-ending spring of new people feeding the thing. LLMs weren't the start of the problem, they were the end - the final wake-up call.
I don't know that a site like SO can exist without the old Google, the old Internet; it is a product of all that, in the same way that mass-market TV shows were a product of 20th-century broadcast technology, or trade paperbacks of a particular intersection of printing tech and reading habits.
[0]: https://www.jerrypournelle.com/reports/jerryp/iron.html
23 replies →
I know the feeling of being happy not being the only one with that same problem (and that somebody bothered to actually ask on SO) and the crushing feeling that the question was closed as off topic (so no reason for me to ask) or marked as duplicate (referencing that is clearly not a duplicate and just showing that the mod took no effort to understand the question)
The moderation definitely got kind of nasty in the last 5 years or so. To the point where you would feel unwelcome for asking a question you had already researched, and felt was perfectly sound to ask. However, that didn't stop millions of people from asking questions every day, it just felt kinda shitty to those of us who spent more time answering, when we actually needed to ask one on a topic we were lacking in. (Speaking as someone who never moderated).
My feeling was always that the super mods were people who had too much time on their hands... and the site would've been better without them (speaking in the past tense, now). But I don't think that's what killed it. LLMs scraping all its content and recycling it into bite-sized Gemini or GPT answers - that's what killed it.
2 replies →
I asked a question for the first time mid last year. It was a question about "default" sizes in HTML layout calculations, with lots of research and links to relevant parts of the spec.
It was immediately closed as off topic, and there were a bunch of extremely vitriolic comments offended that I'd ask such a question on SO. It was briefly reopened weeks (?) later and then I guess closed again and now is deleted, so you can't even view the question any more.
I'd long heard of abusive moderation but... experiencing it first hand is something else. Anecdote of one, but I know I'm never going to ask there again.
In case anyone's wondering, I ended up asking on the WhatWG or W3C or something github project (via an issue?). The TLDR was rather eye opening, that basically the spec only codifies points of contention for browsers and old behaviors are generally undocumented. With some pointers I figured out the default size behavior through code diving, and it was complex (as in, hard to use) and very unintuitive.
41 replies →
Friend in my group was in the public beta back in '08. We all ended up signing up by the end of '09. I used it off-and-on over the years (have some questions and replies with hundreds of upvotes). Though SO had a rap for having what might seem like harsh replies or moderation, it was often imho just blunt/curt, to the point, and often objectively defensible. I also agree with your timeframe that, in the later 2010s, the site became infected with drama, and moderation suddenly started reaching its tendrils into non-technical areas, when it should not have. And on an ostensibly technical site, no less!
I found myself contributing less and less (same with Wikipedia), because I merely wanted to continue honing my craft through learning and contributing technical data with others who shared this same passion... I did not want to have politics shoved in my face, or have every post of mine have to be filtered through an increasingly extreme ideology which had nothing to do with the technical nature of the site. When I had my SO suspended with no warning or recourse for writing "master" in a reply, I knew it was time to leave for good. Most of the admins on the site transformed from technical (yet sometimes brash!) geeks, into political flag-waving and ideology-pushing avatars (including pushing their sexual agendas front and center), and not of the FSF/FLOSS kind, either.
These types of dramas have infected nearly everything online, especially since 2020. Even Linus has lost his mind with pushing politics into what should be purely technical areas https://news.ycombinator.com/item?id=41936049
LLMs were a final blow for many reasons, though I think that a huge part of it is that LLMs won't chide you and suspend/ban you for wanting to stick to strictly technical matters. I don't have to pledge allegiance to a particular ideology and pass a purity test before asking technical questions to an LLM.
1 reply →
Quite frankly you are wrong. Jeff and Joel spoke about their goals for very harsh moderation in their podcast while they were still building SO. The moderation from the very beginning was a direct result of the culture they created and it was completely intentional.
1 reply →
> The fundamental value proposition of SO is getting an answer to a question
I read an interview once with one of the founders of SO. They said the main value stackoverflow provided wasn't to the person who asked the question. It was for the person who googled it later and found the answer. This is why all the moderation pushes toward deleting duplicates of questions, and having a single accepted answer. They were primarily trying to make google searches more effective for the broader internet. Not provide a service for the question-asker or answerer.
Sad now though, since LLMs have eaten this pie.
> This is why all the moderation pushes toward deleting duplicates of questions, and having a single accepted answer.
My personal single biggest source of frustration with SO has been outdated answers that are locking out more modern and correct answers. There are so many things for which there is no permanently right answer over time. It feels like SO started solidifying and failed to do the moderation cleaning and maintenance needed to keep it current and thriving. The over-moderation you described helps people for a short time but then doesn’t help the person who googles much later. I’ve also constantly wished that bad answers would get hidden or cleaned out, and that accepted answers that weren’t very good would get more actively changed to better ones that showed up, it’s pretty common to see newer+better answers than the accepted one.
11 replies →
> This is why all the moderation pushes toward deleting duplicates of questions, and having a single accepted answer.
Having duplicates of the question is precisely why people use LLMs instead of StackOverflow. The majority of all users lack the vocabulary to properly articulate their problems using the jargon of mathematicians and programmers. Prior to LLMs, my use case for StackOverflow was something like this:
30 minutes trying (and failing) to use the right search terms to articulate the problem (remember, there was no contextual understanding, so if you used a word with two meanings and one of those meanings was more popular, you’d have to omit it using the exclusion operator).
30 minutes reading through the threads I found (half of which will have been closed or answered by users who ignored some condition presented by the OP).
5 minutes on implementation.
2 minutes pounding my head on my desk because it shouldn’t have been that hard.
With an LLM, if the problem has been documented at any point in the last 20 years, I can probably solve it using my initial prompt even as a layman. When you’d actually find an answer on StackOverflow, it was often only because you finally found a different way of phrasing your search so that a relevant result came up. Half the time the OP would describe the exact problem you were having only for the thread to be closed by moderators as a duplicate of another question that lacked one of your conditions.
40 replies →
I think that's a great policy. I don't think anyone wants duplicate questions. The problem is moderation marking unrelated questions as duplicates.
I'd say 9/10 times I find a direct match for my question on SO it's been closed as offtopic with links to one or more questions that are only superficially similar.
There are other problems that they don't even try to address. If 10 people ask the same question, why does only the first person to ask it get to choose the answer? Then lots of "XY" questions where the original asker didn't actually have problem X so selects an answer for Y, leaving the original X unsolved, and now all the duplicates only have an answer for Y too.
3 replies →
Sad? No. A good LLM is vastly better than SO ever was. An LLM won't close your question for being off-topic in the opinion of some people but not others. It won't flame you for failing to phrase your question optimally, or argue about exactly which site it should have been posted on. It won't "close as duplicate" because a vaguely-similar question was asked 10 years ago in a completely-different context (and never really got a great answer back then).
Moreover, the LLM has access to all instances of similar problems, while a human can only read one SO page at a time.
The question of what will replace SO in future models, though, is a valid one. People don't realize what a massive advantage Google has over everyone else in that regard. So many site owners go out of their way to try to block OpenAI's crawlers, while simultaneously trying to attract Google's.
3 replies →
Quite often, when my search returned a 'closed as duplicate' reply, I found the allegedly duplicate question did not accurately describe my problem, and the answers to it were often inferior, for my purposes, than those which had been given to my original question before the gate was closed.
3 replies →
The disconnect here is that they built it this way, but still call it a question and answer site and give a lot of power over to the person who created the question. They get to mark an answer as the solution for themselves, even if the people coming from Google have another answer as the solution.
If they were to recreate the site and frame it as a symptom and issue site, which is what the interview described, that would yield many different choices on how to navigate the site, and it would do a lot better. In particular, what happens when two different issues have the same symptom. Right now, that question is closed as a duplicate. Under a symptom and issue site, it's obvious that both should stay as distinct issues.
> They were primarily trying to make google searches more effective for the broader internet
This is mostly how I engaged with SO for a long, long time. I think it’s a testament to SO’s curation of answers that I didn’t ask almost any questions for like 5+ years after starting programming
LLMs also search Google for answers. Hence the knowledge may be not lost even for those who only supervises machines that write code.
If this were true, then treating any question as an X-Y problem shouldn't be allowed at all. I.e. answers should at least address the question as posed before/instead of proposing an alternative approach.
In reality the opposite is encouraged. For countless times, I've landed on questions with promising titles/search extracts, only to find irrelevant answers because people grabbed onto some detail in the question irrelevant to my case and provided X-Y answers.
This often also causes subsequent useful questions to be marked as dups even though they no longer contain that irrelevant detail. The appeal process is so unfriendly that most would not bother.
See also: https://news.ycombinator.com/item?id=36068243
I agree with that and I think it was the right decision. There was grousing about overmoderation but I think a lot of people got unreasonably annoyed when their question was closed. And the result was a pretty well-curated and really useful knowledge base.
> Sad now though, since LLMs have eaten this pie.
By regenerating an answer on command and never caring about the redundancy, yeah.
The DRY advocate within me weeps.
Thinking from first principles, a large part of the content on stack overflow comes from the practical experience and battle scars worn by developers sharing them with others and cross-curating approaches.
Privacy concerns notwithstanding, one could argue having LLMs with us every step of the way - coding agents, debugging, devops tools etc. It will be this shared interlocutor with vast swaths of experiential knowledge collected and redistributed at an even larger scale than SO and forum-style platforms allow for.
It does remove the human touch so it's quite a different dynamic and the amount of data to collect is staggering and challenging from a legal point of view, but I suspect a lot of the knowledge used to train LLMs in the next ten years will come from large-scale telemetry and millions of hours in RL self-play where LLMs learn to scale and debug code from fizzbuzz to facebook and twitter-like distributed system.
> Privacy concerns notwithstanding, one could argue having LLMs with us every step of the way - coding agents, debugging, devops tools etc.
That might work until an LLM encounters a question it's programmed to regard as suspicious for whatever reason. I recently wanted to exercise an SMTP server I've been configuring, and wanted to do it by an expect script, which I don't do regularly. Instead of digging through the docs, I asked Google's Gemini (whatever's the current free version) to write a bare bones script for an SMTP conversation.
It flatly refused.
The explanation was along the lines "it could be used for spamming, so I can't do that, Dave." I understand the motivation, and can even sympathize a bit, but what are the options for someone who has a legitimate need for an answer? I know how to get one by other means; what's the end game when it's LLMs all the way down? I certainly don't wish to live in such a world.
1 reply →
I don't know how others use LLMs, but once I find the answer to something I'm stuck on I do not tell the LLM that it's fixed. This was a problem in forums as well but I think even fewer people are going to give that feedback to a chatbot
2 replies →
Am I the only one that sees this as a hellscape?
No longer interacting with your peers but an LLM instead? The knowledge centralized via telemetry and spying on every user’s every interaction and only available thru a enshitified subscription to a model that’s been trained on this stolen data?
38 replies →
As long as software is properly documented, and documentation is published in LLM-friendly formats, LLMs may be able to answer most of the beyond basic questions even when docs don't explicitly cover a particular scenario.
Take an API for searching products, one for getting product details, and then an API for deleting a product.
The documentation does not need to cover the detailed scenario of "How to delete a product" where the first step is to search, the second step is to get the details (get the ID), and the third step is to delete.
The LLM is capable of answering the question "how to delete the product 'product name'".
To some degree, many of the questions on SO were beyond basic, but still possible for a human to answer if only they read documentation. LLMs just happen to be capable of reading A LOT of documentation a LOT faster, and then coming up with an answer A LOT faster.
If the LLM is also writing the documentation, because the developers surely don’t want to, I’m not sure how well this will work out.
I have some co-workers who have tried to use Copilot for their documentation (because they never write any and I’m constantly asking them questions as a result), and the results were so bad they actually spent the time to write proper documentation. It failed successfully, I suppose.
1 reply →
"In this imaginary world where everything is perfect and made to be consumed by LLMs, LLMs are the best tool for the job".
2 replies →
> I disagree with most comments that the brusque moderation is the cause of SO's problems
The moderation was precisely the reason I stopped using stackoverflow and started looking for answers and asking questions elsewhere. It was nearly impossible to ask anything without someone replying "Why would you even want to do that, do <something completely different that does not solve my problem> instead!". Or someone claiming it's a duplicate and you should use that ancient answer from another question that 1) barely fits and doesnt solve my problem and 2) is so outdated, it's no longer useful.
Whenever I had to ask something, I had to add a justification as to why I have to do it that way and why previous posts do not solve the issue, and that took more space than the question itself.
I certainly won't miss SO.
That's not moderation. That's a lack of moderation. Comments like this weren't posted by mods but by people looking for stuff to answer. A mod would definitely delete such a comment if they were notified about it. But not enough users flagged stuff and preferred to complain about being hurt instead. There's also a problem that not enough mods were present at all times to handle the volume of comments posted.
I will miss it but you are right about moderation. I don't know what the issue is on some platforms, reddit and SO come to mind. Moderators on many other platforms or forums seem to be alright and keep a clear head, even when they have to deal with a lot of vitriol and they get little thanks for their work.
There are probably negative examples as well but some platforms seem to be especially vulnerable. If I had to run reddit or SO, I would limit moderation to one subreddit/subdomain. No idea if that would help, but the problem isn't exactly invisible.
If we're going to diagnose pre-AI Stack Overflow problems I see two obvious ones:
1. The attempt to cut back on the harshness of moderation meant letting through more low-quality questions.
2. More importantly, a lot of the content is just stale. Like you go to some question and the accepted answer with the most votes is for a ten-year-old version of the technology.
> Like you go to some question and the accepted answer with the most votes is for a ten-year-old version of the technology.
This is still a problem with LLMs as a result. The bigger problem is that now the LLM doesn’t show you it was a 10 year old solution, you have to try it, watch it fail, then find out it’s old, and ask for a more up to date example, then watch it flounder around. I’ve experienced this more times than I can count.
6 replies →
> The fundamental value proposition of SO is getting an answer to a question
But the horrible moderation was in part a reason why many SO questions had no answers.
I am not saying poor moderation caused all of this, but it contributed negatively and many people were pissed at that and stopped using SO. It is not the only reason SO declined, but there are many reasons for SO failure after its peak days.
To the extent that moderation ever prevented questions from getting answers, that was by closing them.
When a question gets closed before an answer comes in, the OP has nine days to fix it before it gets deleted automatically by the system.
The value proposition is getting an answer to a question that is useful to a reasonably broad audience. That very often means a question that someone else asked, the answer to which is useful to you. It is not getting an "answer" to a "question" where an individual dumps some code trying to figure out what's wrong.
5 replies →
There was, obviously, only one main reason: LLMs. Anything else makes no sense. Even if the moderation was "horrible" (which sounds to me like a horrible exaggeration), there was nothing which came close to being as good as SO. There was no replacement. People will use the best available platform, even if you insist in describing it as "horrible". It's was not horrible compared to the alternatives, web forums like Reddit and HN, which are poorly optimized for answering questions.
4 replies →
That "Dead Internet" phrase keeps becoming more likely, and this graph shows that. Human-to-human interactions, LLMs using those interactions, less human-to-human interactions because of that, LLMs using... ?
> I disagree with most comments that the brusque moderation is the cause of SO's problems, though it certainly didn't help.
By the time my generation was ready to start using SO, the gatekeeping was so severe that we never began asking questions. Look at the graph. The number of questions was in decline before 2020. It was already doomed because it lost the plot and killed any valuable culture. LLMs were a welcome replacement for something that was not fun to use. LLMs are an unwelcome replacement for many other things that are a joy to engage with.
> I disagree with most comments that the brusque moderation is the cause of SO's problems, though it certainly didn't help. SO has had poor moderation from the beginning.
Overwhelmingly, people consider the moderation poor because they expect to be able to come to the site and ask things that are well outside of the site's mission. (It's also common to attribute community actions to "moderators" who in reality have historically done hardly any of it; the site simply didn't scale like that. There have been tens of millions of questions, versus a couple dozen moderators.)
The kinds of questions that people are getting quick, accurate answers for from an LLM are, overwhelmingly, the sort of thing that SO never wanted. Generally because they are specific to the person asking: either that person's issue won't be relevant to other people, or the work hasn't been done to make it recognizable by others.
And then of course you have the duplicates. You would not believe the logic some people put forward to insist that their questions are not duplicate; that they wouldn't be able, in other words, to get a suitable answer (note: the purpose is to answer a question, not solve a problem) from the existing Q&A. It is as though people think they are being insulted when they are immediately given a link to where they can get the necessary answer, by volunteers.
I agree that Reddit played a big role in this. But not just by answering questions; by forming a place where people who objected to the SO content model could congregate.
Insulting other users is and always has been against Stack Overflow Code of Conduct. The large majority of insults, in my experience, come from new users who are upset at being politely asked to follow procedures or told that they aren't actually allowed to use the site the way they're trying to. There have been many duplicate threads on the meta site about why community members (with enough reputation) are permitted to cast close votes on questions without commenting on what is wrong. The consensus: close reasons are usually fairly obvious; there is an established process for people to come to the meta site to ask for more detailed reasoning; and comments aren't anonymous, so it makes oneself a target.
It seems you deny each problem that everyone sees in SO. The fact is SO repulsed people, so there is a gap between your interpretation and reality.
> It is as though people think they are being insulted when they are immediately given a link to where they can get the necessary answer, by volunteers.
This, for example. Question can be marked as duplicate without an answer. In this case yes, it feels insulting because the other is asked in such a weird way, that no-one will find the old when they search for the new (for example after a library change) and marking it as duplicate of an unanswered answer if a guarantee that the next SEO user won’t see it.
8 replies →
> It is as though people think they are being insulted when they are immediately given a link to where they can get the necessary answer, by volunteers.
Multiple times my questions closed as duplicates of question that was answering a different question.
Even when I explicitly linked that QA in my question and described how it differs from mine.
I don't think "good moderation or not" really touches what was happening with SO.
I joined SO early and it had a "gamified" interface that I actually found fun. Putting in effort and such I able to slowly gain karma.
The problem was as the site scaled, the competition to answer a given question became more and more intense and that made it miserable. I left at that point but I think a lot people stayed with dynamic that was extremely unhealthy. (and the quality of accepted questions declined also).
With all this, the moderation criteria didn't have to directly change, it just had to fail to deal with the effects that were happening.
Agreed. The reputation system was extremely ill considered and never revisited. You may be interested in https://meta.stackexchange.com/questions/387356 .
