This is a large decline. It does not correlate well with the rise of LLMs, which have shifted my own habits away from Stack Overflow recently. Looking through the graphs, the most dramatic shift was the large drop in new visits in the first half of last year, by two thirds or more. That seems late to be a Covid effect. There were some political controversies, but it would surprise me if they had such a large effect.
I wonder if Google started de-ranking them in search results, or changed something similar that caused people to click through to Stack Overflow less frequently.
It is also possible there's a methodology issue. There's no source for where this data comes from.
I just googled a Microsoft tech question and first was a 3rd party provider plugin, then official MS Learn, then "people also asked...", and 4th was the SO block.
I scrolled down a lot and didn't see any scraper site links.
You don't necessarily need to trust them for them to be useful.
I can often evaluate a Copilot autocompletion (check that code looks right at first glance, check that it compiles, hover over method+type signatures to see their docs, run the code) in less time than it would take me to find+read a Stack Overflow answer.
No, no one sane would. Just like no one sane would trust random SO authors with any authority either. You still have to verify the information somehow, but at least in the LLM case I instantly get replies 24/7, and it's been reading more papers than any of the authors on SO.
ChatGPT has gotten quite a bit more accurate over the last few months. It does seem to have lost some creativity in the process, but it's much more true to the training data now.
IMHO people have tired of the smug dismissive assholes that dominate SO and its its insipid gamification and they have, gradually, found alternatives.
For some, it's more welcoming and knowledgeable communities in github issues+discussion.
For others it's special-purpose, interactive forums that provide more guidance and non-hostile support for users of certain platforms/tools (eg Posit Community).
For many, copilot has been fulfilling that need, Does it always give "the correct" answer? No. Does it provide a sketch of a solution that gets you half-way there when dealing with tedious humdrum stuff that you just forgot because it's so boring? YES.
For yet others it's just plain old reddit where you can ask a question and whether it gets smacked down or not depends on how cool the community happens to be.
I was promoted to one of those "mod" type people who reviews new answers, first posts, etc, a few years back.
I can tell you that I quit in the last 6 months. All new answers are answers on questions from first time posters, answering questions that have good solid answers from 8+ years ago (mostly 10 years), and it is pretty obvious, they are just doing it to get points. The question/answer is not something that has changed in the last 8+ years, they are just doing it to get reputation or whatever.
Most good solid questions...I just do not see too many today. I am not saying this is the reason for the decline, but there is a deluge of posts answering very old questions with a slight modification of an old answer. I quit putting effort into it, as it was just non-productive.
I recall that about ten years ago there were reports that edits to Wikipedia were slowing precipitously. People wondered if this meant the imminent death of Wikipedia. With the benefit of hindsight, it seems more plausible that Wikipedia had merely shifted from its adolescent phase to its mature phase, and that all the low-hanging fruit had simply been harvested. If we look at the stats, Wikipedia's edit velocity maxed out in early 2007 (37 days per 10M edits), then fell to half that in late 2014 (73 days per 10M edits), and has actually increased a fair bit since then: https://en.wikipedia.org/wiki/Wikipedia:Time_Between_Edits
It's entirely possible that the same is happening to SO.
I wonder if this might be partly because they have not made an effort to make their software be more intelligent about surfacing answers to questions as well as evaluating answers. I would think that, at some point, it isn't necessarily the fault of the question authors or those responding.
One of the things I say with some frequency is that all failures are engineering problems. Blaming the users/customers, in the end, fails to take advantage of an opportunity to improve the product, learn from mistakes and improve process.
While I understand it's a lot of work, and some (most?) might be posting low quality answers, my main problem with stack overflow has been exactly that over time, the top answer becomes stale.
The question is still relevant, but the best practice or library, platform whatever changed... That is a problem the stack overflow model has a hard time adapting too.. and it just gets worst over time..
From my purview, and again I am a tiny data point of one, the new answers to old questions are really:
Original accepted upvoted to +20 answer:
"Take Y and drive it into X, grep output for Z and place in AZ for processing with the -kombucha switch"
Answer posted yesterday by "NewUserXX123", a first post:
"Take A and drive it into B, grep output for C and place it in DD for processing with the -kombucha switch"
It gets pretty tiring responding with a custom comment and downvoting, only to have "NewUserXX123" then cuss you out in a DM about how you are a total @!#$%()&!@$%.
That's going to depend on your domain, because most fronted d questions I come across end in "there is a jquery plugin for that" which don't help modern sites without jquery.
> answering questions that have good solid answers from 8+ years ago (mostly 10 years),
I hit upon those useless answers all the time. I have no need to know about solutions that might have worked a decade ago and are obsolete. Stack Overflow is useless now.
> and it is pretty obvious, they are just doing it to get points
Gamifying reputation was one of the worst things we've done. I was part of this in previous positions and after a while it became apparent that it brought out the worst in people. There are better ways of incentivizing usage rather than stupid internet points.
Agreed and I don't know this for sure but the points lust, it seems to me, is not limited to questioners and answers, some of the behaviour I've seen from mods seems to be driven by that too.
Just to repeat, I don't know that because I don't know what gets you points but the behaviour is hard to understand otherwise.
I had a question where I had finished with a one word sentence "Thanks." . A mod had removed that part of the question, when I looked at their record they had an awful lot of edits which consisted of similar, tiny, non-substantive changes. It did make me wonder about their motivation.
> "I had a question where I had finished with a one word sentence "Thanks." . A mod had removed that part of the question, when I looked at their record they had an awful lot of edits which consisted of similar, tiny, non-substantive changes. It did make me wonder about their motivation."