This doesn't mean that it's over for SO. It just means we'll probably trend towards more quality over quantity. Measuring SO's success by measuring number of questions asked is like measuring code quality by lines of code. Eventually SO would trend down simply by advancements of search technology helping users find existing answers rather than asking new ones. It just so happened that AI advanced made it even better (in terms of not having to need to ask redundant questions).
> I wonder if, 10 years from now, LLMs will still be answering questions that were answered in the halcyon 2014-2020 days of SO better than anything that came after?
I've wondered this too and I wonder if the existing corpus plus new GitHub/doc site scrapes will be enough to keep things current.
Widespread internet adoption created “eternal September”, widespread LLM deployment will create “eternal 2018”
[dead]
> I disagree with most comments that the brusque moderation is the cause of SO's problems
Just to add another personal data point: i started posting in on StackOverflow well before llms were a thing and moderation instantly turned ne off and i immediately stopped posting.
Moderators used to edit my posts and reword what i wrote, which is unacceptable. My posts were absolutely peaceful and not inflammatory.
Moderation was an incredible problem for stack overflow.
> Moderators used to edit my posts and reword what i wrote, which is unacceptable. My posts were absolutely peaceful and not inflammatory.
99.9% probability the people who made those edits a) were not moderators; b) were acting completely in accordance with established policy (please read: "Why do clear, accurate, appropriately detailed posts still get edited?" https://meta.stackexchange.com/questions/403176)
Why do you think you should be the one who gets to decide whether that's "acceptable"? The site existed before you came to it, and it has goals, purposes and cultural norms established beforehand. It's your responsibility, before using any site on the Internet that accepts user-generated content, to try to understand the site's and community's expectations for that content.
On Stack Overflow, the expectations are:
1. You license the content to the site and to the community, and everyone is allowed to edit it. (This is also explicitly laid out in the TOS.)
2. You are contributing to a collaborative effort to build a useful resource for the programming community: a catalog of questions whose answers can be useful to many people, not just to yourself.
3. Content is intended to be matter-of-fact and right to the point, and explicitly not conversational. You are emphatically not participating in a discussion forum.
6 replies →
"I suspect that the gradual decline, beginning around 2016, is due to growth in a number of other sources of answers."
I think at least one other reason is that a lot of the questions were already posted. There are only so many questions of interest, until a popular new technology comes along. And if you look at mathoverflow (which wouldnt have the constant shocks from new technologies) the trend is pretty stable...until right around 2022. And even since then, the dropoff isn't nearly so dramatic. https://data.stackexchange.com/mathoverflow/query/edit/19272...
>>what happens now?
I'll tell you what happens now: LLMs continue to regurgitate and iterate and hallucinate on the questions and answers they ingested from S.O. - 90% of which are incorrect. LLM output continues to poison itself as more and more websites spring up recycling outdated or incorrect answers, and no new answers are given since no one wants to waste the time to ask a human a question and wait for the response.
The overall intellectual capacity sinks to the point where everything collaboratively built falls apart.
The machines don't need AGI to take over, they just need to wait for us to disintegrate out of sheer laziness, sloth and self-righteous.... /okay.
there was always a needy component to Stack Overflow. "I have to pass an exam, what is the best way to write this algorithm?" and shit like that. A lazy component. But to be honest, it was the giving of information which forced you to think, and research, and answer correctly, which made systems like S.O. worthwhile, even if the questioners were lazy idiots sometimes. And now, the apocalypse. Babel. The total confusion of all language. No answer which can be trusted, no human in the loop, not even a smart AI, just a babbling set of LLMs repeating Stack Overflow answers from 10 years ago. That's the fucking future.
Things are gonna slide / in all directions / won't be nothin you can measure anymore. The blizzard of the world has crossed the threshold and it's overturned the order of the soul.[0]
[0] https://www.youtube.com/watch?v=8WlbQRoz3o4
Labs are spending billions on data set curation and RL from human experts to fill in the areas where they're currently weak. It's higher quality data than SO, the only issue is that it's not public.
3 replies →
The fundamental value proposition of SO is getting an answer to a question
For me, the value was writing answers on topics I was interested in…and internet points as feedback on their quality.
When SE abandoned their app, it broke my habit.
There's another significant forum: GitHub, the rise of which coincided with the start of SO's decline. I bet most niche questions went over to GH repos' issue/discussion forums, and SO was left with more general questions that bored contributors.
The moderation was a lot of the problem, but not the whole problem. Honestly these days a larger part is that old, low-quality answers are usually stuck as the top answer on old questions, despite the fact that the situation has changed massively in the past decade and there are newer answers further down that give the new answer. Or better yet: when the top answer is a decade old and says "that doesn't even make sense, why would you want to do that, it's impossible, but you can look at literally the entire Handbook to see what you can do" (with a link to the frontpage of the FreeBSD Handbook) and you have to scroll down nearly to the bottom to find the one answer that actually answers the question (how to add an on-link route on FreeBSD) (and that it's not actually impossible like the arrogant jerk on top claimed)...
Too bad stack overflow didn't high-quality-LLM itself early. I assume it had the computer-related brainpower.
with respect to the "moderation is the cause" thing... Although I also don't buy moderation as the cause, I wonder if any sort of friction from the "primary source of data" can cause acceleration.
for example, when I'm doing an interenet search for the definition of a word like buggywhip, some search results from the "primary source" show:
> buggy whip, n. meanings, etymology and more | Oxford English Dictionary
> Factsheet What does the noun buggy whip mean? There is one meaning in OED's entry for the noun buggy whip. See 'Meaning & use' for definition, usage, and quotation evidence.
which are non-answer to keep their traffic.
but the AI answer is... the answer.
If SO early on had had some clear AI answer + references, I think that would have kept people on their site.
The meta post describing the policy of banning AI-generated answers from the site (https://meta.stackoverflow.com/questions/421831) is the most popular of all time. Company interference with moderator attempts to enforce that policy lead to a moderator strike. The community is vehemently against the company's current repeated attempts to sneak AI into the system, which have repeatedly produced embarrassing results (see for example https://meta.stackoverflow.com/questions/425081 and https://meta.stackoverflow.com/questions/425162 ; https://meta.stackoverflow.com/questions/427807 ; https://meta.stackoverflow.com/questions/425766 etc.).
What you propose is a complete non-starter.
1 reply →
If my questions become more open ended and challenging as I become more senior, and I am not allowed to ask them on SO, why would I stay on the website?
So you get junior questions only and no seniors to answers.
The decline started in 2016, not 2026, they killed themselves
The newer questions that LLMs can't answer will be answered in forums - either SO, reddit, or elsewhere. There will be a much higher percentage of relevant content with far fewer new pages regurgitating questions about solved problems. So the LLMs will be able to keep up.
> - I know I'm beating a dead horse here, but what happens now? Despite stratification I mentioned above, SO was by far the leading source of high quality answers to technical questions. What do LLMs train off of now? I wonder if, 10 years from now, LLMs will still be answering questions that were answered in the halcyon 2014-2020 days of SO better than anything that came after? Or will we find new, better ways to find answers to technical questions?
To me this shows just how limited LLMs are. Hopefully more people realize that LLMs aren't as useful as they seem, and in 10 years they're relegated to sending spam and generating marketting websites.
Or we just stagnate, as tech no longer can afford to change.
I think the interesting thing here for those of us who use open source frameworks is that we can ask the LLM to look at the source to find the answer (eg. Pytorch or Phoenix in my case). For closed source libraries I do not know.
> What do LLMs train off of now? I wonder if, 10 years from now, LLMs will still be answering questions that were answered in the halcyon 2014-2020 days of SO better than anything that came after? Or will we find new, better ways to find answers to technical questions?
That's a great question. I have no idea how things will play out now - do models become generalized enough to handle "out of distrubition" problems or not ? If they don't then I suppose a few years from now we'll get an uptick in Stackoverflow questions; the website will still exist it's not going anywhere.
Specialized research AI agents are coming at which point we'll have numerous LLMs running and verifying experiments and creating a higher quality text corpus than the 2014-2020 halcyon, which is then used for other LLMs to be trained on.
It will be the reverse I suspect. Eventually we will see that LLM quality is lower when it is training data from 2014-2020 and will chalk it up to human limitations and the data not being written with a laser-focused goal of training better AI.
Instead of having chat-interfaces target single developers, moving towards multiplayer interfaces may bring back some of what has been lost--looping in experts or third-party knowledge when a problem is too though to tackle via agentic means.
Now all our interactions are neatly kept in personalised ledgers, bounded and isolated from one another. Whether by design or by technical infeasability, the issue remains that knowledge becomes increasingly bounded too instead of collaborative.
> SO was by far the leading source of high quality answers to technical questions
We will arrive on most answers by talking to an LLM. Many of us have an idea about we want. We relied on SO for some details/quirks/gotchas.
Example of a common SO question: how to do x in a library or language or platform? Maybe post on the Github for that lib. Or forums.. there are quirky systems like Salesforce or Workday which have robust forums. Where the forums are still much more effective than LLMs.
> The fundamental value proposition of SO is getting an answer to a question; if you can the same answer faster, you don't need SO.
Plus they might find the answer on SO without asking a new question - You probably would expect the # of new questions to peak or plateau even if the site wasn't dying, due to the accumulation of already-answered questions.
We'll get to the point where we can mass moderate core knowledge eventually. We may need to hand out extra weight for verified experts and some kind of most-votes-win type logic (perhaps even comments?), but live training data updates will be a massive evolution for language models.
> will we find new, better ways to find answers to technical questions?
I honestly don't think they need to. As we've seen so far, for most jobs in this world, answers that sound correct are good enough.
Is chasing more accuracy a good use of resources if your audience can't tell the difference anyway?
The LLMs will learn from our interactions with them. That's why they're often free
I stopped because of moderators. They literally killed the site for me.
Google also played a part. After a while, I noticed that for my programming related questions, almost no SO discussions showed up. When they did appear on the first page, they were usually abysmal and unusable for me.
When it started all kinds of very clever people were present and helped even with very deep and complex questions and problems. A few years later these people disappeared. The moderation was ok in the beginning, then they started wooing away a lot of talented people. And then the mods started acting like nazis, killing discussions, proper questions on a whim.
And then bots (?) or karma obsessed/farming people started to upvote batshit crazy, ridiculous answers, while the proper solution had like 5 upvotes and no green marker next to it.
It was already a cesspool before AI took over and they sold all their data. Initial purpose achieved.
> I disagree with most comments that the brusque moderation is the cause of SO's problems
Questions asked on SO that got downvoted by the heavy handed moderation would have been answered by LLMs without any of the flak whatsoever.
Those who had downvoted other's questions on SO for not being good enough, must be asking a lot of such not good enough questions to an LLM today.
Sure, the SO system worked, but it was user hostile and I'm glad we all don't have to deal with it anymore.
As an early user of SO [1], I feel reasonably qualified to discuss this issue. Note that I barely posted after 2011 or so so I can't really speak to the current state.
But what I can say is that even back in 2010 it was obvious to me that moderation was a problem, specifically a cultural problem. I'm really talking about the rise of the administrative/bureaucratic class that, if left unchecked, can become absolute poison.
I'm constantly reminded of the Leonard Nimoy voiced line from Civ4: "the bureaucracy is expanding to meet the needs of the expanding bureaucracy". That sums it up exactly. There is a certain type of person who doesn't become a creator of content but rather a moderator of content. These are people who end up as Reddit mods, for example.
Rules and standards are good up to a point but some people forget that those rules and standards serve a purpose and should never become a goal unto themselves. So if the moderators run wild, they'll start creating work for themselves and having debates about what's a repeated question, how questions and answers should be structured, etc.
This manifested as the war of "closed, non-constructive" on SO. Some really good questions were killed this way because the moderators decided on their own that a question had to have a provable answer to avoid flame wars. And this goes back to the rules and standards being a tool not a goal. My stance was (and is) that shouldn't we solve flame wars when they happen rather than going around and "solving" imaginary problems?
I lost that battle. You can argue taht questions like "should I use Javascript or Typescript?" don't belong on SO (as the moderators did). My position was that even though there's no definite answer, somebody can give you a list of strengths and weaknesses and things to consider.
Even something that does have a definite answer like "how do I efficiently code a factorial function?" has multiple but different defensible answers. Even in one language you can have multiple implementations that might, say, be compile-time or runtime.
Another commenter here talked about finding the nearest point on an ellipse and came up with a method they're proud of where there are other methods that would also do the job.
Anyway, I'd occasionally login and see a constant churn on my answers from moderators doing pointless busywork as this month they'd decided something needed to be capitalized or not capitalized.
A perfect example of this kind of thing is Bryan Henderson's war on "comprised of" on Wikipedia [2].
Anyway, I think the core issue of SO was that there was a lot of low-hanging fruit and I got a lot of accepted answers on questions that could never be asked today. You'll also read many anecdotes about people having a negative experience asking questions on SO in later years where their question was immediately closed as, say, a duplicate when the question wasn't a duplicate. The moderator just didn't understand the difference. That sort of thing.
But any mature site ultimately ends with an impossible barrier to entry as newcomers don't know all the cultural rules that have been put in place and they tend to have a negative experience as they get yelled at for not knowing that Rule 11.6.2.7 forbids the kind of question they asked.
[1]: https://stackoverflow.com/users/18393/cletus
[2]: https://www.npr.org/2015/03/12/392568604/dont-you-dare-use-c...
> This manifested as the war of "closed, non-constructive" on SO. Some really good questions were killed this way because the moderators decided on their own that a question had to have a provable answer to avoid flame wars.
It's literally a Q&A site. Questions need actual answers, not just opinions or "this worked for me".
> This manifested as the war of "closed, non-constructive" on SO. Some really good questions were killed this way because the moderators decided on their own that a question had to have a provable answer to avoid flame wars.
Please point at some of these "really good" questions, if you saved any links. (I have privileges to see deleted questions; deletion is normally soft unless there's a legal requirement or something.) I'll be happy to explain why they are not actually what the site wanted and not compatible with the site's goals.
The idea that the question "should have provable answers" wasn't some invention of moderators or the community; it came directly from Atwood (https://stackoverflow.blog/2011/01/17/real-questions-have-an...).
> I lost that battle. You can argue taht questions like "should I use Javascript or Typescript?" don't belong on SO (as the moderators did). My position was that even though there's no definite answer, somebody can give you a list of strengths and weaknesses and things to consider.
Please read "Understanding the standard for "opinion-based" questions" (https://meta.stackoverflow.com/questions/434806) and "What types of questions should I avoid asking?" (https://stackoverflow.com/help/dont-ask).
4 replies →
Dunno why you are being downvoted - there is a certain type of person who contributes virtually nothing on Wikipedia except peripheral things like categories. BrownHairedGirl was the most toxic person in Wikipedia but she was lauded by her minions - and yet she did virtually no content creation whatsoever. Yet made millions of edits!
Moderation got worse over time
> What do LLMs train off of now?
Perhaps they’ll rely on what was used by people who answered SO questions. So: official docs and maybe source code. Maybe even from experience too, i.e. from human feedback and human written code during agentic coding sessions.
> The fact that the LLM doesn't insult you is just the cherry on top.
Arguably it does insult even more, just by existing alone.
I spent the last 14 days chasing an issue with a Spark transform. Gemini and Claude were exceptionally good at giving me answers that looked perfectly reasonable: none of them worked, they were almost always completely off-road.
Eventually I tried with something else, and found a question on stackoverflow, luckily with an answer. That was the game changer and eventually I was able to find the right doc in the Spark (actually Iceberg) website that gave me the final fix.
This is to say that LLMs might be more friendly. But losing SO means that we're getting an idiot friendly guy with a lot of credible but wrong answers in place of a grumpy and possibly toxic guy which, however, actually answered our questions.
Not sure why someone is thinking this is a good thing.
What I always appreciate about SO is the dialogue between commenters. LLMs give one answer, or bullet points around a theme, or just dump a load of code in your IDE. SO gives a debate, in which the finer points of an issue are thrashed out, with the best answers (by and large) floating to the top.
SO, at its best, is numerous highly-experienced and intelligent humans trying to demonstrate how clever they are. A bit like HN, you learn from watching the back and forth. I don't think this is something that LLMs can ever replicate. They don't have the egos and they certainly don't have the experience.
Whatever people's gripes about the site, I learned a hell of a lot from it. I still find solutions there, and think a world without it would be worse.
The fundamental difference between asking on SO and asking an LLM is that SO is a public forum, and an LLM will be communicated with in private. This has a lot of implications, most of which surround the ability for people to review and correct bad information.
8 replies →
SO also isn't afraid to tell you that your question is stupid and you should do it a better way.
Some people take that as a personal attack, but it can be more helpful than a detailed response to the wrong question.
1 reply →
> What I always appreciate about SO is the dialogue between commenters.
Stack Overflow is explicitly not for "dialogue", recent experiments (which are generally not well received by the regulars on the meta site) notwithstanding. The purpose of the comments on questions is to help refine the question and ensure it meets standards, and in some cases serve other meta purposes like pointing at different-but-related questions to help future readers find what they're looking for. Comments are generally subject to deletion at any time and were originally designed to be visually minimal. They are not part of the core experience.
Of course, the new ownership is undoing all of that, because of engagement metrics and such.
2 replies →
> I don't think this is something that LLMs can ever replicate. They don't have the egos and they certainly don't have the experience
Interesting question - the result is just words so surely a LLM can simulate an ego. Feed it the Linux kernel mailing list?
Isn’t back and forth exactly what the new MoE thinking models attempt to simulate?
And if they don’t have the experience that is just a question of tokens?
2 replies →
I don't know if this is still the case but back in the day people would often redirect comments to some stackoverflow chat feature, the links to which would always return 404 not found errors.
This comment and the parent one make me realize that people who answer probably value the exchange between experts more than the answer.
Perhaps the antidote involves a drop of the poison.
Let an LLM answer first, then let humans collaborate to improve the answer.
Bonus: if you can safeguard it, the improved answer can be used to train a proprietary model.
1 reply →
You can ask an LLM to provide multiple approaches to solutions and explore the pros and cons of each, then you can drill down and elaborate on particular ones. It works very well.
There are so many "great" answers on StackOverflow. Giving the why and not just the answer.
It's flat wrong to suggest SO had the right answer all the time, and in fact in my experience for trickier work it was often wrong or missing entirely.
LLMs have a better hit rate with me.
The example wasn't even finding a right answer so I don't see where you got that..
Searching questions/answers on SO can surface correct paths on situations where the LLMs will keep giving you variants of a few wrong solutions, kind of like the toxic duplicate closers.. Ironically, if SO pruned the history to remove all failures to match its community standards then it would have the same problem.
6 replies →
It entirely depends on the language you were using. The quality of both questions and answers between e.g. Go and JavaScript is incredible. Even as a relative beginner in JS I could not believe the amount of garbage that I came across, something that rarely happened for Go.
No point in arguing with people who bring a snowball into Congress to disprove global warming.
[flagged]
4 replies →
Because what you’re describing is the exception. Almost always with LLM’s I get a better solution, or helpful pointer in the direction of a solution, and I get it much faster. I honestly don’t understand anyone could prefer Google/SO, and in fact that the numbers show that they don’t. You’re in an extreme minority.
> But losing SO means that we're getting an idiot friendly guy with a lot of credible but wrong answers in place of a grumpy and possibly toxic guy which, however, actually answered our questions.
Which by the way is incredibly ironic to read on the internet after like fifteen years of annoying people left and right about toxic this and toxic that.
Extreme example: Linus Torvalds used to be notoriously toxic.
Would you still defend your position if the “grumpy” guy answered in Linus’ style?
> Would you still defend your position if the “grumpy” guy answered in Linus’ style?
If they answered correctly, yes.
My point is that providing _actual knowledge_ is by itself so much more valuable compared to _simulated knowledge_, in particular when that simulated knowledge is hyper realistic and wrong.
Sadly, an accountable individual representing an organization is different from a community of semi-anonymous users with a bunch of bureaucracy that can't or doesn't care about every semis anonymous user
You still get the same thing though?
That grumpy guy is using an LLM and debugging with it. Solves the problem. AI provider fine tunes their model with this. You now have his input baked into it's response.
How you think these things work? It's either a human direct input it's remembering or a RL enviroment made by a human to solve the problem you are working on.
Nothing in it is "made up" it's just a resolution problem which will only get better over time.
How does that work if there's no new data for them to train on, only AI slurry?
I'm hoping increasing we'll see agents helping with this sort of issue. I would like an agent that would do things like pull the spark repo into the working area and consult the source code/cross reference against what you're trying to do.
Once technique I've used successfully is to do this 'manually' to ensure codex/Claude code can grep around the libraries I'm using
Q&A isn't going away. There's still GitHub Discussions.
Not a big surprise once LLMs came along: stack overflow developed some pretty unpleasant traits over time. Everything from legitimate questions being closed for no good reason (or being labeled a duplicate even though they often weren’t), out of date answers that never get updated as tech changes, to a generally toxic and condescending culture amongst the top answerers. For all their flaws, LLMs are so much better.
Agreed. I personally stopped contributing to StackOverflow before LLMs, because of the toxic moderation.
Now with LLMs, I can't remember the last time I visited StackOverflow.
People in this thread are missing another key component in the decline of StackOverflow - the more experienced you become, the less useful it is.
The harder the problem, the less engagement it gets. People who spend hours working on your issue are rewarded with a single upvote. Meanwhile, "how do I concat a string" gets dozens or hundreds of upvotes.
The incentive/reward structure punished experienced folks with challenging/novel questions.
Pair that with the toxic moderation and trigger-happy close-votes, you get a zombie community with little new useful content.
40 replies →
Same here. I just didn't want to expend energy racing trigger happy mods. It was so odd, to this day remember vividly how they cleanup their arguments once proven wrong on the closing vote. Literally minutes before it would the close threshold.
Gen 0: expertsexchange.com, later experts-exchange.com (1996)
Gen 1: stackoverflow.com (2008)
Gen 2: chatgpt.com (2022, sort of)
6 replies →
And you can't delete your post when you realize how awful it was years later! That anti-information sticks around for ages. Even worse when there are bad answers attached to it, too.
6 replies →
The dumbest part of SO is how the accepted answer would often be bad, and sometimes someone had posted a better answer after the fact, and if the all-powerful moderators had the power to update it, they sure never did. Likewise, there were often better insights in comments. Apparently if you have the right mod powers, you can just edit an answer (such as the accepted one) to make it correct, but that always struck me as a bizarre feature, to put words in other people’s mouths.
I think overall SO took the gamification, and the “internet points” idea, way too far. As a professional, I don’t care about Reddit Karma or the SO score or my HN karma. I just wanted answers that are correct, and a place to discuss anything that’s actually interesting.
I did value SO once as part of the tedious process of attempting to get some technical problem solved, as it was the best option we had, but I definitely haven’t been there since 2023. RIP.
1 reply →
The same is true for reddit imo, it became impossible to post anything to a subreddit way before LLMs
Seemed like for every other question, I received unsolicited advice telling me how I shouldn't be doing it this way, only for me to have to explain why I wanted to do it this way (with silence from them).
This is called the XY problem https://meta.stackexchange.com/a/66378 . You ask for X, I tell you that what you really want is Y, I bully you, and I become more convinced that you and people that ask for X want Y.
Oh I love that game! (At least I think it's a game)
You ask how to do X.
Member M asks why you want to do X.
Because you want to do Y.
Well!? why do you want to do Y??
Because Y is on T and you can't do K so you need a Z
Well! Well! Why do you even use Z?? Clearly J is the way it is now recommended!
Because Z doesn't work on a FIPS environment.
...
Can you help me?
...
I just spent 15 minutes explaining X, Y and Z. Do you have any help?
...(crickets)
13 replies →
To avoid going insane the mindset should be to produce something useful for future readers.
Long before LLMs. Setting aside peak-COVID as a weird aberration, question volume has been in decline since 2014 or maybe 2016.
Stack Overflow would still have a vibrant community if it weren't for the toxic community.
Imagine a non-toxic Stack Overflow replacement that operated as an LLM + Wiki (CC-licensed) with a community to curate it. That seems like the sublime optimal solution that combines both AI and expertise. Use LLMs to get public-facing answers, and the community can fix things up.
No over-moderation for "duplicates" or other SO heavy-handed moderation memes.
Someone could ask a question, an LLM could take a first stab at an answer. The author could correct it or ask further questions, and then the community could fill in when it goes off the rails or can't answer.
You would be able to see which questions were too long-tail or difficult for the AI to answer, and humans could jump in to patch things up. This could be gamified with points.
This would serve as fantastic LLM training material for local LLMs. The authors of the site could put in a clause saying that "training is allowed as long as you publish your weights + model".
Someone please build this.
Edit: Removed "LLMs did not kill Stack Overflow." first sentence as suggested. Perhaps that wasn't entirely accurate, and the rest of the argument stands better on its own legs.
15 replies →
It doesn't have anything to do with LLMs. It has to do with shifting one's focus from doing good things to making money. Joel did that, and SO failed because of it.
Joel promised the answering community he wouldn't sell SO out from under them, but then he did.
And so the toxicity at the top trickled down into the community.
Those with integrity left the community and only toxic, selfcentered people remained to destroy what was left in effort to salvage what little there was left for themselves.
Mods didn't dupe questions to help the community. They did it to keep their own answers at the top on the rankings.
How did Joel sell out? Curious as I’m not aware of any monetary changes. I watched Joel several times support completely brain dead policies in the meta discussions which really set the rules and tone. So my respect there is low.
2 replies →
I also wonder if GitHub Discussions was also a (minor) contributing factor to the decline. I recall myself using GitHub Discussions more and more when it came to repo specific issues.
The timeline also matches:
https://github.blog/changelog/2020-12-08-github-discussions-...
https://github.blog/news-insights/product-news/github-discus...
Do we have any stats for the number of GitHub discussions created each month to compare to this?
It seemed to me that pre-llm, google had stopped surfacing stackoverflow answers in search results.
My memory is there were a spate of SO scraping sites that google would surface above SO and google just would not zap.
It would have been super trivial to fix but google didn’t.
My pet theory was that google were getting doubleclick revenue from the scrapers so had incentives to let them scrape and to promote them in search results.
3 replies →
Because nobody was clicking on them
> legitimate questions being closed for no good reason
They are closed for good reasons. People just have their own ideas about what the reasons should be. Those reasons make sense according to others' ideas about what they'd like Stack Overflow to be, but they are completely wrong for the site's actual goals and purposes. The close reasons are well documented (https://meta.stackoverflow.com/questions/417476) and well considered, having been exhaustively discussed over many years.
> or being labeled a duplicate even though they often weren’t
I have seen so many people complain about this. It is vanishingly rare that I actually agree with them. In the large majority of cases it is comically obvious to me that the closure was correct. For example, there have been many complaints in the Python tag that were on the level of "why did you close my question as a duplicate of how to do X with a list? I clearly asked how to do it with a tuple!" (for values of X where you do it the same way.)
> a generally toxic and condescending culture amongst the top answerers.
On the contrary, the top answerers are the ones who will be happy to copy and paste answers to your question and ignore site policy, to the constant vexation of curators like myself trying to keep the site clean and useful (as a searchable resource) for everyone.
> For all their flaws, LLMs are so much better.
I actually completely agree that people who prefer to ask LLMs should ask LLMs. The experience of directly asking (an LLM) and getting personalized help is explicitly the exact thing that Stack Overflow was created to get away from (i.e., the traditional discussion forum experience, where experts eventually get tired of seeing the same common issues all the time and all the same failures to describe a problem clearly, and where third parties struggle to find a useful answer in the middle of along discussion).
You seem to have filled this thread with a huge number of posts that try to justify SO's actions. Over and over, these justifications are along the lines of "this is our mission", "read our policy", "understand us".
Often, doing what your users want leads to success. Stamping authority over your users, and giving out a constant air of "we know better than all of you", drives them away. And when it's continually emphasized publicly (rather than just inside a marketing department) that the "mission" and the "policy" are infinitely more important than what your users are asking for, that's a pretty quick route to failure.
When you're completely embedded in a culture, you don't have the ability to see it through the eyes of the majority on the outside. I would suggest that some of your replies here - trying to deny the toxicity and condescension - are clearly showing this.
1 reply →
> "why did you close my question as a duplicate of how to do X with a list? I clearly asked how to do it with a tuple!" (for values of X where you do it the same way.)
This is a great example of a question that should not be closed as a duplicate. Lists are not tuples in Python, regardless of how similar potential answers may be.
1 reply →
Oh yeah.
My favorite feature of LLMs, is the only dumb question, is the one I don't ask.
I guess someone could train an LLM to be spiteful and nasty, but that would only be for entertainment.
If you say the wrong thing to grok, it will go off on you. It's quite entertaining!
I suppose all sites that have a voting component run the risk of becoming unpleasant.
Hacker News, and we who frequent it, ought to have that in mind.
dang and the other HN moderators do a heroic job to set the tone, which has second- and third-order effects on behavior.
1 reply →
I think it has more to do with the fact that when you offer zero salary for moderators, you have to take what you can get, and it ain't good. I don't really see a connection to the voting mechanic.
6 replies →
The irony is that the LLMs are trained on stack overflow and should inherit a lot of those traits and errors.
Yeah, but they don't inherit their rules and attitude.
Really, if we could apply some RLHF to the Stack Overflow community, it would be doing a lot better.
>For all their flaws, LLMs are so much better
But LLMs get their answers from StackOverflow and similar places being used as the source material. As those start getting outdated because of lack of activity, LLMs won't have the source material to answer questions properly.
I regularly use Claude and friends where I ask it to use the web to look at specific GitHub repos or documentation to ask about current versions of things. The “LLMs just get their info from stack overflow” trope from the GPT-3 days is long dead - they’re pretty good at getting info that is very up to date by using tools to access the web. In some cases I just upload bits and pieces from a library along with my question if it’s particularly obscure or something home grown, and they do quite well with that too. Yes, they do get it wrong sometimes - just like stack overflow did too.
4 replies →
StackOverflow answers are outdated. Every time I end up on that site these days, I find myself reading answers from 12 years ago that are no longer relevant.
2 replies →
Now they can read the documentation and code in the repo directly and answer based on that.
1 reply →
I think the industry is quickly moving to syntheticly derived knowledge, or custom/systematic knowledge production from humans.
That depends on what you mean by "came along". If you mean "once that everyone got around to the idea that LLMs were going to be good at this thing" then sure, but it was not long ago that the majority of people around here were very skeptical of the idea that LLMs would ever be any good at coding.
What you're arguing about is the field completely changing over 3 years; it's nothing, as a time for everyone to change their minds.
LLMs were not productified in a meaningful way before ChatGPT in 2022 (companies had sufficiently strong LLMs, but RLHF didn't exist to make them "PR-safe"). Then we basically just had to wait for LLM companies to copy Perplexity and add search engines everywhere (RAG already existed, but I guess it was not realistic to RAG the whole internet), and they became useful enough to replace StackOverflow.
I dont think this is true. People were skeptical of agi / better than human coding which is not the case. As a matter of fact i think searching docs was one of the first manor uses of llms before code.
That's because there has been rapid improvement by LLMs.
Their tendency to bullshit is still an issue, but if one maintains a healthy skepticism and uses a bit of logic it can be managed. The problematic uses are where they are used without any real supervision.
Enabling human learning is a natural strength for LLMs and works fine since learning tends to be multifaceted and the information received tends to be put to a test as a part of the process.
How can we be sure that LLMs won't start giving stale answers?
We can't. I don't think the LLMs themselves can recognize when an answer is stale. They could if contradicting data was available, but their very existence suppresses the contradictory data.
1 reply →
They will, but model updates and competition help solve the problem. If people find that Claude consistently gives better/more relevant answers over GPT, for example, people will choose the better model.
The worst thing with Q/A sites isn't they don't work. It's that they there are no alternatives to stackoverflow. Some of the most upvoted answers on stackoverflow prove that it can work well in many cases, but too bad most other times it doesn't.
They still use the official documentation/examples, public Github Repos, and your own code which are all more likely to be evergreen. SO was definitely a massive training advantage before LLMs matured though.
LLMs are just statistics, eventually they kill themselves with feedback loop by consuming their own farts (literally)
all true, but i still find myself ask questions there after llm gave wrong answers and wasted my time
Yep, LLMs are perfect for the "quick buy annoying to answer 500 times" questions about writing a short script, or configuring something, or using the right combination of command line parameters.
Quicker than searching the entirety of Google results and none of the attitude.
Indeed. StackOverflow was by far the most unpleasant website that I have regularly interacted with. Sometimes, just seeing how users were treated there (even in Q&A threads that I wasn’t involved in at all) disturbed me so much it was actually interfering with my work. I’m so, so glad that I can now just ask an AI to get the same (or better) answers, without having to wade through the barely restrained hate on that site.
not only stackoverflow, but also reddit.com/r/aws reddit.com/r/docker reddit.com/r/postgresql all 3 of them have extremely toxic communities. ask a question and get downvoted instantly! Noo!! your job is to actually upvote the question to maximize exposure for the algorithm unless it is a really really stupid question that a google search could fix
> For all their flaws, LLMs are so much better.
For now. They still need to be enshitted.
Models are check-pointed. You can save one you like and use it forever.
7 replies →
Yeah just wait for the ads
This change was happening well before LLMs. People were tired of being yelled at and treated poorly.
A cautionary tale for many of these types of tech platforms, this one included.
They will no doubt blame this on AI, somehow (ChatGPT release: late 2022, decline start: mid 2020), instead of the toxicity of the community and the site's goals of being a knowledgebase instead of a QA site despite the design.
PS - This comment is closed as a [duplicate] of this comment: https://news.ycombinator.com/item?id=46482620
Right. I often end up on Stack Exchange when researching various engineering-related topics, and I'm always blown away by how incredibly toxic the threads are. We get small glimpses of that on HN, but it was absolutely out of control on Stack Exchange.
At the same time, I think there was another factor: at some point, the corpus of answered questions has grown to a point where you no longer needed to ask, because by default, Google would get you to the answer page. LLMs were just a cherry on top.
I agree there was some natural slow down as the corpus grew - the obvious questions were answered. But if the community was healthy, that should not have caused growth to stop. New technologies get created all the time, each starting with zero SO questions. (Or Google releases v2.0 which invalidates all answers written about v1.)
SO just stopped being fun for me. I wish more systems would use their point systems though.
2 replies →
> I'm always blown away by how incredibly toxic the threads are.
They are not "threads" and are not supposed to be "threads". Thinking about them as if they were, is what leads to the perception of toxicity.
It's funny that people blame the site for this.
That toxicity is just part of software engineering culture. It's everywhere.
1 reply →
[flagged]
13 replies →
People overestimate the impact of toxicity on number of monthly questions. The initial growth was due to missing answers. After some time there is a saturation point where all basic questions are already answered and can be found via Google. If you ask them again they are marked as dups.
That would be true if no new technologies were created every year (even more often).
2 replies →
The downward trend seems to start ~2017, and was interrupted by a spike during the early months of COVID-19. I'd be interested to know what drove that jump, perhaps people were less hesitant to post when they were working from home?
More people spent lot more time learning new tech skills (at every experience level).
The excess time available (less commute or career pause etc) and more interest (much more new opportunities) were probably leading reasons why they spent more time I would imagine.
2 replies →
A huge amount of people were just starting to learn programming, because they were stuck at home and had the time to pick something up.
If you look at the trends tag by tag, you can see that the languages, libraries, technologies etc. that appeal to beginners and recreational coders grew disproportionately.
Other tech support forums are terrible in other ways. AI is a godsend.
Typical response:
I am RJ, an Independent Advisor and Microsoft Gold Certified Support Specialist Enthusiast.
I know how your system is not functioning as desired! Rest assured, I am here to help you resolve this today.
Please follow these steps in order. Do not skip any steps.
Step 1: Reboot your computer Step 2: Reinstall windows Step 3: Contact Microsoft support
Did this resolve your issue? [ Yes ] [ No ]
If this helped, please mark this as the Answer and give me a 5-star rating so I can continue providing high-quality, scripted responses to other users!
Standard Disclaimer: I do not work for Microsoft. I am an independent volunteer who enjoys copying and pasting from a manual written in 2014.
> the site's goals of being a knowledgebase instead of a QA site despite the design.
A Q&A site is a knowledge base. That's just how the information is presented.
If you want a forum — a place where you ask the question to get answered one-on-one — you have countless options for that.
Stack Overflow pages have a different design from that explicitly to encourage building a knowledge base. That's why there's a question at the top and answers underneath it, and why there are not follow-up questions, "me too" posts, discussion of annoyances related to the question, tangential rants, generic socialization etc.
Jeff Atwood was quite clear about this from the beginning.
If you ignore the early pandemic bump, it even looks like the decline started in late 2017, though it's more variable than after the bump
I wonder what is the role of moderating duplicate questions. More time passes - more existing data there is and less need for new questions. If you moderate duplicate questions, will they disappear from these charts? Is this decline actually logical?