"Do not use signature, taglines, greetings, thanks, or other chitchat.
"Every post you make is already “signed” with your standard user card, which links directly back to your user page. Your user page belongs to you, so fill it with information about your interests, links to stuff you’ve worked on, or whatever else you like!"
"Thanks and other statements of appreciation are unnecessary, and, like other chitchat, should not be included."
"If you use signatures, taglines, greetings, thanks, or other chitchat, they will be removed to reduce noise in the questions and answers."
----
Removing it doesn't get the editor any points, they are spending their time cleaning up your question to site standards for no reward at all.
What would you think if you clicked on a Wikipedia article about Mozart and instead of the current introduction which says:
> Wolfgang Amadeus Mozart[a][b] (27 January 1756 – 5 December 1791) was a prolific and influential composer of the Classical period. Despite his short life, his rapid pace of composition resulted in more than 800 works of virtually every genre of his time.
it opened with:
> Hello everyone, I studied classical music in college but have been out of the scene for a while and now I'm getting back in, I was wondering about the history of Mozart and I remember that his middle name was Allen or Almond or something? Can anyone help? Thanks for any information. || Hi, I learned about him in middle school and we all joked that his middle name was Armadillo but I think that's wrong, haha. || Greets all, it's in Olivier Hallengrunsch's Classical Composers as 'Amadeus', and that's well regarded. It also says he lived (27 January 1756 – 5 December 1791). I think we can all agree he was prolific, right? Thanks and regards, Jason [xxKiller; AMD Ryzen2 32GB RAM 2TB Western Digital SSD; BMW 330 2l aircooled] || Why's nobody mentioning how short that life was, smh || Good evening all and sundry, m'lady (tips hat), forsooth would anyone speak to how many works he is believed to have composed, all considered? Methinks such knowledge would be a most hearty addition to this esteemed gentleman's biography - Martin, [Fort Lauderdale TX Ren. Faire organiser 1997-1997] || etc.
StackOverflow isn't a forum, it's a collaborative reference work. Meta-chat would be edited out of a wikipedia page and goes on a separate 'talk' page (equivalent: meta stackexchanges or the stackexchange chat). What if you then went to the Wikipedia talk page and said "Is overzealous moderation of questions ruining Wikipedia? I want to be able to edit questions and greetings into pages but there are hoards of awful literal-minded jobsworths cruising the site just looking for the slightest reason to edit a spelling or grammar mistake or revert my changes. It just leaves me feeling unwelcome"?
Why would you expect to feel welcome when you're spoiling what others are trying to build up and insulting them for caring??
Stackoverflow has https://meta.stackoverflow.com/ and https://chat.stackoverflow.com/ (you see it when the comments on a question or answer go on too long, there's an automatic "comments are not for extended discussion, take it to chat" reply, or you can invite people to a chatroom about a question).
I hear your point and thank you for your service as one. That said, how much do we think this is more attributed to the existence of coding AIs like Co-pilot or ChatGPT. I hardly use stack overflow anymore thanks to OpenAI
Why was there such a heavy handed approach to moderation in the first place though? Why not let the community ask the same question over and over again, and let the answers naturally shift with the time?
I'm just not sure what value there was to have a team of people volunteer their time to making sure there were absolutely zero duplicated questions/answers? What problem is that solving? If anything, it is adding problems by not letting the content naturally evolve in time.
If people want to assess the quality of answers, it's better to have multiple data points, and SO could have invested those resources in algorithmically linking similar questions and making it easy to navigate by both answer reputation and time.
Along similarly lines, it seems best to tackle the people trying to game the system algorithmically as well - if content is word for word duplicate, that's a problem which can be solved by computers instead of people (similar text is a solved problem).
It seems the only use case for moderators on SO is for removing truly inappropriate content - it's wild to me that moderators were spending significant amounts of time actually removing technical questions and answers.
It reminds me a bit of reddit moderation, where reddit communities enforce these non-sensical rules and by extension require hugely heavy handed moderation to 'curate' their communities. Like, the headphones subreddit disallows pictures of headphones in boxes. Why? If the community is interested in headphones, what's the difference if it's a box or not? It's not like if you take it out of the box the headphones look different than any picture you can find on the internet.
Seems like many of the problems of moderation are artificial rules endlessly being enforced by real people which ends up just being pseudo 'make busy' work.
When it comes to software, it’s even less sensical to have one archival correct answer, given that software evolves over time, and what was an acceptable or necessary workaround 5-10 years ago could be a complete anti pattern today.
It would more sense to establish lineage. Lock older questions from more answers after some point in time, and instead of closing and linking to original, do it the other way. Keep new questions open and link to past variations. Then in the old threads indicate newer guidance may exist and link forward.
They talked about it in the Stack Exchange podcast years ago.
IIRC, the vision of the perfect question was one with a canonical answer. They didn’t want to be Quora. I think that concept made and makes a lot of sense, but doesn’t capture the “meta” issues surrounding how you litigate the form of the question, especially as questions get more nuanced.
I couldn't figure out how to gain any points other than answering my own question, but I needed points to contribute in any way. It's like SO is set up to be difficult to use. I can't imagine why anyone keeps using it.
For me, a few months ago, when I had a question, I used to google/bing it and try to find my solution between 2-3 different SO answers (even the questions themselves were often helpful).
Now, AI kinda replaced for me, it's copilot first, then ChatGPT if copilot is not enough, then google/bing if ChatGPT is not enough and maybe finally SO.