2020 there was new CEO and moderator council was formed: https://stackoverflow.blog/2020/01/21/scripting-the-future-o...
Many people are pointing out the toxicity, but the biggest thing that drove me away, especially for specific quantitative questions, was that SO was flat out wrong (and confidently so) on many issues.
It was bad enough that I got back in the habit of buying and building a library of serious reference books because they were the only reliable way to answer detailed technical questions.
If you do not mind my asking, what sorts of questions were you asking that were resulting in wrong answers?
There is an obvious acceleration of the downwards trend at the time ChatGPT got popular. AI is clearly a part of this, but not the only thing that affects SO activity.
Ironically they could probably do some really useful deduplication/normalization/search across questions and answers using AI/embeddings today, if only they’d actually allowed people to ask the same questions infinite different ways, and treated the result of that as a giant knowledge graph.
I was into StackOverflow in the early 2010s but ultimately stopped being an active contributor because of the stupid moderation.
I wonder if we can attribute some $billion of the investment in LLMs directly to the toxicity on StackOverflow.
It is sort of because of AI - it provided a way of escaping StackOverflow's toxicity!
Could view it as push/pull dynamics: pushed away by toxicity, pulled to good answers from AI.
Use of GPT3 among programmers started 2021 with GitHub Copilot which preceded ChatGPT.
I agree the toxic moderation (and tone-deaf ownership!) initiated the slower decline earlier that then turned into the LLM landslide.
Tbf SO also suffered from its own success as a knowledgebase where the easy pickings were long gone by then.
Toxic community is mostly a meme myth. I have like 30k points and whatever admins were doing was well deserved as 90% of the questions were utterly impossible to help with. Most people wanted free help and couldn't even bother to put in 5 minutes of work.
Actual analysts here that have looked at this graph like... a lot, so let me contextualize certain themes that tend to crop up from these:
- The reduction of questions over time is asymptomatic of SO. When you have a library of every question asked, at some point, you asked most of the easy questions. Have a novel question becomes hard. - This graph is using the Posts table, not PostsWithDeleted. So, it only tells you of the questions that survived at this point in time, this [0] is the actual graph which while describes a curve that shows the same behavior, it's more "accurate" of the actual post creation. - This is actually a Good Thing™. For years most of the questions went unanswered, non-voted, non-commented, just because there was too many questions happening all the time. So the general trend is not something that the SO community needs to do anything about. Almost 20% of every question asked is marked as duplicate. If people searched... better™ they wouldn't ask as many questions, and so everyone else had more bandwidth to deal with the rest. - There has been a shift in help desk style of request, where people starting to prefer discord and such to get answers. This is actually a bad thing because that means that the knowledge isn't public nor indexed by the world. So, information becomes harder to find, and you need to break it free from silos. - The site, or more accurately, the library will never die. All the information is published in complete archives that anyone can replicate and restart if the company goes under or goes evil. So, yeah, such concerns, while appreciated, are easily addressed. At worst, you would be losing a month or two of data.
[0]: https://data.stackexchange.com/stackoverflow/query/edit/1926...
> When you have a library of every question asked, at some point, you asked most of the easy questions. Have a novel question becomes hard
This would be true if programming were a static field, but given that new programming languages/frameworks/technologies/techniques/etc. are constantly coming out and evolving, that argument doesn't make sense.
Programming is not a static field in the answers side, but it's in the question side. "How to print characters on a terminal with python?" is the same problem today as it was 25 years ago. The answer changed but the problem remained. That's what people saying that programming isn't static is missing: the problem space grows significantly slower than the solution space.
1 reply →
OP here: I had the same thought, but noticed a very similar trend in both [0]; I think this graph is more interesting because you'd expect the number of new users to be growing [1], but this seems to have very little effect on deleted questions or even answers
[0]: https://data.stackexchange.com/stackoverflow/query/1927371#g...
[1]: https://data.stackexchange.com/stackoverflow/query/1927375#g...
The second graph here ([1]) is especially interesting because the total montly number of new users seems completely unrelated to number of posts, until you filter for a rep > 1 which has a close to identical trend
At my place of work we use an indexing service for discord that creates an index of searchable static pages for all discord interactions.
So while I agree the help desk style system isn’t really better it also doesn’t necessarily mean that it is lost forever in a silo.
Before you ask, we use https://www.linen.dev/ but I’m sure there are other similar solutions by now
> which while describes a curve that shows the same behavior, it's more "accurate" of the actual post creation.
I would say that this graph looks a lot more extreme, actually!
Your post formatting is making this very difficult to read
I am aware, I tried to do some kind of bullet point as I've seen other posts, but I don't understand how to "activate it"
1 reply →
What exactly do you mean by "asymptomatic"? The proper meaning of the word does not fit into what you wrote.
"Asymptomatic" means you have a cold but you show none of the symptoms, hence a-symptom-atic, no symptoms.
parent is meaning asymptote as in maths where a line is approaching to a value but never touching it. maybe it was a typo or auto correct...
https://en.wikipedia.org/wiki/Asymptote
I guess I'm the only one that was a fan of SO's moderation. I never got too deep into it (answered some TypeScript questions). But the intention to reduce duped questions made a lot of sense to me. I like the idea of a "living document" where energy is focused on updating and improving answers to old versions of the same question. As a user looking for answers it means I can worry less about finding some other variation of the same question that has a more useful answer
I understand some eggs got cracked along the way to making this omelette but overall I'd say about 90% of the time I clicked on a SO link I was rewarded with the answer I was looking for.
Just my two cents
The problem with duplicate questions is that they weren't duplicates at all, and mods weren't competent enough to tell a difference.
Show me one that was closed by a moderator. Just one. And I will tell you exactly what happened.
16 replies →
I also agreed with this vision. It was meant to be more like Wikipedia rather than Reddit.
Those saying that StackOverflow became toxic are absolutely correct. But we should not let that be it's legacy. It is IMO still today one of the greatest achievements in terms of open data on the internet. And it's impact on making programming accessible to a large audience cannot be understated.
[flagged]
what the fuck are you talking about? SO, remarkably, did not get enshittified.
5 replies →
Many users left because they had had overly strict moderation for posting your questions. I have 6k reputation, multiple gold badges and I will remember StackOverflow as a hostile place to ask a questions, honestly. There were multiple occasions when they actually prevented me from asking, and it was hard to understand what exactly went wrong. To my understanding, I asked totally legit questions, but their asking policy is so strict, it's super hard to follow.
So "I'm not happy he's dead, but I'm happy he's gone" [x]
I have around 2k points, not something to brag about, but probably more than most stackoverflow users. And I know what I am talking about given over a decade of experience in various tech stacks.
But it requires 3,000 points to be able to cast a vote to reopen a question, many of which incorrectly marked as duplicate.
I said to myself, let it die.
I was an early adopter. Have over 30k reputation because stack overflow and my internship started at the same time. I left because of the toxic culture, and that it's less useful the more advanced you get
> many of which incorrectly marked as duplicate.
Please feel free to cite examples. I'll be happy to explain why I think they're duplicates, assuming I do (in my experience, well over 90% of the time I see this complaint, it's quite clear to me that the question is in fact a duplicate).
But more importantly, use the meta site if you think something has been done poorly. It's there for a reason.
3 replies →
It's also was a bit frustrating for me to answer. There was time when I wanted to contribute, but questions that I could answer were very primitive and there were so many people eager to post their answer that it demotivated me and I quickly stopped doing that. Honestly there are too many users and most of them know enough to answer these questions. So participating as "answerer" wasn't fun for me.
Once StackOverflow profiles, brief as they were, became a metric they ceased to be worth a helluva lot. Back in the early 2010s I used to include a link to my profile. I had a low 5-figure score and I had more than one interviewer impressed with my questions and answers on the site. Then came point farmers.
I remember one infamous user who would farm points by running your questions against some grammar / formatting script. He would make sure to clean up an errant comma or a lingering space character at the end of your post to get credit for editing your question, thereby “contributing.”
To their early credit, I once ran for and nearly won a moderator slot. They sent a nice swag package to thank me for my contributions to the community.
1 reply →
I spent a lot of time answering rather primitive questions, but since it was on a narrow topic (Logstash, part of the ELK stack), there wasn't many other people eager to post answers. Though it often ended up with the same type of issues, not necessarily duplicates, but similar enough that I got bored with it.
> To my understanding, I asked totally legit questions, but their asking policy is so strict, it's super hard to follow.
I think https://meta.stackoverflow.com/questions/417476 is pretty straightforward. If you can show a question of yours that was closed, I'll be happy to try to explain why.
25k here, stopped posting cause you'd spend 10m on a reply to a question just to have the question closed on you by some mod trying to make everything neat.
Maybe it was a culture clash but I came from newsgroups where the issue was is that someone needed help. However SO had the idea that the person who needed help wasn't as important as the normalisation of the dataset.
I sometimes wonder how much time I could have saved for those whose questions got closed before I could answer them. But yeah, that and the swearing culture clash were issues I struggled with, and ultimately meant I stopped contributing.
> Maybe it was a culture clash but I came from newsgroups where the issue was is that someone needed help. However SO had the idea that the person who needed help wasn't as important as the normalisation of the dataset.
Yes, because doing things that way was explicitly part of the goal, from the beginning. There are countless other places where you can directly respond to people who need help (and if you like doing that, you should stick to those places). Doing things that way has negative consequences in terms of making something that's useful for on-lookers, and causing a lot of experts to burn out or get frustrated. This is stuff that Jeff Atwood was pointing out when explaining the reason for creating SO in the first place.
> I sometimes wonder how much time I could have saved for those whose questions got closed before I could answer them.
It would be better to focus on saving time for yourself, by understanding the goal. Please read https://meta.stackoverflow.com/questions/254770 and https://meta.stackoverflow.com/questions/429808 .
3 replies →
I once published a method for finding the closest distance between an ellipse and a point on SO: https://stackoverflow.com/questions/22959698/distance-from-g...
I consider it the most beautiful piece of code I've ever written and perhaps my one minor contribution to human knowledge. It uses a method I invented, is just a few lines, and converges in very few iterations.
People used to reach out to me all the time with uses they had found for it, it was cited in a PhD and apparently lives in some collision plugin for unity. Haven't heard from anyone in a long time.
It's also my test question for LLMs, and I've yet to see my solution regurgitated. Instead they generate some variant of Newtons method, ChatGPT 5.2 gave me an LM implementation and acknowledged that Newtons method is unstable (it is, which is why I went down the rabbit hole in the first place.)
Today I don't know where I would publish such a gem. It's not something I'd bother writing up in a paper, and SO was the obvious place were people who wanted an answer to this question would look. Now there is no central repository, instead everyone individually summons the ghosts of those passed in loneliness.
The various admonitions to publish to a personal blog, while encouraging, don't really get at the 0xfaded's request which I'd summarize as follows:
With no one asking questions these technical questions publicly, where, how and on what public platform will technical people find the problems that need solving so they can exercise their creativity for the benefit of all?
> The various admonitions to publish to a personal blog, while encouraging, don't really get at the 0xfaded's request
They also completely missed the fact that 0xfaded did write a blog post and it’s linked in the second sentence of the SO post.
> There is a relatively simple numerical method with better convergence than Newtons Method. I have a blog post about why it works http://wet-robots.ghost.io/simple-method-for-distance-to-ell...
Clearly we need something in between the fauxpen-access of journals and the wilde west of the blogosphere, probably. Why wouldn't the faded ox publish in a paper? Idk, but I guess we need things similar to those circulars that British royal society members used to send to each other...except not reserved for a club. The web should be a natural at this. But it's either centralized -> monetized -> corrupted, or decentralized -> unindexed/niche -> forgotten fringe. What can come between?
14 replies →
You can (and always were encouraged to) ask your own questions, too.
And there are more sites like this (see e.g. https://codidact.com — fd: moderator of the Software section). Just because something loses popularity isn't a reason to stop doing it.
5 replies →
Seriously where will we get this info anymore? I’ve depended on it for decades. No matter how obscure, I could always find a community that was talking about something I needed solved. I feel like that’s getting harder and harder every year. The balkanization of the Internet + garbage AI slop blogs overwhelming the clearly declining Google is a huge problem.
11 replies →
> where, how and on what public platform will technical people find the problems that need solving so they can exercise their creativity for the benefit of all?
The same place people have always discovered problems to work on, for the entire history of human civilization. Industry, trades, academia, public service, newspapers, community organizations. The world is filled with unsolved problems, and places to go to work on them.
Einstein was literally a patent clerk.
This is a perfect example of an element of Q&A forums that is being lost. Another thing that I don't think we'll see as much of anymore is interaction from developers that have extensive internal knowledge on products.
An example I can think of was when Eric Lippert, a developer on the C# compiler at the time, responded to a question about a "gotcha" in the language: https://stackoverflow.com/a/8899347/10470363
Developer interaction like that is going to be completely lost.
This type of thing often lives in the issues / discussion tab of a github repo now a days, for better and worse.
6 replies →
The second answer cites Lippert's pre-existing blog post on the subject: https://ericlippert.com/2009/11/12/closing-over-the-loop-var...
I agree that there will be some degradation here, but I also think that the developers inclined to do this kind of outreach will still find ways to do it.
I believe the community has seen the benefit of forums like SO and we won’t let the idea go stale. I also believe the current state of SO is not sustainable with the old guard flagging any question and response you post there. The idea can/should/might be re-invented in an LLM context and we’re one good interface away from getting there. That’s at least my hope.
I used to look at all TensorFlow questions when I was on the TensorFlow team (https://stackoverflow.com/tags/tensorflow/info). Unclear where people go to interact with their users now....Reddit? But the tone on Reddit is kind of negative/complainy
I had a similar beautiful experience where an experienced programmer answered one of my elementary JavaScript typing questions when I was just starting to learn programming.
He didn't need to, but he gave the most comprehensive answer possible attacking the question from various angles.
He taught me the value of deeply understanding theoretical and historical aspects of computing to understand why some parts of programming exist the way they are. I'm still thankful.
If this was repeated today, an LLM would have given a surface level answer, or worse yet would've done the thinking for me obliviating the question in the first place.
I wrote a blog post about my experience at https://nmn.gl/blog/ai-and-learning
Had a similar experience. Asked a question about a new language feature in java 8 (parallell streams), and one of the language designers (Goetz) answered my question about the intention of how to use it.
An LLM couldn't have done the same. Someone would have to ask the question and someone answer it for indexing by the LLM. If we all just ask questions in closed chats, lots of new questions will go unanswered as those with the knowledge have simply not been asked to write the answers down anywhere.
3 replies →
You can prompt the LLM to not just give you the answer. Possibly even ask it to consider the problem from different angles but that may not be helpful when you don't know what you don't know.
For every example of that, there were 999 instances of people having their question closed, criticised, or ignored.
You can write a paper, submit the arxiv, and you can also make a blog post. At any rate, I agree - SO was (is?) a wonderful place for this kind of thing.
I once had a professor mention that they knew me from SO because I posted a few underhanded tricks to prevent an EKF from "going singular" in production. That kind of community is going to be hard to replace, but SO isnt going anywhere, you can still ask a question and answer your own question for permanent, searchable archive.
I would imagine the endorsement requirement reduces submissions by a few orders of magnitude.
4 replies →
Has anyone tried building a modern Stack Overflow that's actually designed for AI-first developers? The core idea: question gets asked → immediately shows answers from 3 different AI models. Users get instant value. Then humans show up to verify, break it down, or add production context. But flip the reputation system: instead of reputation for answers, you get it for catching what's wrong or verifying what works. "This breaks with X" or "verified in production" becomes the valuable contribution. Keep federation in mind from day one (did:web, did:plc) so it's not another closed platform. Stack Overflow's magic was making experts feel needed. They still do—just differently now.
Oh, so it wasn't bad enough to spot bad human answers as an expert on Stack Overflow... now humans should spend their time spotting bad AI answers? How about a model where you ask a human and no AI input is allowed, to make sure that everyone has everyone else's full attention?
6 replies →
Am I reading an AI trying to trick me into becoming its subordinate?
6 replies →
AI is generally setup to return the "best" answer as defined as the most common answer, not the rightest, or most efficient or effective answer, unless the underlying data leans that way.
It's why AI based web search isn't behaving like google based search. People clicking on the best results really was a signal for google on what solution was being sought. Generally, I don't know that LLMs are covering this type of feedback loop.
That seems like a horrible core idea. How is that different from data labeling or model evaluation?
Human beings want to help out other human beings, spread knowledge and might want to get recognition for it. Manually correcting (3 different) automation efforts seems like incredible monotone, unrewarding labour for a race to the bottom. Nobody should spend their time correcting AI models without compensation.
1 reply →
I think this could be really cool, but the tricky thing would be knowing when to use it instead of just asking the question directly to whichever AI. It’s hard to know that you’ll benefit from the extra context and some human input unless you already have a pretty good idea about the topic.
1 reply →
thanks for sharing that, it was simple, neat, elegant.
this sent me down a rabbit hole -- I asked a few models to solve that same problem, then followed up with a request to optimize it so it runs more efficiently.
chatgpt & gemini's solutions were buggy, but claude solved it, and actually found a solution that is even more efficient. It only needs to compute sqrt once per iteration. It's more complex however.
Claude's trick: instead of calling sin/cos each iteration, it rotates the existing (cos,sin) pair by the small Newton step and renormalizes:
See: https://gist.github.com/achille/d1eadf82aa54056b9ded7706e8f5...
p.s: it seems like Gemini has disabled the ability to share chats can anyone else confirm this?
Thanks for pushing this, I've never gone beyond "zero" shotting the prompt (is it still called zero shot with search?)
As a curiosity, it looks like r and q are only ever used as r/q, and therefore a sqrt could be saved by computing rq = sqrt((rxrx + ryry) / (qxqx + qyqy)). The if q < 1e-10 is also perhaps not necessary, since this would imply that the ellipse is degenerate. My method won't work in that case anyway.
For the other sqrt, maybe try std::hypot
Finally, for your test set, could you had some highly eccentric cases such as a=1 and b=100
Thanks for the investigation:)
Edit: BTW, the sin/cos renormalize trick is the same as what tx,ty are doing. It was pointed out to me by another SO member. My original implementation used trig functions
5 replies →
I can relate. I used to have a decent SO profile (10k+ reputation, I know this isnt crazy but it was mostly on non low hanging fruit answers...it was a grind getting there). I used to be proud of my profile and even put it in my resume like people put their Github. Now - who cares? It would make look like a dinosaur sharing that profile, and I never go to SO anymore.