Conservative figure from my side would be 15-25% less stackoverflow in a regular day.
Not sure about the statistics, but an anecdote for me: I had a specific programming issue at work I need to solve, and I tried googling for answers in Stack Overflow. There are some answers but ultimately none of what I found can actually solve my issue. Then I ask ChatGPT to suggest a solution, and in about 5 seconds it gave me an answer. Not perfect, but with little editing it is solved.
Now I'm sure that ChatGPT probably scraped StackOverflow, perhaps using the very same answers I got from StackOverflow, but combined with other answers from other sources resulting in the instant answer. That does not necessary mean AI will replace StackOverflow, it just means that people wouldn't ask redundant questions in StackOverflow anymore, just questions that AI can't solve.
Where were these data sourced from? As far as I know, StackOverflow does not publicize internal analytics with such granularity. If these figures are real, are they leaked?
We can also see that lately there are more questions than answers, which shows that most experts are no longer that active, or that there are more beginners and fewer experts overall.
Stack Overflow used to release their data archives quarterly on BigQuery. Looking at the BQ datasets, they were last updated Nov 2022, which doesn't have the latest 2023 info in the submission.
Thanks for sharing, good to see alternative options popping up. My wish is that the Stack Exchange dataset could one day be provided as a streaming parquet or arrow table, as underfunded grads and post-grads could then more easily/selectively sample the datasets (similar to how Huggingface provides some of its datasets)[1][2].
The Hugginface repo unfortunately prefilters some of the tables/rows according to some criteria, making it less usable for general analytical queries that the BQ or SEDE datasets enable. If anyone knows of an 'XML-streaming' solution that directly samples from the Internet Archive's data dumps, I am all ears.
I think the problem is google losing the fight with spammers.
It's being a while for me that I have to put "stackoverflow" in the search query to avoid sites with scraped content
especially YouTube links. sadly, it would not surprise me if these people are earning a decent enough money from ads to make it worth their while to be "content creators" solely from search results from Googs
It is both infuriating and sad that Google can‘t figure out a way to compensate for this SEO spam. Is there an easier problem than doing it for SO (and yes, coding is a big enough problem for Google imho to be worth investing a little here).
Which means they aren't applying any sort of primacy to the information.
If three segments of the internet think the same piece of information is relevant, that should affect the score of all 3 copies, not just the largest segment.
I'm not sure I'm reading you right--you're suggesting it should work this way?
When content republished on some bizarre/sketchy/unaccountable ~adfarm outranks the site where it first appeared, users of Google's search service end up at higher risk of getting phished or infected with malware.
Is there some benefit you see here that outweighs this downside risk?
Applying SEO to a copy of someone else's content gets you highly ranked on Google. I'm saying that at this point Google is doing enough processing that they should be able to detect duplicates after a fashion, and weight the oldest copy more heavily than duplicates.
That's my guess too; I'm sure Google drives the overwhelming majority of SO traffic.
A few years ago, my programming-related queries would hit Stack Overflow as the first or second result. Now it's very frequently spammy garbage in the top 2-3 slots.
What kind of spam do you get when searching something specific and technical? Who is trying to SEO their way to the top for "how to set redis max memory"? A lot of comments here saying the spam is beating out SO, but what spam and from who and why??
What’s interesting is that Google was known for how hard it was to figure out the Google algorithm.
Remember, when people were hired because they knew the secret sauce on how to get the best Google ranking. Google experts?
Well, it turns out that the person at Google that was responsible for keeping the algorithm fresh and the search results fresh retired and everything went to shit when they left.
Actually, I’m betting that person did leave the company, but the real damage happened when someone came along and convinced everyone they knew the real trick to better search results and we have the shit that is now Google. Nice work new guy! Let me rephrase that. Nice work to the guy that thinks they are smarter than everyone else and still thinks their approach is the best, yet evidence to the contrary.
Really sounds as if your made up story is deeply rooted in your own experience. I am sorry if something like a new guy taking your position and claiming to be smarter has happened to you but creating imaginary stories is not quite what this comment section needs and you'd probably be better off dealing with this in a different way
Not at all. Nothing personal. Although, it looks like you are the one self-projecting here.
It’s simply how times change and people with it. Knowledge is lost when people move on and the reasons why certain decisions were made are not transferred.
At any rate, I imagine people at Google are trying to figure out why there is such a negative opinion on their search results lately.
Matt Cutts was instrumental in community outreach and helping SEO differentiate from spam. When he left, Search pivoted to stuff like using Twitter data and lifting content directly from websites into results. While it’s probably hard to attribute all the changes to one person, Matt Cutts made a huge impact on the product.
Well if they served up a high quality site it you'd just go there and might not have ads even. Where the dozen SEO garbage sites they do serve up are all hosting ads google gets a cut of.
That's a very short sighted business strategy if true. Simply liquidating their reputation. Those junk AI results have certainly led to me using Google less.
> That's a very short sighted business strategy if true. Simply liquidating their reputation.
Why should they care? They're too big to fail…
Google controls almost the whole end-user realm through Chrome & clones, and Android, the dominating end-user OS by a wide margin.
At the same time end-users are completely helpless and can't do anything against Googles liking because they don't understand anything about IT tech.
Computers are black magic to most people so they're trapped. This never changed! Especially millennials and gen-z are completely clueless as they didn't had the chance to use personal computers ever, where you had at lest some control over the device and needed to know at least some basics about its inner working. All the younger people know are the tightly sealed black-box devices you don't have any control over, called mobiles, which are fully operated by big-tech. Google search + Android apps are "the internet" for most people. They mostly don't even know there is something else beyond that, so Google can do whatever they want, and this will have exactly zero consequences for them by now.