I too, around 2012 was too much active on so, in fact, it had that counter thing continuously xyz days most of my one liners, or snippets for php are still the highest voted answers. Even now when sometimes I google something, and an answer comes up, I realize its me who asked the same question and answered it too.
I have had this experience -- twice with the same answer. There is nothing so amusing in quite this way.
I often forget just how much smaller and less siloed the internet was just ~13 years ago.
Please, start a blog! Hugo + GitHub hosting makes it laughably simple. (Or pick a different stack; that’s just mine.)
Even if you’re worried it’ll be sparse and crappy, isn’t an Internet full of idiosyncratic personal blogs what we all want?
If you want help or encouragement, reach out: zellyn@ most places
> Please, start a blog!
The second sentence of the SO post is a link to their blog where it was posted originally. The blog is not a replacement for the function SO served.
It's been a long time, but here is the writeup https://blog.chatfield.io/simple-method-for-distance-to-elli...
I don't disagree completely by any means, it's an interesting point, but in your SO answer you already point to your blog post explaining it in more detail, so isn't that the answer, you'd just blog about it and not bother with SO?
Then AI finding it (as opposed to already trained well enough on it, I suppose) will still point to it as did your SO answer.
Looks like solid code. My only gripe is the shadowing of x. I would prefer to see `for _ in range`. You do redefine it immediately so it's not the most confusing, but it could trip people up especially as it's x and not i or something.
Hahaha thanks, I never noticed that. If I ever print it out and frame it I'll be sure to fix it
That's pretty nice ;)
I once wrote this humdinger, that's still on my mostly dead personal website from 2010... one of my proudest bits of code besides my poker hand evaluator ;)
The question was, how do you generate a unique number for any two positive integers, where x!=y, such that f(x,y) = f(y,x) but the resulting combined id would not be generated by any other pair of integers. What I came up with was a way to generate a unique key from any set of positive integers which is valid no matter the order, but which doesn't key to any other set.
My idea was to take the radius of a circle that intersected the integer pair in cartesian space. That alone doesn't guarantee the circle won't intersect any other integer pairs... so I had to add to it the phase multiple of sine and cosine which is the same at those two points on the arc. That works out to:
(x^2+y^2)+(sin(atan(x/y))*cos(atan(x/y)))
And means that it doesn't matter which order you feed x and y in, it will generate a unique float for the pair. It reduces to:
x^2+y^2+( (x/y) / (x^2+y^2) )
To add another dimension, just add it to the process and key it to one of the first...
x^2+y^2+z^2+( (x/y) / (x^2+y^2) )+( (x/z) / (x^2+z^2) )
It looks like you have typos? (x^2+y^2)+(sin(atan(x/y))*cos(atan(x/y))) reduces to x^2+y^2+( (x/y) / (x^2/y^2 + 1) ) - not the equation given? Tho it's easier to see that this would be symmetrical if you rearrange it to: x^2+y^2+( (xy) / (x^2+y^2) )
Also, if f(x,y) = x^2+y^2+( (x/y) / (x^2+y^2) ) then f(2,1) is 5.2 and f(1,2) is 5.1? - this is how I noticed the mistake. (the other reduction gives the same answer, 5.4, for both, by symmetry, as you suggest)
There's a simpler solution which produces integer ids (though they are large): 2^x & 2^y. Another solution is to multiply the xth and yth primes.
I only looked because I was curious how you proved it unique!
10 replies →
You should write it up and submit it to some journal officially. Doesn't matter if it mostly duplicates your own (technically unpublished) work.
SO in 2013 was a different world from the SO of the 2020's. In the latter world your post would have been moderator classified as 'duplicate' of some basic textbook copy/pasted method posted by a karma grinding CS student and closed.
My experience as well:
Stack Overflow used to (in practice) be a place to ask questions and get help and also help others.
At some point it became all about some mission and not only was it not as useful anymore but it also became a whole lot less fun.
I have a similar story about an interesting little advance in computing that I haven't formally published anywhere, but it's at https://cs.stackexchange.com/a/171695/50292
The question boils down to: can you simulate the bulk outcome of a sequence of priority queue operations (insert and delete-minimum) in linear time, or is O(n log n) necessary. Surprisingly, linear time is possible.
On the other hand, I once implemented something to be told later it was novel and probably the optimal solution in the space.
An AI might be more likely to find it...
> Today I don't know where I would publish such a gem.
In the same blog you published it originally, then mentioning it on whatever social media site you use? So same?
Then let me quickly say: thank you! I used that algorithm three times in different projects during my academic "career" :-)
Reddit is my current go-to for human-sourced info. Search for "reddit your question here". Where on reddit? Not sure. I don't post, tbh, but I do search.
Has the added benefit of NOT returning stackoverflow answers, since StackOverflow seems to have rotted out these days, and been taken over by the "rejection police".
Sounds like this should live in Wikipedia somewhere on https://en.wikipedia.org/wiki/Ellipse...or maybe a related but more CS focused related page.
Naive question maybe but how haven’t the models been trained on your answer if it’s on SO?
Models are NOT search engines.
Even if LLMs were trained on the answer, that doesn't mean they'll ever recommend it. Regardless of how accurate it may be. LLMs are black box next token predictors and that's part of the issue.
This is a really method for solving that problem! I wouldn’t have thought to use the tangents but that makes perfect sense
If you ask me your blog post is basically a paper, I’d publish to arxiv.
That algorithm reminds me of raymarching signed distance functions.
Really great write-up, thanks for sharing it again!
Amazing work!
Very cool!
Why did SO decide to do that to us? to not invest in ai and then, iirc, claim our contributions their ownership. i sometimes go back to answers i gave, even when answered my own questions.
Decide to do what?
SO didn't claim contributions. They're still CC-BY-SA
https://stackoverflow.com/help/licensing
AFAICT all they did is stop providing dumps. That doesn't change the license.
I was very active, In fact I'm actually upset at myself for spending so much time there. That said, I always thought I was getting fair value. They provided free hosting, I got answers and got to contribute answers for others.
While AI might have amplified the end, the drop-off preceded significant AI usage for coding.
So some possible reasons:
- Success: all the basic questions were answered, and the complex questions are hard to ask.
- Ownership: In its heyday, projects used SoF for their support channel because it meant they don't have to answer twice. Now projects prefer to isolate dependencies to github and not lose control over messaging to over-eager users.
- Incentives: Good SoF karma was a distinguishing feature in employment searches. Now it wouldn't make a difference, and is viewed as being too easy to scam
- Demand: Fewer new projects. We're past the days of Javascript and devops churn.
- Community: tight job markets make people less community-oriented
Some non-reasons:
- Competition (aside from AI at the end): SoF pretty much killed the competition in that niche (kind of like craigslist).
> - Success: all the basic questions were answered, and the complex questions are hard to ask.
I think this is one major factor that is not getting enough consideration in this comment thread. By 2018-2020, it felt like the number of times that someone else had already asked the question had increased to the point that there was no reason to bother asking it. Google also continued to do a better and better job of surfacing the right StackOverflow thread, even if the SO search didn't.
In 2012 you might search Google, not find what you needed, go to StackOverflow, search and have no better luck, then make a post (and get flamed for it being a frequently-asked question but you were phrasing yours in a different / incorrect way and didn't find the "real" answer).
In 2017, you would search Google and the relevant StackOverflow thread would be in the top few results, so you wouldn't need to post and ask.
In 2020, Google's "rich snippets" were showing you the quick answers in the screen real estate that is now used by the AI Overview answers, and those often times had surfaced some info taken from StackOverflow.
And then, at the very end of 2022, ChatGPT came along and effectively acted as the StackOverflow search that you always wanted - you could phrase your question as poorly as you want, no one would flame you, and you'd get some semblance of the correct answer (at least for simple questions).
I think StackOverflow was ultimately a victim of it's own success. Most of the questions that would be asked by your normal "question asker" type of user were eventually "solved" and it was just a matter of how easy it was to find them. Google, ChatGPT, "AI Overviews", Claude Code, etc have simply made finding those long-answered questions much easier, as well as answering all of the "new" questions that could be posed - and without all of the drama and hassle of dealing with a human-moderated site.
The volume of basic questions is unlimited. There are new technologies every year.
Not sure. As software becomes a commodity I can see the "old school" like tech slowing down (e.g. programming languages, frameworks frontend and backend, etc). The need for a better programming language is less now since LLM's are the ones writing code anyway more so these days - the pain isn't felt necessarily by the writer of the code to be more concise/expressive. The ones that do come out will probably have more specific communities for them (e.g. AI)
The graph is scary, but I think it's conflating two things:
1. Newbies asking badly written basic questions, barely allowed to stay, and answered by hungry users trying to farm points, never to be re-read again. This used to be the vast majority of SO questions by number.
2. Experiencied users facing a novel problem, asking questions that will be the primary search result for years to come.
It's #1 that's being canibalized by LLM's, and I think that's good for users. But #2 really has nowhere else to go; ChatGPT won't help you when all you have is a confusing error message caused by the confluence of three different bugs between your code, the platform, and an outdated dependency. And LLMs will need training data for the new tools and bugs that are coming out.
The newbies vastly outnumber the experienced people (in every discipline), and have more to ask per-capita, and are worse at asking it. Category 2 is much smaller. The volume of Stack Overflow was never going to be sustainable and was not reasonably reflective of its goals.
We are talking about a site that has accumulated more than three times as many questions as there are articles on Wikipedia. Even though the scope is "programming languages" as compared to "literally anything that is notable".
But there are other places people can go, such as https://software.codidact.com (fd: I am a moderator there).
I’m going to argue the opposite. LLMs are fantastic at answering well posed questions. They are like chess machines evaluating a tonne of scenarios. But they aren’t that good at guessing what you actually have on your mind. So if you are a novice, you have to be very careful about framing your questions. Sometimes, it’s just easier to ask a human to point you in the right direction. But SO, despite being human, has always been awful to novices.
On the other hand, if you are experienced, it’s really not that difficult to get what you need from an LLM, and unlike on SO, you don’t need to worry about offending an overly sensitive user or a moderator. LLMs never get angry at you, they never complain about incorrect formatting or being too lax in your wording. They have infinite patience for you. This is why SO is destined to be reduced to a database of well structured questions and answers that are gradually going to become more and more irrelevant as time goes by.
Yes, LLMs are great at answering questions, but providing reasonable answers is another matter.
Can you really not think of anything that hasn't already been asked and isn't in any documentation anywhere? I can only assume you haven't been doing this very long. Fairly recently I was confronted with a Postgres problem, LLMs had no idea, it wasn't in the manual, it needed someone with years of experience. I took them IRC and someone actually helped me figure it out.
Until "AI" gets to the point it has run software for years and gained experience, or it can figure out everything just by reading the source code of something like Postgres, it won't be useful for stuff that hasn't been asked before.
"well posed questions"
And that is exactly why so many people gripe about SO being "toxic". They didn't present a well posed question. They thought it was for private tutoring, or socializing like on reddit.
All I can say to these is: Ma'am, this is a Wendy's.
1 reply →
I think you overestimate 2 by a longshot most problems only appear novel because they couched in a special field, framework or terminology, otherwise it would be years of incremental work. Some are, they are more appropriately put in a recreational journal or BB.
The reason the "experts" hung around SO was to smooth over the little things. This create a somewhat virtuous cycle, but required too much moderation and as other have pointed out, ultimately unsustainable even before the release of LLMs.
The first actually insightful comment under the OP. I agree all of it.
If SO manages to stay online, it'll still be there for #2 people to present their problems. Don't underestimate the number of bored people still scouring the site for puzzles to solve.
SE Inc, the company, are trying all kinds of things to revitalize the site, in the service of ad revenue. They even introduced types of questions that are entirely exempt from moderation. Those posts feel literally like reddit or any other forum. Threaded discussions, no negative scores, ...
If SE Inc decides to call it quits and shut the place down and freeze it into a dataset, or sell it to some SEO company, that would be a loss.
Seems like the exit was very well timed.
https://meta.stackoverflow.com/questions/408138/what-will-ha...
Another plausible explanation is that the new owners didn’t develop the community in a good way. Instead of fixing the myriad of issues that were obvious to almost all contributors they instead basically let it die?
The new owners (well, not really new any more) are focused on adding AI to SO because it's the current hotness, and making other changes to try to extract more money that they're completely ignoring the community's issues and objections to their changes, which tend to be half-assed and full of bugs.
Love this comment [1] under the post
> $1.8 billion? So do those of us who contribute get any of that?
1. https://meta.stackoverflow.com/questions/408138/what-will-ha...
Somewhere out there, there's an alternate universe in which the Stackoverflow community was so friendly, welcoming, helpful, and knowledgeable that this seems like a tragedy and motivates people to try to save it.
But in this universe, most people's reaction is just "lol".
Lots of the comments here are attributing the decline to a toxic community or overly-strict moderation, but I don't think that that is the main reason. The TeX site [0] is very friendly and has somewhat looser moderation, yet it shows the exact same decline [1].
[0]: https://tex.stackexchange.com/
[1]: https://data.stackexchange.com/tex/query/1926661#graph
A similar but not as strong decline. Taking the one but last datapoint for both (stackoverflow/tex respectively): 4436 and 394. If you compare this to how it looked like between 2015-2020 you get (my guess from scanning): 160,000 and 1700. So Stackoverflow as a whole went from 160K -> ~4.4K. That's like a 35x drop, compared to tex, where it's a 1700 -> 394, 4x drop.
Hate to argue with people on the internet, but your graph doesn't actually show what you claim. The TeX data was stable until late 2021, whereas the SO decline started in 2017. I also would expect some correlation so that SO was a drag on the TeX site.
I would ascribe that to these communities evolving differently. There is no reason to assume that the popularity of LaTeX tracks the popularity of programming languages. It's a type setting system. And that doesn't even take into account communities that exist parallel to SO/SE. Surely there exist communities today for LaTeX that have been around since before SO began its life.
1 reply →
The obvious culprit here are the LLMs, but I do wonder whether Github's social features, despite its flaws, have given developers fewer reasons to ask questions on SO?
Speaking from experience, every time I hit a wall with my projects, I would instinctively visit the project's repo first, and check on the issues / discussions page. More often than not, I was able to find someone with an adjacent problem and get close enough to a solution just by looking at the resolution. If it all failed, I would fall back to asking questions on the discussion forum first before even considering to visit SO.
I used to contribute a ton to Stack Overflow at the beginning in 2009 and 2010 and then stopped cold turkey. One of the senior product execs emailed me to see what turned me off.
What killed it for me was community moderation. People who cannot contribute with quality content will attempt to contribute by improperly and excessively applying their opinion of what is allowed.
Unfortunately it happens to every online technical community once they become popular enough. I even see it happening on HN.
Fellow OG! And it's been happening on HN since the mid-2010s, too. Moderation went out of control everywhere, but at least this site isn't branded as some strictly technical site. Can't believe I'm even saying this, but moderation on a site that encompasses a cornucopia of topics is their prerogative. The mind-boggling thing to me about SO was that the moderation used non-technical criteria (such as failing to recognize why certain problems were asked and were not dupes, and later to shoehorn in political and sexual ideologies) to shape a technical site.
HN is a sponsored marketing outlet for YC first and foremost.
HN has full time staff that provides moderation and does an excellent job. Nonetheless there are numerous users who take it upon themselves to determine what content should be available to the rest of us, as if they were heroes in their own mind.
It’s a form of narcissism. While they think of themselves as community saviors everyone else thinks they are censoring assholes. Just let the moderators do their job. Unwanted content will naturally fall off either by downvoting or it will be ignored.
All the rest of ask for is just don’t be an asshole.
1 reply →
I feel like I'm taking crazy pills, reading some of these comments.
It looks like a pretty clear divide between the people that wanted to ask questions to get solutions for their own specific problems; and those who were aware of what the site wanted to be and how it actually operated, and were willing to put in the time and answer questions, etc.
The sheer amount of garbage that used to get posted every day required some pretty heavy moderation. Most of it was not by actual moderators, it was by high-reputation users.
(I have 25K reputation on StackOverflow, and was most active between 2011 and 2018.)
I think 95% of comments earnestly using the word "toxic" can be disregarded.
They were unaware of or unwilling to follow the rules of the site. They mistook SO for reddit, a place for socializing.
Garbage was never moderated on StackOverflow, it was always ignored.
Moderation was used by the insiders to keep new people out.
And half the garbage is from people "moderating"! You are literally rewarded points for doing moderating activities, so of course every post is flooded with BS edits, votes to close, etc.. Cobra effect and whatnot.
4 replies →
As someone that spent a fair bit of time answering questions on StackOverflow, what stood out years ago was how much the same thing would be asked every day. Countless duplicates. That has all but ceased with LLMs taking all that volume. Honestly, I don't think that's a huge loss for the knowledge base.
The other thing I've noticed lately is a strong push to get non-programming questions off StackOverflow, and on to other sites like SuperUser, ServerFault, DevOps, etc.
Unfortunately, what's left is so small I don't think there's enough to sustain a community. Without questions to answer, contributors providing the answers disappear, leaving the few questions there often unanswered.
I do use Claude a lot, but I still regularly ask questions on https://bioinformatics.stackexchange.com/. It's often just too niche, LLMs hallucinate stuff like an entire non-existent benchmarking feature in Snakemake, or can't explain how I should get transcriptome aligners to give me correct quantifications for a transcript. I guess it's too niche. And as a lonely Bioinformatician it can be nice to get confirmation from other bioinformaticians.
Looking back at my Stack Exchange/Stack Overflow (never really got the difference) history, my earlier, more general programming questions from when I just started are all no-brainers for any LLM.
The decline is not surprising. I am sure AI is replacing Stackoverflow for a lot of people. And my experience with asking questions was pretty bad. I asked a few very specific questions about some deep detail in Windows and every time I got only some smug comments about my stupid question or the question got rejected outright. That while a ton of beginner questions were approved. Definitely not a very inviting club. I found i got better responses on Reddit.
Would you care to link to any of those, even ones that might appear to be deleted?
I joined Stackoverflow early on since it had a prevalence towards .NET and I’ve been working with Microsoft web technologies since the mid 90’s.
My SO account is coming up to 17 years old and I have nearly 15,000 points, 15 gold badges, including 11 famous questions and similar famous answer badges, also 100 silver and 150 bronze. I spent far much time on that site in the early days, but through it, I also thoroughly enjoyed helping others. I also started to publish articles on CodeProject and it kicked off my long tech blogging “career”, and I still enjoy writing and sharing knowledge with others.
I have visited the site maybe once a year since 2017. It got to the point that trying to post questions was intolerable, since they always got closed. At this point I have given up on it as a resource, even though it helped me tremendously to both learn (to answer questions) and solve challenging problems, and get help for edge cases, especially on niche topics. For me it is a part of my legacy as a developer for over 30 years.
I find it deeply saddening to see what it has become. However I think Joel and his team can be proud of what they built and what they gave to the developer community for so many years.
As a side note it used to state that was in the top 2% of users on SO, but this metric seems to have been removed. Maybe it’s just because I’m on mobile that I can’t see it any more.
LLM’s can easily solve those easy problems that have high commonality across many codebases, but I am dubious that they will be able to solve the niche challenging problems that have not been solved before nor written about. I do wonder how those problems get solved in the future.
Apparently it was removed to reduce the load on the database see [0], [1].
The top voted response points out that SO are [2]:
> destroying a valuable feature for users.
Kinda wild they allowed it. As that answer also suggests, perhaps rather than remove it entirely, they could just compute those stats at a lesser frequency to reduce load.
[0] https://meta.stackexchange.com/q/399661/389220
[1] https://meta.stackoverflow.com/q/437862/5783745
[2] https://meta.stackexchange.com/a/400024/389220
i've been using SO for 17 years as well but ultimately gave up out of frustration, and a lot of comments here are correctly pointing at the toxicity but the real-time chats were on a next level, it was absolutely maddening how toxic and aggressive these moderators were.
A lesson can be learned here. If you don't introduce some form of accountability for everyone that influences the product, it eventually falls apart. The problem, as we all know now, is that the moderators screwed things up, and there were no guardrails in place to stop them from killing the site. A small number of very unqualified moderators vandalized the place and nobody with common sense stepped in to put an end to it.
Interestingly, stagnation started around 2014 (in the number of questions asked no longer rising,) and a visible decline started in 2020 [1]: two years before ChatGPT launched!
It’s an interesting question if the decline would have happened regardless of LLMs, just slower?
[1] An annotated visualization of the same data I did: https://blog.pragmaticengineer.com/are-llms-making-stackover...
Some commenters suggest it's not the moderation. I think it is the key problem, and the alternative communities were the accumulated effect. Bad questions and tough answer competition is part of it, but moderation was more important, I think. Because in the end what kept SO relevant was that people made their own questions on up to date topics.
Up until mid-2010s you could make a seriously vague question, and it would be answered, satisfactory or not. (2018 was when I made the last such question. YMMV) After that, almost everything, that hadn't snap-on code answer, was labelled as offtopic or duplicate, and closed, no matter what. (Couple of times I got very rude moderators' comments on the tickets.)
I think this lead some communities to avoid this moderator hell and start their own forums, where you could afford civilized discussion. Discourse is actually very handy for this (Ironically, it was made by the same devs that created SO). Forums of the earlier generation, have too many bells and whistles, and outdated UI. Discourse has much less friction.
Then, as more quality material was accumulated elsewhere, newbies stopped seeing SO on top of search, and gradually language/library communities churned off one by one. (AI and other summaries, probably did contribute, but I don't think they were the primary cause.)
One might even say that Joel and Jeff created their very own puts on shades coding horror yeaaahhhh with SO, and it is indeed ironic that them building Discourse has created far better communities than SO ever was.
This is horrifying.
Given the fact that when I need a question answered I usually refer to S.O., but more recently have taken suggestions from LLM models that were obviously trained on S.O. data...
And given the fact that all other web results for "how do you change the scroll behavior on..." or "SCSS for media query on..." all lead to a hundred fake websites with pages generated by LLMs based on old answers.
Destroying S.O. as a question/answer source leaves only the LLMs to answer questions. That's why it's horrific.
SO has lost against LLMs because it has insistently positioned itself as a knowledge base rather than a community. The harsh moderation, strict content policing, forbidden socialization, lack of follow mechanics etc have all collectively contributed to it.
They basically made a bet because they wanted to be the full anti-thesis of ad-ridden garbage-looking forums. Pure information, zero tolerance for humanity, sterile looking design.
They achieved that goal, but in the end, they dug their own grave too.
LLMs didn’t admonish us to write our questions better, or simply because we asked for an opinion. They didn’t flag, remove our post with no advance notice. They didn’t forbid to say hello or thanks, they welcomed it. They didn’t complain when we asked something that was asked many times. They didn’t prevent us from deleting our own content.
Oh yeah, no wonder nobody bothers with SO anymore.
It’s a good lesson for the future.
IMO people underestimate the value of heavy moderation. But moderation heavy or light, good or bad.
Why wait hours for an answer when an LLM gives it in seconds?
Do I read that correctly — it is close to zero today?!
I used to think SO culture was killing it but it really may have been AI after all.
Not zero, but it is smaller than when it launched originally. And this is questions asked, not how many people are visiting and reading posts.
Still a couple thousand away from 0.
But yea the double whammy of toxic culture and LLMs did the trick. Decline already set in well before good enough LLMs were available.
I wonder how reddit compares, though its ofc pretty different use case there
Reddit is a forum morphed into social media. I usually use "question + reddit" on Google to confirm my suspicions about a subject. It is a place to discuss things rather than find answers. It is extremely politicized (leftist/liberal), but that's a whole other story.
It's surely both.
Look at the newest questions: https://stackoverflow.com/questions?tab=Newest
Most questions have negative karma.
Even if somehow that is "deserved", that's not a healthy ecosystem.
All that is left of SI are clueless questioners and bitter, jaded responders.
SO worked when "everyone" was new to it, and they felt energized to ask questions ( even "basic" questions, because they hadn't been asked before ), and felt energized to answer them.
SO solved a real problem - knowledge being locked into forum posts with no follow-up, or behind paywalls.
Most? 3 out of 15 is most? What's wrong with youngsters today?!
2 replies →
It is not "karma". It is not to be taken personally. It represents the objective usefulness of the question, not the personal worth of the person asking it.
It's both. I stopped asking questions because the mods were so toxic, and I stopped answering questions because I wasn't going to train the AI for free.
It can be both. Push and pull factors work better together than either does individually.
It's not zero but it's very low. You can glance at the site now for confirmation.
I was using the site recently (middle of a US workday) and the "live stats" widget showed 10s of questions asked per hour and ~15K current users. I have not done the work to compare these values to historical ones but they're _low_.
The last data point is from January 2026, which has just begun. If you extrapolate the 321 questions by multiplying by 10 to account for the remaining 90 % of the month, you get to within the same order of magnitude as December 2025 (3862). The small difference is probably due to the turn of the year.
There are tabs to change to a table view. I see a peak of 207k in 2014 and the last month was only 3,710.
The decline has been pretty surprising: more questions asked in May 2021 (133,914) than in the whole of 2025 (129,977).
Maybe the graph doesn’t include questions that get closed by moderators?
The steep decline started way before llms
AI didn't necessarily kill SO because it was strictly better at giving technical answers (and it certainly wasn't better when GPTs initially burst onto the mass-appeal scene several years ago), but that it provided an alternative (even if subpar) where users could actually get responses to their questions (and furthermore not be ridiculed by psychopaths while doing so was the cherry on top).
This is a huge loss.
In the past people asked questions of real people who gave answers rooted in real use. And all this was documented and available for future learning. There was also a beautiful human element to think that some other human cared about the problem.
Now people ask questions of LLMs. They churn out answers from the void, sometimes correct but not rooted in real life use and thought. The answers are then lost to the world. The learning is not shared.
LLMs have been feeding on all this human interaction and simultaneously destroying it.
StackOverflow was a pub where programmers had fun while learning programming. The product of that fun was valuable.
Instead of cultivating the pub, the owners demanded that the visitors be safe, boring and obedient witers of value. This killed the pub and with it the business.
The most visible aspect was the duplicate close. Duplicate closes scare away fresh patrons, blocking precisely the path that old timers took when they joined. And duplicates allow anyone with a grudge to take revenge. After all, there are no new questions, and you will always find a duplicate if you want to.
To create a new Stack Overlflow, create a pub where programmers enjoy drinking a virtual beer, and the value will appear by itself.
Another exemple being "Comments are not for extended discussion ! if you want to actively bring value by adding information, later updates, history, or just fun that cultivates a community, please leave and go do that somewhere else like our chat that doesn't follow at all the async functionnality of this platform and is limited to the regular userbase while scaring the newcomers."
"comments are not for extended discussion" is one of the biggest own goals of SO product development. Like, they had a feature that people were engaging with actively, and the discussions were adding value and additional context to posts, and they decided "yeah, let's kill this".
The people who run SO have some sort of control-freak complex. If there's anything I've learned from the SO saga, it is that oftentimes just letting a community do what it wants (within reasonable boundaries, of course) leads to a better and more successful product than actively trying to steer things in a certain direction.
1 reply →
hahaha, I almost forgot about that! "stop talking about edge cases and other things pertinent to this topic in comments about this topic!! reeeeeeeee!!!!"
>StackOverflow was a pub where programmers had fun while learning programming. The product of that fun was valuable.
I really like this description. I and others here who are talking about negative experiences there seem to decry how we enjoy programming (you see words like "fun" and "passion" used in these posts), and how SO decided to take this good faith and cheer and bludgeon users for often opaque reasons, just so they could power trip. As much as I have many reservations about LLMs, I can ask LLMs to be as emotionless (or even emotional but chipper/happy) as I want. On SO, you needed to prostrate yourself and self-criticize to even have the opportunity to be bludgeoned further by the moderators. Who tf would want to spend their time contributing there? Even if you contributed a decent or even great amount to the site, you would still get whacked over the head if you dared to ask a question of your own.
This is why people jumped to LLMs, even when they were far less capable than they are now. Most people (SO moderators don't view others as "people", as is apparent in this thread) would rather receive mid-tier answers from an LLM (though LLMs have now exceeded this level of quality) while still having fun, than get castigated and "closed as duped" on SO.
SO was built to disrupt the marriage of Google and Experts Exchange. EE was using dark patterns to sucker unsuspecting users into paying for access to a crappy Q&A service. SO wildly succeeded, but almost 20 years later the world is very different.
I recall when they disabled the data export a few years ago [0], March 2023. Almost certainly did this in response to the metrics they were seeing, but it accelerated the decline [1].
[0] https://meta.stackexchange.com/questions/389922/june-2023-da...
[1] https://data.stackexchange.com/stackoverflow/query/edit/1926...
As one of my good friends pointed out back in 2012, most people don't know how to ask questions[0].
I'm feeling a bit sorry for zahlman in the comment section here, they're doing a good job of defending SO to a comment section that seems to want SO to bend to their own whims, no matter what the stated aims and goals of SO really were. There does seem to be a lot of people in the comments here who wanted SO to be a discussion site, rather than the Q&A site that it was set out to be.
I do think it's very unfair of many of you who are claiming SO was hostile or that they unfairly closed questions without bringing the citations required. I'm not saying at all that SO was without it's flaws in leadership, moderators, community or anything else that made the site what it was. But if you're going to complain, at least bring examples, especially when you have someone here you could hold somewhat accountable.
The problem is, you still see a lot of it today, whether it's in IRC channels, Discord chats, StackOverflow or GitHub issues. People still don't know how to ask questions:
* [1] * [2] * [3]
[0]: https://blog.adamcameron.me/2012/12/need-help-know-how-to-as... [1]: https://github.com/swagger-api/swagger-ui/issues/10670 [2]: https://github.com/swagger-api/swagger-ui/issues/10649 [3]: https://github.com/usebruno/bruno/issues/6515
One factor I haven't seen mentioned is the catastrophic decline in quality of Google search. That started pre-llm and now the site is almost unusable to search web. You can access something you know exists and you know where it exists, but to actually search..?
Most SO users are passive readers who land there using search, but these readers are also the feed of new active users. Cut off the influx, and the existing ones will be in decline (the moderation just accelerates it).
As everyone is saying, it was already down-trending before AI, and probably experts exchange traffic and whatever came before looks similar
Also not sure exactly when they added the huge popup[0] that covers the answer (maybe only in Europe as it's about cookies?) but that's definitely one of the things that made me default reach for other links instead of SO.
[0] https://i.imgur.com/Z7hxflF.png
Those popups were a big contributor for me to stop using SO. I stopped updating my uBlock origin rules when LLMs became good enough. I am now using the free Kimi K2 model via Groq over CLI, which is much faster.
Here’s how SO could still be useful in the LLM era:
User asks a question, llm provides an immediate answer/reply on the forum. But real people can still jump in to the conversation to add additional insights and correct mistakes.
If you’re a user that asks a duplicate question, it’ll just direct you to the good conversation that already happened.
A symbiosis of immediate usually-good-enough llm answers PLUS human generated content that dives deeper and provides reassurances in correctness
Users could upvote whether Claude, Gemini or ChatGPT provided the best answer. The best of three is surfaced, the others are hidden behind a "show alternatives."
However, I can see how this would be labelled "shoving AI into everything" and "I'm not on SO for AI."
Or they can start claiming copyright on the training content
Should probably email this to the CEO of SO
StackExchange forgot who made them successful long ago. This is what they sowed. I don't have any remorse, only pity.
When Hans Passant (OGs will know) left, followed by SE doing literally nothing, that was the first clue for me personally that SE stopped caring.
That said, it is a bit shocking how close to zero it is.
Stackoverflow is like online gaming--lots of toxic people, but I still get value out of it. Ignore the toxic people, get your questions answered and go home to your family with your paycheck.
It's surprisingly tame still given it interests tens (hundreds?) of millions of people at varying age and background and mostly when the mind is occupied by a problem. I always found it surprising there's not more defacing and toxicity.
I was tasked to add OpenOffice's hyphenation lib to our software at work back in 2010 when I was a junior dev. I had to read the paper and the C code/documentation to understand how it works but got stuck in one particular function.
It was such an obscure thing (compare to web dev stuffs) that I couldn't find anything on Google.
Had no choice but to ask on Stackoverflow and expected no answers. To my surprise, I got a legit answer from someone knowledgable, and it absolutely solve my problem at the time. (The function has to do with the German language, which was why I didn't understand the documentation)
It was a fond memory of the site for me.
AI is a vampire. Coming to your corner of the world, to suck your economic blood, eventually. It’s hard to ignore the accelerated decline that started in late 2022/early 2023.
LLMs absolutely body-slammed SO, but anyone who was an active contributor knows the company was screwing over existing moderators for years before this. Writing was on the walls
If by "body-slammed" you mean "trained on SO user data while violating the terms of the CC BY-SA license", then sure.
In the best case scenario, LLMs might give you the same content you were able to find on SO. In the common scenario, they'll hallucinate an answer and waste your time.
What should worry everyone is what system will come after LLMs. Data is being centralized and hoarded by giant corporations, and not shared publicly. And the data that is shared is generated by LLMs. We're poisoning the well of information with no fallback mechanism.
> If by "body-slammed" you mean "trained on SO user data while violating the terms of the CC BY-SA license", then sure.
You know that's not what they meant, but why bring up the license here? If they were over the top compliant, attributing every SO answer under every chat, and licensing the LLM output as CC BY-SA, I think we'd still have seen the same shift.
> In the best case scenario, LLMs might give you the same content you were able to find on SO. In the common scenario, they'll hallucinate an answer and waste your time.
Best case it gives you the same level of content, but more customized, and faster.
SO being wrong and wasting your time is also common.
So the question for me is how important was SO to training LLMs? Because now that the SO is basically no longer being updated, we've lost the new material to train on? Instead, we need to train on documentation and other LLM output. I'm no expert on this subject but it seems like the quality of LLMs will degrade over time.
Yep, exactly. Free data grabbing honeypots like SO won't work anymore.
Please mark all locations on the map where you would hide during the uprise of the machines.
Why publish anything for free on the internet if it's going to be scanned into some corporation's machine for their free use? I know artists who have stopped putting anything online. I imagine some programmers are questioning whether or not to continue with open source work too.
It has often been claimed, and even shown, that training LLMs on their own outputs will degrade the quality over time. I myself find it likely that on well-measurable domains, RLVR improvements will dominate "slop" decreases in capability when training new models.
Are we in the age of all CS problems being solved and everything being invented? Even if so, do LLM incorporate all that knowledge?
A lot of my knowledge in CS come from books and lectures, LLMs can shine in that area by scraping all those sources.
However SO was less about academic knowledge but more about experience sharing. You won't find recipes for complex problems in books, e.g. how to catch what part of my program corrupts memory for variable 'a' in gdb.
LLMs know correct answer to this question because someone shared their experience, including SO.
Are we Ok with stopping this process of sharing from one human to another?
There's https://cstheory.stackexchange.com/ for the more academic CS questions btw.
it is indeed a shame. if you are doing anything remotely new and novel, which is essential if you want to make a difference in an increasingly competitive field, LLMs confidently leave you with non-working solutions, or sometimes worse they set you on the wrong path.
I had similar worries in the past about indexable forums being replaced by discord servers. the current situation is even worse.
Web has been solved for a decade imo.
LLMs caused this decline. Stop denying that. You don't have to defend LLMs from any perceived blame. This is not a bad thing.
The steep decline in the early months of 2023 actually started with the release of ChatGPT, which is 2022-11-30, and its gradually widening availability to (and awareness of) the public from that date. The plot clearly shows that cliff.
The gentle decline since 2016 does not invalidate this. Were it not for LLMs, the site's post rate would now probably be at around 5000 posts/day, not 300.
LLMs are to "blame" for eating all the trivial questions that would have gotten some nearly copy-pasted answer by some eager reputation points collector, or closed as a duplicate, which nets nobody any rep.
Stack Overflow is not a site for socializing. Do not mistake it for reddit. The "karma" does not mean "I hate you", it means "you haven't put the absolute minimum conceivable amount of effort into your question". This includes at least googling the question before you ask. If you haven't done that, you can't expect to impose on the free time of others.
SO has a learning curve. The site expects more from you than just to show up and start yapping. That is its nature. It is "different" because it must be. All other places don't have this expectation of quality. That is its value proposition.
For those who miss SO, check out Stack Overflow Simulator: A functional museum for developers to relive the good ol' days of asking innocent questions and being told to "RTFM"
https://sosimulator.xyz/
Good.
This is what Stack Overflow wanted. They ban anyone who asks stupid questions, if not marking everything off topic.
LLMs are a solid first response for new users, with Reddit being a nice backup.