Google's move regarding rolling out "browser DRM", the next "trusted computing" initiative, regardless of what anybody thinks about is is very telling.
Now they will violently reap the fruits of their monopoly, and likely nobody will be able to stop them in the next decade. People where warned about the consequences of this monopoly for many many years. Nobody cared. Now it's payout day for Google.
When do you think millennials were born? The very youngest millennials were in their tweens when the first iPhone came out, and the oldest were pushing 30. They definitely experienced pre-smartphone computing. In fact, it's probably the defining characteristic of the generation: millennials grew up with modern computing, but before the smartphone. Gen Z grew up in a world were smartphones were ubiquitous.
I think we're rapidly approaching a point where any content that come with ads is suspect. The fact that only Wikipedia has managed to largely escape deterioration (or as some call it, "enshitification") is testament to this. A search engine that can selectively search non-sponsored content or soft-paywalled content would be potentially quite popular. However, monetising such a service without ads will be a challenge.
Google is not "losing" any fight. Google is deliberately letting spam thrive because that spam may contain Google Ads/analytics and increases engagement on the SERP page as people who click on the spam go back to try something else (potentially one of the sponsored results). All these contribute to Google's bottom-line.
Problem is that in addition to people whose salary depends on it, there seems to be plenty of people out here defending Google and spreading misinformation despite having no obvious profit motive.
I agree, but I'm afraid the downvotes are inevitable. SO for a long time had by a very long shot the best signal-to-noise ratio of any free-as-in-beer Q&A channel. Basically any live comms channels (IRC, Matrix, and so on) disincentivise clear, thoughtful responses in favour of streams of consciousness. And forums is where information goes to die, as every question turns into a huge back-and-forth until the question is sufficiently refined to receive a useful answer, which is usually extremely specific ("run this command") rather than generally applicable ("you need to frobnicate the foon, for example by running this command").
Unfortunately anything popular is a target, and once they stopped trying to innovate to keep ahead of systems gaming they were going to lose eventually. It's amazing it took this long.
Interesting that this seems to have started around Spring 2021 when posts + votes started tailing off, followed by traffic starting to decline around spring of 2022.
I can think of a few theories that I don't think hold water:
1. The rise of ChatGPT to answer many questions that StackOverflow would previously have been used for. This seems unlikely, since the timing doesn't really work out.
2. The perennial complaints about StackOverflow's culture of closing everything as duplicates or offtopic. This seems unlikely as well, since those complaints have been common for a decade or more.
3. The prevalence of SEO-optimized scrape sites - the ones that pop up with a "blog post" merely reposting a stack overflow question + answer in a different font. I've seen these for a while, and anecdotally they feel more common that they used to, but I couldn't give any real timeline for that vague feeling.
4. StackOverflow internal politics? I've seen the occasional stack-overflow meta thread pop up periodically on HN or social media, but I don't recall anything earth-shattering recently.
5. Most questions have good answers now and there's less need for new ones. I'd have bought that answer 10 or so years ago when StackOverflow's pile of questions + answers reached maturity. I don't think it suddenly hit some sort of answer saturation point in 2021.
My guess is that it's a slow shift in the culture of the StackOverflow userbase:
- Being a top answerer confers some cachet and makes you more employable in some places
- People notice this and start looking for the most effective ways to become a top answerer
- The most effective way is fast, low-effort answers
- There's been a rise in such low-effort answers over the last 5 years or so
- As a result, the cachet of being an StackOverflow top answerer is a lot lower
- The really good, deep, technical answerers (as well as the mods) are leaving as that cachet goes away
- Post quality starts dropping around 2021 and views start declining as people react to that in 2022.
As much as SEO experts would love to do that for their clients, I don't see any way they could've realistically crossed that moat. If you had a programming question, and you went to Google to answer it and found some answers - would you choose Stack Overflow, or PythonIsSuperFun.com? I think the behaviour of Google searching for programming issues decreased overall, and I don't think Stack Overflow suffered more than any other site. Disclosure: I've nearly abandoned SO for ChatGPT myself.
> would you choose Stack Overflow, or PythonIsSuperFun.com
I would choose first result - people are lazy. The problem comes and goes but it is true, spam sites, with content copied from SO, are ranking higher than SO itself. And it is often hard to tell you are viewing a copy on first glance, the layouts are usually more like a forum.
If I were building a Q&A site that genuinely wanted to encourage high-quality answers, I'd implement something like a 1-hour window where all answers are invisible immediately after the question is posted. This is to give people time to work on a good quality answer, without racing to be the first and gain those precious early upvotes. The UI could still indicate how many other people have answered / are answering, and if answer volume became a problem you could perhaps block new answers during that window after the first 20 or so have landed. When the hour is over, answers are displayed in random order and the existing site mechanics around upvotes etc kick in.
It feels like that could potentially address the problem of low-effort answers killing off the good ones.
Comment: "You shouldn't do A, it's better to do B"
Closed, duplicate of "How to do C".
Yeah sure, in the beginning, there were many more basic questions (how to increase number by 1, how to get division remainder, how to check if file exists,...), and if you're coding in Perl, you can still find all the answers... if you're working in python, you'll find answers for python2, some for python3, some specific to python2.6, etc... if you ask again, it'll be closed as a duplicate.
I know it's anecdotal, but after a few bad experiences, people just decide not to use that specific site anymore.
Echos my experience perfectly. It's actually impossible to ask a question and get an answer, also impossible to ask a clarifying question on an existing question/answer.
That makes it much, much less useful than it otherwise could be.
Perl is a good example because it's stagnant. Googling Python or Java? If the result is up to date, it's a low quality content farm. If it isn't from a content farm, it's a page from 2008 that is no longer relevant at all!
My experience was that there was a long period where your complaint was quite true.
I believe that they have heard this complaint enough and have lightened up a little. There are still more and stricter rules than most sites, but the obsession with pruning duplicates seems to have cooled.
And my opinion as someone with high reputation is that letting in silly questions has ruined the site and made me stop using it. SO is not a support forum, it's not your teacher, it's a place for interesting knowledge sharing and if the knowledge is drowned in noise, well it's just noise then.
Stupid people don't realize they are stupid. When they can't understand a question they conclude there's something wrong with the question, not with them. And since they believe the question is silly, they happily close-vote the question.
This way they close questions which require specialized knowledge, experience, or deep understanding of the subject. Precisely the kind of questions I consider interesting to answer.
Sometimes I managed to undelete such questions, but SO made it easy to delete and hard to undelete.
The first question for example is what would be a low-quality question in SO, and could easily have been a homework assignment. It doesn't show any effort from your side, you just pose the question and expect other people to answer it fully, without explaining what you attempted, why it worked/didn't work, where you got stuck, etc.
> It doesn't show any effort from your side, you just pose the question
You probably assumed I asked that question? If so, the assumption was incorrect, I answered it.
> without explaining what you attempted, why it worked/didn't work, where you got stuck
I don't think any of that is possible to do.
The question, and my answer, are too simple to decompose into smaller parts. Not enough runway to start and get stuck.
> expect other people to answer it fully
My expectation was rather different. I wrote my answer on Jul 16, 2019, and expected it to stay there.
Instead, on the next day some people have decided the question was bad, and closed it. Then on the next month, some other people have deleted it, along with my answer.
To close it, 5 people clicked once/each. To undelete it, I spent quite a few hours. Sadly, that's not a rare exception: moderation on SO is horrible.
Ah sorry, if you answered it it's totally different, sorry for assuming the question was yours! Some of my best answers are also on "simple questions".
The question is asking about the best way of doing X, which has a clear substep of "any way of doing X". Not only that, it's asking about a code solution, not just maths/combinatronics. Not a single line of code in a question about finding the optimal (not any) solution for doing some algebra in code is, IMHO, reason enough to close it.
Edit: for example, I'm not asking and so have no stakes on the question but can already try to think about brute-forcing it, which already shows more effort than the author SHOWS at attempting a solution.
To be fair, English is not my native language. But when I see a question “what’s the best way of doing X”, when it doesn’t have any criteria for what’s the best would be, and no other ways of doing X in the question, I consider “what’s the best way” part a redundant figure of speech. I view such questions an equivalent of “what’s a good enough way of doing X”.
> it's asking about a code solution, not just maths/combinatronics
Please read this: https://stackoverflow.com/help/on-topic According to that article, questions about math which don’t imply a code solution are offtopic on stackoverflow.com. They should be closed, and possibly moved to other stachexchange sites. According to that article, the OP’s question is good. The question was about a specific programming problem, and is a practical, answerable problem unique to software development.
Again, they are asking how to solve a math problem, in code. That's a two big-step problem, no attempt to solve it on their own. Big problems:
- Does not show any attempt or willingness to try to solve the problem first on their own.
- Does not even give any indication of where the problem comes from, why it might be interesting, etc., it's just a "how to calculate X?", which could easily be a homework problem.
- It is about finding a (possibly) mathematical solution, and then implement it in C++. Two very big and different problem, asking the audience to do them both. Again no attempt to fix either of these two problems on their own.
- I'll concede the optimal thing might be a language issue.
> I’m not sure that’s actually possible to do.
But that's not my point, my point is that I already showed more willingness to try to solve this problem than OP. And THAT is a big problem. It's not on topic about any of those points, in fact if you remove the bit where OP is asking us to give them the full solution in C++ it could be a good question for the Mathematics SE!
> attempt or willingness to try to solve the problem first on their own… indication of where the problem comes from, why it might be interesting
None of that is required to ask questions on stackoverflow. For details, read “How do I ask a good question?” and “What types of questions should I avoid asking?” help articles. You’re inventing arbitrary restrictions.
Another thing is, “why it might be interesting” is subjective. Personally, I found the question interesting, that’s why I have answered it. You probably think otherwise, but note it only takes 3-5 votes to kill the question. Any question at all is guaranteed to have at least 3-5 people on that site who find it uninteresting, opinion-based, need more focus, duplicate, etc.
> Two very big and different problem, asking the audience to do them both.
Two big problems don’t have solutions which can be both explained in 3 short sentences. As you can see from my answer, the problem formulated in that question has such solution.
> I already showed more willingness to try to solve this problem than OP
You have not. However, you have demonstrated willingness to delete interesting questions based on arbitrary and subjective criteria, despite the question is perfectly in line with the stackoverflow guidelines. Which BTW is very on-topic, because I think that’s the main reason for the fall of SO being discussed here.
> SO is not a support forum, it's not your teacher
This, to me, is the mistake that SO made. Developers helping developers is the engine that runs the site, that's why people come.
The body of interesting knowledge is an emergent property of the support forum/peer-to-peer teacher.