Makes for a good conspiracy theory. Bad actors intentionally making the internet hostile. https://youtu.be/qOTYgcdNrXE
Game over. I didn’t notice all the toxicity mentioned in the other comments, although I did stop using it around 2016 maybe. It had its days, it was fundamentally a verb at some point. Its name is part of web history, and there’s no denying that.
They're desperately trying to save it e.g. by introducing "discussions" which are just questions that would normally have been closed. The first one I saw, the first reply was "this should have been a question instead of a discussion".
Let's never forget that Stackoverflow was killed by its mods. Sure, it needed AI as an alternative so people could actually leave, but the thing that actually pushed them away was the mods.
StackOverflow cemented my fears of asking questions. Even though there were no results for what I needed, I was too afraid to ask.
Good riddance, now I’m never afraid to ask dumb questions to LLM and I’ve learned a lot more with no stress of judgement.
Maybe it's a mix of me using the site less, or questions I previously answered not being as relevant anymore, however as it stands, it's just not fun to visit the site any more.
I have about ~750 answers and 24K rep after almost 12 years of being a member. The site was a great way to spend some free cycles and help people. My favorite bounty answer lead to me finding a bug in the Java compiler! I even got recruited into my current role from the old Stack Overflow Jobs board.
With AI, not only did the quality and frequency of posts go down, but the activity on my existing posts are basically zero now. I used to have a few notifications a week with either comments on my past answers/questions or a few upvotes (for those fun little serotonin boosts). Looking at my past stats.. in 2023 I had ~170 notifications, in 2024 that dropped to ~100, and in 2025 it went down to ~50 (with only 5 notifications since September).
I don't feel engaged with the community, and even finding new questions to answer is a struggle now with (the unanswerable) "open-ended questions" being mixed into the normal questions feed.
People are still asking questions, it's no longer on the public internet. Google, Anthropic, OpenAI etc get to see and use them.
This is concerning on two fronts. The questions are no longer open (SO is CC-BY-SA) and if Q&A content dies then this herds even more people towards LLM use. It's basically draining the commons.
Yup. This, to me, provides another explanation for why the social contract is being used as toilet paper by the owner class. They literally see the writing on the wall.
For this occasion, I just logged in to my SO profile; I've been a member for 9 years now.
To me, back when I started out learning web dev, as a junior with no experience and barely knowing anything, SO seemed like a paradise for programmers. I could go on there and get unblocked for the complex (but trivial for experts) issues I was facing. Most of the questions I initially posted, which were either closed as duplicates or "not good enough," really did me a lot of discouragement. I wasn't learning anything by being told, "You did it wrong, but we're also not telling you how you could do it better." I agree with the first part; I probably sucked at writing good questions and searching properly. I think it's just a part of the process to make mistakes but SO did not make it better for juniors, at least on the part of giving proper guidance to those who "sucked".
They had pretty neat infra, maybe it still runs in the same clever way. https://meta.stackexchange.com/questions/10369/which-tools-a...
Recently they've chucked out their own servers:
https://stackoverflow.blog/2025/12/24/the-great-unracking-sa...
Notably the default Redis client for most .NET developers is still StackExchange.Redis.
I fit a Bass product lifetime model on earlier related StackOverflow data, it looked bad at the time. https://win-vector.com/2025/03/02/best-before-dates-by-bass/
SO peaked long, long before LLMs came along. My personal experience is that GitHub issues took over.
You can clearly see the introduction of ChatGPT in late 2022. That was the final nail in the coffin.
I am still really glad that Stack Overflow saved us from experts-exchange.com - or “the hyphen site” as it is sometimes referred to.
Seems like the sharp decline started shortly after they were sold to a private equity firm.
Was already dying a decade ago, but AI pretty much guarantees we'll never see a public forum that useful ever again.
AI may be fine for people asking the basic stuff or who don't care about maintenance, but for a time SO was the place to find extraordinary and creative solutions that only a human can come up with.
When you were in a jam and found a gem on there it not only solved your problem, but brought clarity and deep knowledge to your entire situation in ways that I've never seen an LLM do. It inspired you to refactor the code that got you into that mess to begin with and you grew as a developer.
This timeline shows the death of a lot more than just the site itself.
Code golfing on SO was fun, too!
I just looked it up, and "Note: This tag is currently blacklisted and can no longer be used." lmfao, what a braindead site, so glad I left years ago, after many years of greatly reduced activity. Source (at the top): https://stackoverflow.com/questions/tagged/code-golf?tab=New...
Look at fun stuff like this from 2010: https://stackoverflow.com/questions/2440314/code-golf-%cf%80...
You do realize that they made an entire site for CodeGolf that is reasonably active (and has its own culture... and lead to the creation of specialized languages for it... and even pushed the bounds of OEIS a few times - https://codegolf.meta.stackexchange.com/q/5318)?
https://codegolf.stackexchange.com
The issue is that code golf didn't fit well into the intended design of the library and it split off 15 years ago https://stackexchange.com/sites?view=list#oldest
1 reply →
Wow. I was expecting a decline but not to that extent.
Whenever I see mention of stack overflow’s decline I think of “StackOverflow does not want to help you”
https://news.ycombinator.com/item?id=42246333
Everything we have done and said on the internet since its birth has just been to train the future AI.
I ended up having a high reputation on SO. Not sure why, but it’s over 7000.
I also experienced many of the issues I see described here. The most egregious was when I asked a completely valid question for R: How to fit a curve through a set of points, with each point having an error associated.
This is something completely normal in a physics experiment. Each measurement had its own error interval. But, for people using R, this seemed like something completely new. So, they just downvoted the question and told me I was wrong.
I ended up answering my own question… but was also told that was wrong and that all points must have the same error interval.
Instead of answering a programming question, people just went around denying experimental physics.
I think that was the beginning of the end of SO for me.
This entire thread is fantastic. I felt nostalgic, angry and then concerned all at once.
I love LLMs. But I miss SO. I miss being able to have that community. How do we bring it back?
If anyone from the Stack Overflow team is reading this (I assume you are): what’s the plan?
My take: stop optimizing for raw question volume and start optimizing for producing and maintaining “known good” public knowledge. The thing SO still has that Discord and LLMs don’t is durable, linkable, reviewable answers with accountable humans behind them. But the workflow needs to match how devs work now.
A concrete idea: make “asking” a guided flow that’s more like opening a good GitHub issue. Let me paste my error output, environment, minimal repro, what I tried, and what I think is happening. Then use tooling (including an LLM if you want) to pre check duplicates, suggest missing details, and auto format. Crucially: don’t punish me for being imperfect. Route borderline questions into a sandbox or draft mode where they can be improved instead of just slammed shut.
Second idea: invest hard in keeping answers current. A ton of SO is correct but stale. Add obvious “this is old” signaling and make it rewarding to post updates, not just brand new answers.
Last thing that I don’t see an easy answer to: LLMs are feasting on old SO content today. But LLMs still need fresh, high quality, real world edge cases tomorrow. They need the complexity and problem solving that humans provide. A lot of the answers I get are recycled. No net new thinking. If fewer people ask publicly, where does that new ground truth come from? What’s the mechanism that keeps the commons replenished?
So… TLDR…my question to this group of incredibly intelligent people: how does SO save itself?
One thing you won’t get with in an LLM is genuine research. I once answered a 550 point question by researching the source code of vim to see how the poster’s question could be resolved. [0]
[0] https://stackoverflow.com/questions/619423/backup-restore-th...
Wow, that's not just collapsing, that's collapsed.
People are mentioning the politicization of moderation. But also don’t forget when Joel broke the rules to use the site to push his personal political agenda.
The number of active users at StackOverflow started dropping in the middle of 2020, i.e. long time before ChatGPT release in the end of 2022.
https://data.stackexchange.com/stackoverflow/revision/192836...
Are there any publicly available options to actually interact with real people about software development anymore? There doesn't seem to be anywhere that's accessible with something like a google search... Sure there are derelict IRC/Discord/$language forums, but of the handful I've been part of they aren't active or in the case of discord, weirdly disjointed.
AI is great and all, but somewhere with a little bit of an opinion and push back to carelessly thrown out questions would be nice (as a thrower of careless questions).
SO obviously went off the toxic deep end, but has that culture of shared problem solving just died completely online?
I find a lot of good stuff in GitHub issues
Local meet-ups I guess?
Good riddance. There were some ok answers there, but also many bad or obsolete answers (leading to scrolling down find to find the low-ranked answer that sort of worked), and the moderator toxicity was just another showcase of human failure on top of that. It selected for assholes because they thought they had a captive, eternally renewing audience that did not have any alternative.
And that resulted in the chilling effect of people not asking questions because they didn't want to run the moderation gauntlet, so the site's usefulness went even further down. Its still much less useful for recent tech, than it is for ancient questions about parsing HTML with regex and that sort of thing.
LLMs are simply better in every way, provided they are trained on decent documents. And if I want them to insult me too, just for that SO nostalgia, I can just ask them to do that and they will oblige.
Looking forward to forgetting that site ever existed, my brain's health will improve.
If nobody is on StackOverflow, What will LLM's train on for new problems?
GitHub Issues and Disscussions + searching the code base, fetching the docs and some reasoning on top. Maybe even firing up a sandbox VM and testing some solutions.
"firing up a sandbox VM and testing some solutions"
If the LLM can start up a VM and test a solution, to identify a new unique problem, and find it's own solution. That would be pretty impressive. I'm not sure they are really to that point. But some AI's are winning the Math Olympiad, so maybe it is happening. I'm sure this is the overall goal.
Everyone agrees their community and moderators turned toxic. But why? Was it inevitable that people would turn bitter / jaded after answering questions for years? Was it wrong incentives from StackOverflow itself? The outside tech environment becoming worse?
The precipitous decline was already happening long before LLM's dealt the final blow.
Yes, it was intended by SO itself. Basically moderate mercilessly. See posts by Jeff Atwood:
> Avoid asking questions that are subjective, argumentative, or require extended discussion. This is not a discussion board, this is a place for questions that can be answered!
https://stackoverflow.blog/2010/01/04/stack-overflow-where-w...
> Certainly on Stack Overflow and Stack Exchange we are very much pro-moderation -- and more so with every passing year.
https://stackoverflow.blog/2012/01/31/the-trouble-with-popul...
> Stack Overflow – like most online communities I’ve studied – naturally trends toward increased strictness over time. It’s primarily a defense mechanism, an immune system of the sort a child develops after first entering school or daycare and being exposed to the wide, wide world of everyday sneezes and coughs with the occasional meningitis outbreak. It isn’t always a pleasant process, but it is, unfortunately, a necessary one if you want to survive. > All the content on the site must exist to serve the mission of learning over entertainment – even if that means making difficult calls about removing some questions and answers that fail to meet those goals, plus or minus 10 percent.
https://blog.codinghorror.com/new-programming-jargon/
I think the biggest issue, what lead to the toxicity, came down to the question/answer format not suiting the problem it was trying to solve — The answer could only be as good as the original question, and the platform gave little leeway to "get to the bottom" of the problem. Getting to a high-quality question/response required a back-and-forth that the platform made difficult by burying the discovery/definition work in comments and edits instead of a clear discussion mechanism.
All of this meant the learning-curve on how to participate was high, and this spurred gate-keeping and over-zealous moderation. High-quality but out-of-date information was preferred over lower-quality but more recent updates. When combined with the rapid shifts brought on with mobile development and web frameworks, the answers could easily get out-of-date months after being answered.
I remember a time when StackOverflow dominated every search query. Now we're seeing searches take you to a dedicated forum/discussion board, which feels more appropriate for the current state of the industry.
The question askers got stupider and stupider.
The corresponding answers graph: https://data.stackexchange.com/stackoverflow/query/1927992/a...
The founders made the right move selling it when they did. No way that site is worth $1.8 billion now.
Approaching 0 is wild
Moderator team must be over the moon
Spolsky and co. sold SO in 2021 - timing IS everything.
I still would like to get other humans' experiences and perspectives when it comes to solving some problems, I hope SO doesn't go away entirely.
With LLMs, at least in my experience, they'll answer your question best they can, just as you asked it. But they won't go the extra step to make assumptions based on what they think you're trying to do and make recommendations. Humans do that, and sometimes it isn't constructive at all like "just use a different OS", but other times it could be "I don't know how to solve that, but I've had better lack with this other library/tool".
Don't lose sight of one of the dreams of the early Internet: How do we most effectively make a marketplace for knowledge problems and solutions that connects human knowledge needs with AI and human responses?
It should be possible for me to put a question out there (not on any specific forum/site specific to the question), and have AI resource answer it and then have interested people weigh in from anywhere if the AI answer is unsatisfactory. Stackoverflow was the best we could do at the time, but now more general approach is possible.
IMHO Good Riddance to such a toxic community.
I think the disallowing of “controversial” technical questions might have helped as much as the AI boom.
So frustrating to be reading a deeply interesting technically and intense debate to be closed down by an admin.
I'm glad I learned how to program when you could coax useful answers from Google searches.
Whenever a Stack Overflow result comes up now the answer is years old and wrong, you might as well search archive.org.
I have a SO profile and I both contributed and used the site for some time.
I use the site from time to time to research something. I know a lot more about software than 15 years ago.
I used to ask questions and answer questions a lot, but after I matured I have no time and whatever I earn is not worth my time.
So perhaps the content would grow in size and quality if they rewarded users with something besides XP.
I don't use AI for research so far. I use AI to implement components that fit my architecture and often tests of components.
StackOverflow didn't feel like a welcoming and humane place the last 10+ years, at least for me.
Actually I think it never did.
It started when I was new there and couldn't write answers, just write comments and then got blasted for writing answer-like comments as comments. What was I supposed to do? I engaged less and less and finally asked them to remove my account.
And then it seems like the power-users/moderators just took over and made it even more hostile.
I hope Wikipedia doesn't end up like this despite some similarities.
I don't think the reputation system ever worked that way - new users could always answer questions, but comments required more reputation.
OK, you might be right and I got it backwards. It still felt wrong at the time before I got enough points.
Where will LLMs be trained if no-one generates new posts and information like this? Do we sort of just stop innovating here in 2026? Probably not but it's a serious consideration.
Ideally, you'd train them on the core documentation of the language or tool itself.
Hopefully, LLMs lead to more thorough documentation at the start of a new language, framework, or tool. Perhaps to the point of the documentation being specifically tailored to read well for the LLM that will parse and internalize it.
Most of what StackOverflow was was just a regurgitation of knowledge that people could acquire from documentation or research papers. It obviously became easier to ask on SO than dig through documentation. LLMs (in theory) should be able to do that digging for you at lightning speed.
What ended up happening was people would turn to the internet and Stack Overflow to get a quick answer and string those answers together to develop a solution, never reading or internalizing documentation. I was definitely guilty of this many times. I think in the long run it's probably good that Stack Overflow dies.
Why do you think people stop creating new posts just because SO collapsed? People on GitHub issues and Reddit answer programming questions everyday.
SO was dying even before ChatGPT was released. LLMs just accelerated that process.
Before writing the comment I had in my head I did a CTRL+F search for "toxic" in the comment section here. 42 occurences. It says everything about what's happening to SO.
Between 2017 and 2022 (pre-LLM), it appears to show a clear downward trend, ignoring the covid surge. Any ideas why this might be?
The query also filters to PostTypeId = 1, what does this refer to?
Incompetent moderation and the air of hostility towards contributing users.
PostTypeId = 1 means "only select questions."
2 would be answers.
There is a bunch more of further post types: https://meta.stackexchange.com/a/2678
A decline in number of questions asked can also be because most people's questions are already answered in the database.
How would you query this for post views over time?
Now imagine what happens when a new programming language comes along. When we have a question, we will no longer be able to Google it and find answers to it on Stack Overflow. We will ask the LLMs. They will work it out. From that moment, the LLM we used has the knowledge for solving this particular problem. Over time, this produces huge moat for the largest providers. I believe it is one of the subtler reasons why the AI race is so fierce.
Stagnation started around 03/2014 and downward trend started around 03/2017. Looking at dates, it doesn't seem like AI caused those trend changes.
It's a very toxic place, you ask a doubt, and someone will abuse you, down vote you, make you feel you are not for to be a human. Better it's dead.
I find this quite worrying: with this much decline SO might end up disappearing. This would be a very bad thing because in some answers there are important details and nuances that you only see by looking at secondary answers and comments. Also, this seems to imply that most people will just accept the solutions proposed by LLMs without checking them, or ever talking about the subject with other humans.
It's unfortunate that SO hasn't found a way to leverage LLMs. Lots of questions benefit from some initial search, which is hard enough that moderators likely felt frustrated with actual duplicates, or close enough duplicates, and LLMs seem able to assist. However I hope we don't lose the rare gem answers that SO also had, those expert responses that share not just a programming solution but deeper insight.
I think that SO leveraging LLMs implicitly. Like I'll always ask LLM first, that's the easiest option. And I'll only come to SO if LLM fails to answer.
Interesting timing. I just analyzed TabNews (Brazilian dev community) and ~50% of 2025 posts mention AI/LLMs. The shift is real. The 2014 peak is telling. That's before LLMs, before the worst toxicity complaints. Feels like natural saturation, most common questions were already answered. My bet, LLMs accelerated the decline but didn't cause it. They just made finding those existing answers frictionless.
Good times. Although, I have to say, I was getting sick of SO before the LLM age. Modding felt a bit tyrannical, with a fourth of all my questions getting closed as off topic, and a lot of aggressive comments all around the site (do your homework, show proof, etc.)
Back when I was an active member (10k reputation), we had to rush to give answers to people, instead of angrily down voting questions and making snark comments.
How is Experts Exchange doing?
SO has been a curse on technology. I've met teams of people who decide whether to adopt some technology based solely on if they can find SO answers for it. They refuse to read documentation or learn how the technology works; they'll only google for SO answers, and if the answer's not there, they give up. There's an entire generation like this now.
I used to joke that when SO goes under, I will move professions. The joke came from my experience of how many common issues in technology could not be solved with knowledge found via a search engine. I don’t see that niche as gone, so I wonder what is satisfying that requirement such that new questions do not show up at SO?
Why did SO traffic halve from it's maximum till the ChatGPT release date? Also, for a long time after initial release, ChatGPT was pretty much useless for coding questions. It only became more or less useful ~2 years ago. So there's about 4x decline from peak to explain with reasons that do not involve LLMs. What these could be?
There's no doubt that generally LLMs are better. In addition SO had its issues. That being said I can't help but worry about losing humans asking questions and humans answering questions. The sentimentality aside, if humans aren't posing questions and if humans aren't recommending answers, what are the models going to use?