Eventually they tried to put the cart before the horse and traffic has dropped.
It's OK that you're done answering "how do I change font color with jQuery" for the thousandth time and are only interested in the occasional very interesting question, because there are people behind you who do want to answer that question. That will help them grow to get where you are.
If we don't allow new generations of users to go through that process we went through, then StackOverflow has an expiration date.
I found the people who took on Rust questions to be actively hostile on SO when I started learning Rust. The Reddit community was much better, so I’ve tended to hang out there for Rust q&a. Most of my time in the Stack Overflow world these days is mostly in the more specialized StackExchange sites. tex.se tends to be pretty good, as does latin.se japanese.se seems dedicated to stomping out anything that is remotely a translation question.
Reminds me of being on IRC in 97 asking questions about Assembly. When someone did finally respond it was "Did you read the manual yet"? It's ironic that SO became what it was replacing.
But did you even _try_ to read the manual? Did you lift a finger to figure out what the issue is? Some of the best answers are ones that directly quote the manual and then add clarity since manual language can be terse.
Yes I showed him the exact page that talked about my issue and he never responded. Still a silly response. He could have answered it in 30 seconds and saved me 10 hours.
I think ChatGPT is amazing foe this because you will rarely have to deal with anyone who has this type of attitude when you need help in the future
Honestly still much better than asking a forum. The hostility from forum regulars always seems more intense--less masked as a "terse down to brass tacks" attitude and more-so blatant laziness, misplaced frustration, etc. Also, forum posters tend to waste much more time as every other post is likely to be a joke or a tangent idea, as if it were casual dialogue in a chatroom.
The same mechanisms that make SO kind of brutal have also helped revolutionize asynchronous online Q&A.
Interestingly, just a few weeks before launching ChatGPT, ZDNet magazine made an interview [0] with StackOverflow’s CEO to discuss how the site became the world's most popular programming site. Unfortunately, this doesn’t seem to be the case anymore.
I wonder how much traffic has been siphoned away from those copy-cats that mirror SO? Out of principle alone, I will immediately hit the back button and look for the genuine article, but I could believe many do not.
The other aspect that grinds my gears is the closing of duplicate question. Fine in principle, except when the original was answered 10 years ago and all of the answers are jQuery.
All, but all, my recent questions there were flagged and deleted immediately. My account there is 11 year old and has over 11k reputation - not much but not nothing. The site, for the past 2-3 years seems moderated by nazis and bots. Dunno if my timeline matches these graphs but I noticed the decline quite early.
Even though StackOverflow in the common use case has been taken over by ChatGPT, I sincerely hope it keeps operating, stays strict (even if it causes collateral) and keeps ban on LLM-generated content.
The wheels of this kind of stuff turn slowly, but obviously ChatGPT was trained partly with data only gainable from a healthy StackOverflow-kind of site with users actively asking unique questions and enough people answering those unique questions with well-though-out answers. The shittyfuture outcome is that StackOverflow goes out of business and LLM's stagnate on this front, while being capable of answering in fuwwy speawk when prwompted, would still be limited to / biased towards older versions of libraries, software, languages, tools etc.
Hopefully so. As I mentioned in my other comment somewhere here, my optimistic prediction would be that StackOverflow will eventually still keep operating, but only by questions that can't be solved by AI, so hopefully leading into a more high quality discussions.
It has for me, and several developer friends, and considering the fame ChatGPT has gotten, and that StackOverflow's fall has accelerated, it's obvious the milkshake's migrating. Not all of it of course, as I stated, "for the common use case"
It's painfully slow. I can Google the question, click one of the top results, skip to the relevant part and read it faster than GPT can generate two sentences. You also have to build an elaborate prompt instead of throwing two/three keywords into it.
It doesn't help that GPT is insistent on replying in the three paragraph format, meaning that the first 30-40 words it creates are just trash to be ignored.
I found it useful once - when I had to write an essay about ISO 27001 for college and just wanted it to go away. Took what it generated and spent 20 minutes editing it to look closer to my style. For real work it isn't as useful.
I find it incredible you find a LLM slower and less full of useless chitchat about a question than stack overflow.
I don’t even open SO anymore; if it has a direct answer to your question, the LLM almost certainly does too; and asking new questions on SO is basically impossible.
If you manage to survive the gauntlet of “too specific, already answered, not general interest, arbitrary moderator activity”, the chances of getting an answer that answers your question can take forever; most likely you’ll get a stupid answer that doesn’t answer it, upvoted by idiots who don’t understand that it not an answer the the actual question, and, ultimately, because it “already has an answer”, ignored, never to receive an answer.
Maybe one day, a passing savant will answer in a comment.
…and yet, you find it faster and more reliable?
You, and I, have had different experiences on stack overflow in the last two years.
I think maybe you haven't been using GPT4 (the one where you have to pay money). Or else you're coming at it with a very strong prior, or you're not asking it about software engineering questions, or you're not phrasing your questions carefully. GPT4 is demonstrably extremely useful for technical questions in the realm of software engineering, and in addition to surfacing useful answers, it (obviously) presents a completely unprecedented conversational interface.
> I can Google the question, click one of the top results, skip to the relevant part and read it faster than GPT can generate two sentences.
Ironically, this is why people like me prefer LLMs (when they're accurate). With Google, about 50% of the times the top SO hit is not answering my question. So I have to click 5-10 SO links, parse each one to see if:
1. The question being asked is relevant to my problem.
2. The answer actually answers it.
I may be able to do it quickly, but it is a tedious burden on my brain. While GPT doesn't always work, the nice thing about it is that when it does work, it has taken care of this burden for me.
Also, GPT's pretty much memorized a lot of the answers. I once asked it an obscure question involving openpyxl. It gave a perfectly working answer. I wondered: Did it reason it and generate the code, or is there a SO post with the exact same answer? So I Googled it, and sure enough, there was an SO question with the same code!
Except GPT's solution was superior in one tiny respect: The SO answer had some profanity in the code (in a commented line). GPT removed the profanity :-)
Can you give an example of a technical software question where you found it wasn't helpful? I'll see if I can get a good answer and post the permalink for you. I suspect you're not phrasing your questions well.
Sure, but that doesn’t mean it’s not true, and for many of us its truth is prima facie because it’s true of both our own usage and the people we work with and talk to.
I have 110% replaced it with ChatGPT. Perhaps SO would still have a chance back in its glory days but there's no comparison to having a direct, specific, instant answer vs having to fight against SEO or moderators for hours.
I always read here ChatGPT is amazing. Can you give a link on how to use it? Every time I tried to google it returns lots of different results and when I tried it it:s not even usable for basic things I want. Is the ChatGPT you:re talking about on their website? Do I have to pay for it?
You're mostly right in your experience. I have spent quite a bit of time trying to get ChatGPT to be a worthwhile piece of my workflow, and I guess sometimes it is, but most of the time the basic code or config or content I try to generate, it gets very fundamental things incorrect. It feels like it's mostly just hype these days.
Can you please specify whether you use (paid) GPT-4? Would you kindly provide links to a few examples of very fundamental things incorrect?
My experience - the free version made up a lot of things but still felt very useful - enough to want to upgrade to the paid version. With the paid version, I notice very rarely that it hallucinates. It does make errors but it can correct them when I provide feedback. It is possible that I just do not notice the errors you would notice, it is also possible that we use it differently. I would like to know.
> Can you please specify whether you use (paid) GPT-4?
Paid.
> Would you kindly provide links to a few examples of very fundamental things incorrect?
No, definitely not.
> I notice very rarely that it hallucinates.
Unsure of what "hallucinates" means in this case. Some examples of things I've used it for: docker configuration, small blocks of code, generating a cover letter, proofreading a document, YAML validation, questions about various software SDKs. The outcome is usually somewhere on the spectrum of "not even close/not even valid output" to "kind of close but not close enough to warrant a paid service". When I ask for a simple paragraph and I get a response that isn't grammatically correct/doesn't include punctuation, I'm not sure what I'm paying for.
>> Unsure of what "hallucinates" means in this case
The term "hallucinations" is now commonly used for instances of AI making stuff up - like when I asked ChatGPT (before I had paid account) to recommend 5 books about a certain topic and two of the recommended books looked totally plausible, but when I tried to find them, I discovered there are no such books. This is where I see a big difference between GPT-3.5 and GPT-4.
>> I get a response that isn't grammatically correct/doesn't include punctuation
What punctuation? If you mean stuff like commas separating complex sentences, my English is definitely not good enough to spot that. But your mention of punctuation reminded me of problems that ChatGPT has with my native language... any chance you are using ChatGPT in a language other than English?
Can you give an example of a technical software question where you found it wasn't helpful? I'll see if I can get a good answer and post the permalink for you. I suspect you're not phrasing your questions well.
> Can you give a link on how to use it? Every time I tried to google it returns lots of different results and when I tried it it:s not even usable for basic things I want.
Here is an example of using it to write simple powershell scripts:
You need to pay for it if you want access to the latest version of the model, along with some beta features like plugins. Plugins are extremely useful and it is worth paying just to get access to them. For instance, you need to have a certain plugin to get it to read links.
In that JSON example you're honestly losing more time with ChatGPT that doing it yourself. It seems more like mentoring a junior than a helpful assistant. Most of my interactions with it have been this way.
I knew/know very little about 3D printers and the fields didn't mean much to me so I didn't want to have to research every one of them. It wouldn't have been difficult, just tedious.
I haven't. Because (free, as in free beer) chatGPT is extremely slow, I have to make a rather extensive proompt to get the result I want to, and then I still have to debug most code.
That's not very convenient, atleast for now. I got so used to search engines by now, that it only takes a few keywords to get the expected result. Be it a SO-answer or a documentation page. And as people have mentioned, chatGPT was learned on the stuff that's on the internet, so if there will never be any new stuff, because people just use AI, then it will not learn and won't answer your new questions. For some edge cases I might try AI here or there, but usually it's not for me.
Hell, there comes even an example to my mind. I recently just asked chatGPT what a single-issue 5 stage pipeline on a CPU actually means. I wanted to know if, especially, the "single-issue" meant that only one instruction is present in the pipeline at a time, or if a new one gets shifted in on every clock cycle (if there is no hazard). It just couldn't answer it straight-forward. It was also kinda hard to find the exact definition on the internet. I found it in a book from the 90s which was chilling in my book shelf (Computer architecture and parallel processing by Kai Hwang). Hint: Single-issue just means that only one instruction can be in one stage at a time, but still multiple get processed inside the pipeline. The keyword is 'underpipelined'
Yes, someone tested it on GPT-4 for me too and that actually gave a quite decent reply. Still, there are always some cases somewhere where it messes up.
I'll just keep an eye on AI progress, but will probably not make it my goto for some time. Maybe later (whenever that is)
> I haven't. Because (free, as in free beer) chatGPT is extremely slow, I have to make a rather extensive proompt to get the result I want to, and then I still have to debug most code.