Stack Overflow set out to be a better Q&A site but has turned into a user-unfriendly, gatekeeping platform where questions are often marked as duplicates because a similar question was answered 15 years ago. Everything and every question is banned, gatekeep, or marked as a duplicate.
For me, my usage of SO started declining as LLMs rose. Occasionally I still end up there, usually because a chat response referenced a SO thread. I was willing to put up with the toxicity as long as the site still had technical value for me.
But still, machines leave me wanting. Where do people go to ask real humans novel technical questions these days?
> Where do people go to ask real humans novel technical questions these days?
I don't think such generic place exists. I just do my own research or abandon the topic. I think that in big companies you probably could use some internal chats or just ask some smart guy directly? I don't have that kind of connections and all online communities are full of people whose skill is below mine, so it makes little sense to ask something. I still do sometimes, but rarely receive competent answer.
If you have some focused topic like a question about small program, of course you can just use github issues or email author directly. But if you have some open question, probably SO is the only generic platform out there.
To put it differently, find some experts and ask which online place to the visit to help strangers. Most likely they just don't do it.
So for me, personally, LLMs are the saviour. With enough forth and back I can research any topic that doesn't require very deep expertise. Sure, access to an actual expert willing to guide me would be better, but I just don't have that luxury.
Find the relevant discord and search.
I think one of the phenomenon that people haven't mentioned is that the question space was heavily colonized by 2016.
I was one of the top 30 or 50 answerers for the SVG tag on SO, and I found that the question flow started to degrade around 2016, because so many of the questions asked had been answered (and answered well) already.
While I generally agree with the narrative of the negative arc that stack overflow took, I found (and have as recently as a few months ago) that I could have enjoyable interactions on the math, Ux, written language, and aviation exchanges. The OS ones in the middle (always found the difference between Linux and superuser confusing).
While SO is mostly dead, narrower stackexchange communities may be very much alive. E.g. the Emacs community is responsive.
Obviously LLMs ate StackOverflow, but perhaps developers could keep it alive for much longer if they wanted to. LLMs provide answers, but only humans provide human contact.
And that last part is where SO failed by allowing a few people power trip over the rest of us. Kind of like reddit does at times, but harder.
I'm not sad.
LLMs did not eat SO, it was SO that fed the LLMs too well.
https://meta.stackexchange.com/questions/399619/our-partners...
Clicked on the link and got stopped by cloudflare. Guess I won't be giving them any more traffic either.
I am surprised at the amount of hate for Stack Overflow here. As a developer I can't think of a single website that has helped me as much over the last ten years.
It has had a huge benefit for the development community, and I for one will mourn its loss.
I do wonder where answers will come from in the future. As others have noted in this thread, documentation is often missing, or incorrect. SO collected the experiences of actual users solving real problems. Will AI share experiences in a similar way? In principle it could, and in practice I think it will need to. The shared knowledge of SO made all developers more productive. In an AI coded future there will need to be a way for new knowledge to be shared.
It makes sense to see the number of questions decline over time as people google questions and get results. It would be interesting to look at the number of comments and views of questions over time to see if that has declined as LLMs have driven declining engagement and discussion.
There are still airgapped places in the world where transferring information to offsite LLMs is expressly forbidden, but the offline LLMs available perform so terribly that they’re not worth using. An SO type application can be immensely helpful for engineering teams working in these environments.
But surely transmitting information to actual SO is just as forbidden?
And if you're making an internal-only site, it doesn't really need to be name-brand SO.
Stack overflow was useful with a fairly sanitized search like “mysql error 1095”. Agentic LLMs do there best work when able to access your entire repository or network environment for context, which is impossible to sanitize. For a season, private environments will continue to be able to use SO. But as LLMs capture all the good questions and keep them private, public SO will become less and less relevant. It’s sad to see a resource of this class go.
1 reply →
https://stackoverflow.co/internal/
I think the bigger point we should realize is LLMs offer the EXACT same thing in a better way. Many people are still sharing answers to problems but they do it through an AI which then fine tunes on it and now that problem solution is shared with EVERYONE.
Far better method of automated sharing of content
I wonder what the April 2020 spike is about... maybe lockdowns meant people started learning new stuff?
Wonder if this is a good proxy for '# of Google Searches'. Or perhaps a forward indicator (sign of things to come), since LLMs are adopted by the tech-savvy first, then the general public a little later, so Stack Overflow was among the first casualties.
I recently wrote a blog post similar to this situation: https://ertu.dev/posts/ai-is-killing-our-online-interaction/
I'd still use SO at times if it weren't for how terribly it was managed and moderated. It offers features that LLMs can't, and I actually enjoyed answering questions enough to do it quite often at one time. These days I don't even think about it.
It is always good to see other cultured people who structure their SQL queries the right way.
On what will the LLMs train, now?
On the same 14 year old Java questions like the rest of us.
user chat logs clearly. They are not much diffent than the SO Q&A format.
Seems like there are "blocked by cloudflare" number of questions per month.
Their blocking of everyone not using chrome/etc from accessing their website probably contributed quite a bit to the implied downturn I'm reading in other comments.
Even if you get on the site you'll end up dealing with Cloudflare's BBBaaS (breaking back buttons as a service).
Surprising to see it bottom out so hard.
I imagine at least some of the leveling off could be due to question saturation. If duplicates are culled (earnestly or overzealously) then there will be a point where most of the low hanging fruit is picked.
The SO mission is complete. It's now an LLM training set.
Things would be different if we didn't.
This is the elephant in the room most commenters chose to turn their blind eye to.
[dead]
There will be a generation of coders that will never have heard of stack overflow.
When you see AI giving you back various coding snippets almost verbatim from SO, it really makes you wonder what will happen in the future with AI when it can't depend on actual humans doing the work first.
I certainly use it less now that I get a CloudFlare check every time I go and sometimes it fails or loads forever. I usually just go back to search results and look elsewhere after a second or two.
This is a great example of how free content was exploited by LLMs and used against oneself to an ultimate destruction.
Every content creator should be terrified of leaving their content out for free and I think it will bring on a new age of permanent paywalls and licensing agreements to Google and others, with particular ways of forcing page clicks to the original content creators.
What if we filter out all the questions closed as dupes, off topic, etc?
Man after reading some of the comments and looking at the graph I have learned a lesson. I went to SO all the time to find answers to questions, but I never participated. I mean they made it hard, but given the amount of benefit I gained I should've overcome that friction. If I and people like me had, maybe we could have diluted the moderation drama that others talk about (and that I, as a greedy user, never saw). Now it's a crap-shoot with an LLM instead of being able to peruse great answers from different perspectives to common problems and building out my own solution.
So it seems all the questions have now been answered– Great!
I suspect a lot of the traffic shift is from Google replacing the top search result, which used to be Stack Overflow for programming questions, with a Gemini answer.
Someone needs to archive the entirety of StackOverflow and make it available over torrent so that it can be preserved when the site shuts down. Urgently.
ArchiveTeam got alerted to this thread on IRC and may do something. Current non-updated docs about SO.
https://wiki.archiveteam.org/index.php/Stack_Exchange
https://archive.org/details/stackexchange Found it
https://archive.org/details/stackexchange_20250930
> As of (and including) the 2025-06-30 data dump, Stack Exchange has started including watermarking/data poisoning in the data. At the time of writing, this does not appear to apply to the 2025-09-30 data dump. The format(s), the dates for affected data dumps, and by extension how the garbage data can be filtered out, are described in this community-compiled list: https://github.com/LunarWatcher/se-data-dump-transformer/blo.... If the 2025-09-30 data dump turns out to be poisoned as well, that's where an update will be added. For obvious reasons, the torrent cannot be updated once created.
It's amazing to think that in the next few years, we may have software engineers entering the workforce who don't know what StackOverflow is...
Everyone is saying LLMs did this site in, but what if we just asked all the questions already? We should be celebrating how we solved programming!
I misread the title at first and thought it was hacker news questions [comments] that were being graphed. That’s what I would be interested in seeing
Acquired in June 2021 for $1.8 billion usd. Hurts but acquirer Naspers is a prolific tech investor, its stake in TenCent is worth > $150B usd today.
The result is not surprising! Many people are now turning to LLMs with their questions instead. This explains the decline in the number of questions asked.
Signs of over-moderation and increasing toxicity on Stack Overflow became particularly evident around 2016, as reflected by the visible plateau in activity.
Many legitimate questions were closed as duplicates or marked off-topic despite being neither. Numerous high-quality answers were heavily edited to sound more "neutral", often diluting their practical value and original intent.
Some high-profile users (with reputation scores > 10,000) were reportedly incentivized by commercial employers to systematically target and downvote or flag answers that favored competing products. As a result, answers from genuine users that recommended commercial solutions based on personal experience were frequently removed altogether.
Additionally, the platform suffers from a lack of centralized authentication: each Stack Exchange subdomain still operates with its own isolated login system, which creates unnecessary friction and discourages broader user participation.
Everyone here talks about LLMs, but for me, the reason why StackOverflow became totally irrelevant is because of dedicated Discord servers and forums.
I really admire that they publicly posted this data, and hope that the platform can find a new type of pivot or draw to bring back a community.
Has AI summarization led to people either getting their answer from a search engine directly, and failing that, just giving up?
RTFM
TFMs are not a thing anymore. Most of them are merely collections of sparse random dots one might join by sheer luck only, granted no other knowledge of the system being attempted to document.
2 replies →
Looks like they sold right before the end. Wonder whether the AI deals they've struck make up for the difference
Couldn’t have happened to a meaner community
You're clearly excluding gaming communities such as DotA2.
Maybe the average question will be more "high level" now that all simple questions are answered by LLMs ?
While the decline started a decade ago in 2014 and accelerated in 2020, the huge drop since 2023 is remarkable
Good riddance.
I stopped using SO before LLM's were a thing because the community was such a pain in the ass to deal with.
Monica has the last laugh it seems
they pretend like everything is fine at HN too wouldn't surprise me looking similar in the future.
It's funny to see people's new year's resolution to learn how to code in the graph
It was a good idea ruined by the compulsively obtuse and pedantic, not unlike Reddit.
It was a good 16 year-ish run.
I've never once asked a question on there Mostly because you can't unless your account has X something-points. Which you get by answering questions.
This threw me off so much when I got started with programming. Like why are the people who have the most questions, not allowed to ask any...?
Are you sure? You can post questions even with a completely new blank account. It's comments that require some reputation, maybe you were thinking about those?
You don't need any reputation to ask questions, you only need to create an account.
And still last month one of my questions on SO got closed because it was - "too broad". I mean it was 2025 and how many very precise software engineering questions are there that any flagship models couldn't answer in seconds?
Although I had moderate popularity on SO I'm not gonna miss it; that community had always been too harsh for newcomers. They had the tiniest power, and couldn't handle that well.
Probably similar for google. My first line of search is always chatgpt
Good riddance to bad rubbish (TLDR: Questions are now almost never being asked on Stack overflow).
The most annoying example I can think of (but can’t link to, alas) is when I Googled for an answer to a technical question, and got an annoying Stack Overflow answer which didn’t answer the question, telling the person to just Google the answer.
When StackOverflow dies, who will train the LLMs?
Not surprising. It's very often a toxic, unhelpful, stubborn community. I think maybe once or twice in years of use did I ever find it genuinely welcoming and helpful. Frequently instead I thought "Why should I even bother to post this? It'll just get either downvoted, deleted, or ignored."
A death graph.
Kind of sad that they ran out of ideas how to fix SO.
End of an era. :-(
Gee...I wonder why it's almost dead (again)?
I fairly recently tried to ask a question on SO because the LLMs did not work for that domain. I’m no beginner to SO, having some 13k points from many questions and answers. I made, in my opinion, a good question, referenced my previous attempts, clearly stating my problem and what I tried to do. Almost immediately after posting I got downvoted, no comments, a close- suggestions etc. A similar thing happened the last two times I tried this too. I’m not sure what is going on over there now, but whatever that site was many years ago, it isn’t any more. It’s s shame, because it was such a great thing, but now I am disincentivized to use it because I lose points each time I tip my toes back in.
Someone turn off the lights on the way out
I wonder if google search saw a similar hit
My personal bet is that traditional search engines face a -70% usage drop at the moment.
I doubt it. If I want to ask AI a simple question I type it into Google now.
anecdotally, i personally stopped using google a lot in the last few years
LLMs are dogshit in many ways but when it comes to programming they are faster than people, respond instantaneously to further information, and can iterate until they understand the problem fully.
Bonus is that you don’t get some dipshit being snarky.
Now that StackOverflow has been killed (in part) by LLMs, how will we train future models? Will public GitHub repos be enough?
Precise troubleshooting data is getting rare, GitHub issues are the last place where it lives nowadays.
They would just use documentation. I know there is some synthesis they would lose in the training process but I’m often sending Claude through the context7 MCP to learn documentation for packages that didn’t exist, and it nearly always solves the problem for me.
The brilliance of StackOverflow was in being the place to find out how to do tricky workarounds for functionality that either wasn't documented or was buggy such that workarounds were needed to make it actually work.
Software quality is now generally a bit better than it was in 2010, but that need is ultimately still there.
1 reply →
Aren't a lot of projects using LLMs to generate documentation these days?
They pay lots of humans to train the LLMs..
>This post was not virtue signaling enough and therefore closed as duplicate.
SO had the greatest minds but the shitiest moderation
StackOverflow was immediately dead for me the day they declared that AI sellout of theirs.
Pathetic thieves, they won't even allow deleting my own answers after that. Not that it would make the models unlearn my data, of course, but I wanted to do so out of principle.
https://meta.stackexchange.com/questions/399619/our-partners...
Eternal September is finally over =)
It was impossible to ask certain programming questions. Asking there was truly last resort.
Derivative of S curve
I mean, I kinda of miss it. But man it was a hostile place for newcomers.
Only ever asked one question and I tried to answer more than a handful but never really clicked with the site.
I do wonder if it would have faired better under the original ownership before it was sold in 2021-06-02.
I not even hearing stack overflow survey for 2025
damn bro, its sad how "tradition" is gone now
edit: I know they still doing it but usually there is "viral" post,yt video etc for developer talking about it in my feed
now??? less people talk about it anymore
I’m glad it’s dead. They were super rude.
llm killed stackoverflow
This is incredible. Anyone who claims LLMs aren't useful will need to explain how come almost every programmer can solve 95% of his problems with an LLM without needing anything else. This is real usefulness right here.
EDIT: I'm not saying I'm loving what happened and what is becoming of our roles and careers, I'm just saying things have changed forever; there's still a (shrinking) minority of people who seem to not be convinced.
Maybe we had too many programmers who weren’t capable of actually solving their own problems. Maybe only one in twenty programmers were ever actually any good at their jobs.
Now the real question is...
Which AI company will acquire whats left of StackOverflow and all the years of question/answer data?
It is already acquired.
https://meta.stackexchange.com/questions/399619/our-partners...
What an incredible graph
Flag it as off topic
Another note to add here: The whole system was stupid, too! What do you mean, I can only give answers, but not comment?
While there is much more to say about SO's demise, the "interaction" on the platform was definitely not one of its strengths, either.
Comments have less visibility in moderation. This has made them spam / link farming targets in the past.
A lot of people come to Stack Overflow with the mindset that it is a forum to discuss something and have tangential discussions in the comments.
https://stackoverflow.com/tour
> This site is all about getting answers. It's not a discussion forum. There's no chit-chat.
The "no comments until you get a little bit of rep" is to try to help people realize that difference from the start.
Cool, https://news.ycombinator.com/item?id=46344599
Wonder why my submission wasn't featured and this one went to #1 immediately ... oh wait I actually know :^)!
Why?
For those who have historically wondered about or objected to "moderation" (people usually mean curation here; as the overwhelming majority of the actions they're talking about are not performed by moderators) on Stack Overflow, here's a hand-picked list of important discussions from the meta site explaining some policy basics:
https://meta.stackoverflow.com/questions/251758 Why is Stack Overflow so negative of late?
https://meta.stackoverflow.com/questions/254262 If your question was not well received, read this before you post your next question
https://meta.stackoverflow.com/questions/254358 Why the backlash against poor questions?
https://meta.stackoverflow.com/questions/254770 What is Stack Overflow’s goal?
https://meta.stackoverflow.com/questions/260263 How long should we wait for a poster to clarify a question before closing?
https://meta.stackoverflow.com/questions/261592 How much research effort is expected of Stack Overflow users?
https://meta.stackoverflow.com/questions/262446 Are we being "elitist"? Is there something wrong with that?
https://meta.stackoverflow.com/questions/262791 The rudeness on Stack Overflow is too damn high (N.B.: linked specifically for the satire in the top-voted answer)
https://meta.stackoverflow.com/questions/284236 Why is "Can someone help me?" not a useful question?
https://meta.stackoverflow.com/questions/309208 Are there questions that are too trivial to answer?
https://meta.stackoverflow.com/questions/357436 Why isn't it required to provide comments/feedback for downvotes, and why are proposals suggesting this so negatively received?
https://meta.stackoverflow.com/questions/366757 On the false dichotomy between quality and kindness
https://meta.stackoverflow.com/questions/366889 Can we make it more obvious to new users that downvotes on the main site are not insults and in fact can help them help themselves?
https://meta.stackoverflow.com/questions/368072 Comments asking for clarification or an MCVE are not rude/abusive
https://meta.stackoverflow.com/questions/370792 Is this really what we should consider "unwelcoming"?
https://meta.stackoverflow.com/questions/417476 Question Close Reasons - Definitions and Guidance
https://meta.stackoverflow.com/questions/429808 Why should I help close "bad" questions that I think are valid, instead of helping the OP with an answer? (fd: my self-answered Q&A)
Note that IDs are in chronological order. The rate of new meta.stackoverflow.com posts fell off dramatically at some point because of the formation of a network-wide meta.stackexchange.com. The earliest entries listed here are from 2014.
Some of the comments in these links are hilariously elitist. They are actively embracing a hostile environment, especially towards newcomers, but how do they expect to grow and maintain a community when they are scaring users away?
Holy shit.
[dead]
Stackoverflow bureaucracy and rule mongering are insane. I recommend participation just to behold the natives in their biom. Its like a small european union laser focused on making asking snd answering a question the largest pain point of a site that is mainly about asking and answering questions.
Since the trend must go on, we expect StackExchange to now offer answers, and the user responses need to be questions. We could even make a quiz game show out of that! /s
[dead]
[dead]
[dead]
it died
The most toxic, degrading, and insulting forum for people. My questions, as well as my answers, always got poisonous criticism. Good.