That's because you are comparing asking ChatGPT to write full code to searching for a question on Stack Overflow and adapting their answer (which is comparing apples and oranges).
Try using ChatGPT like you use Stack Overflow instead (i.e. the question is "How would I record an audio stream to disk in Python" rather than "write me an application / function which...").
As an aside, try "How would I record an audio stream to disk in Python"" in both GPT4 and searching for an answer on Stack Overflow and see what has the better answer! (Clue: GPT4, and if you don't like GPT4's answer just ask it to clarify/change it)
>Try using ChatGPT like you use Stack Overflow instead (i.e. the question is "How would I record an audio stream to disk in Python" rather than "write me an application / function which...").
That's my point though. I get, that it can produce quite good results, if you are specific enough. And for some applications it makes sense to take your time and describe that as much as possible.
Most of the time I just need some small snippet though and usually I can get that with just a few keywords in my favorite search engine, which is way faster. So the conclusion is: There is no one or the other. They should be used complementary, or atleast that's what I am doing (as in use the search engine for quick hints and chatGPT for some more verbose stuff 'write me a parser for this csv in awk'.)
Personally ChatGPT generally gives me a quicker, better, simpler and ad-free result for the snippet (At least with GPT4).
Plus I can ask follow-up questions in a context-driven way ("Can I do this without importing a library?").
I'm aware that different people will have different feelings on this though and personal tastes will differ, but while search engines stagnate I suspect the needle will continue to shift towards AI.
Anecdata: I've started asking Bing these questions instead of SO. E.g., it recently gave me a very helpful answer for debugging a Spring issue and cited its sources. What it didn't do was present me with a whole lot of moderation cruft.
I can ask for recommendations for tools and libraries, which IIRC SO disallows.
I also don't have to pray my question will get enough vote attention or worry that I posted it at the wrong time of day.
On the whole, going the GPT route has been more satisfying in all ways.
small nitpick but I thought it was just an icon but it turns out to be the button for switching light/dark mode. It would be great if you could replace it.
> I can ask for recommendations for tools and libraries, which IIRC SO disallows.
Bing Chat almost always is useless for me with these kinds of queries. A few days ago I asked for a tool that monitors to see if a website is up. I told it I needed the tool to be something I'd run locally - not an online service and not something I need to sign up for.
If you loosen up your definition of of LLM the moderators and posters are really just LLMs that have been jailbroken to insult you and close your post.
Stack Overflow is the programmer's internet bloodsport.
How is this being enforced? It's either bots banning bots in a digital game of whack-a-mole; or humans arbitrarily trying to asses whether something has been written by an LLM or a human.
Yeah, I've been using chatGPT quite intensively. And while pure LLM output is relatively easy to spot, human edited LLM output is almost impossible to detect. Most of my message above has actually been written by GPT4 (3 prompts + some light editing).
It's human judgement. Definitely not perfect, but something has to be done to prevent SO from being overrun.
There are some subjective signs that a post is LLM generated, like being overly verbose and making unrelated assumptions, or mix of horrible and perfect grammar. Those bans are hard to justify because the false positive rate is high.
But other signs are pretty obvious. My favorite is the use of APIs that should exist but don't. Passing parameters that neatly solve the problem but have never been accepted, or importing non existent libraries. I'm happy to flag those.
And I think Stack Exchange needs a new CEO. Maybe new owners, which is since 2021 Prosus. My impression is that they don’t understand what is the purpose for developers.
If the lights are on at SO then they must be in the process of training their own AI with their own dataset and documentation for the topics covered. That is what it would take for me to make SO my first stop again. It should be very doable for that talented group.
These figures show an extremely precipitous (and permanent) decline in traffic over the course of a few days in May of 2022 [0], during which the number of daily new visits dropped from ~1M to ~300K, the number of total daily page views dropped from ~20M to ~14M, and the number of daily sessions dropped from ~9.4M to ~6.1M.
However, there is no commensurate decrease in posts/votes during the same time period. Posts/votes remained relatively constant through 2022 (modulo normal seasonal fluctuations), until February 2023 when both fell off a cliff (I assume due to the rise of LLMs). Traffic data are sourced from Google Analytics, while post/vote data are computed internally by StackOverflow [1]. I wonder if the apparent precipitous drop in traffic in May 2022 is simply an artifact of Google Analytics suddenly changing how it tracks traffic/visitors.
the post doesn't explain where they got these traffic numbers, and it seems unlikely they have access to stackoverflow's real traffic stats. they're using some sort of estimation here. there's always a chance that their estimates are wrong - especially if they're showing implausible shifts like this.
That makes sense. "New visits" are first time users, likely young coders who are looking up answers to things on a search engine, find what they're looking for on Stack Overflow, maybe click on an ad, and leave. They probably don't vote or post much. A sudden die-off there suggests something very bad happened to organic traffic (change in Google? Terrible new SEO scheme? A sudden stop in ad buys?)
The new content rate been dropping at a dismally constant rate for a long time, but the first few months in 2023 were awfully grim. I wonder what might've corresponded to that.
If SO was worried about that drop I think they would have bought back some of that traffic. More likely something has changed how they count the visits or they blocked some bad traffic. Traffic data is often sampled as well.
The fall in the beginning of 2023 may be the introduction of ChatGPT. A more worrying idea is that the numbers reflect not just the decline of SO but a decline of the whole IT business.
Which would make sense, right? You are more likely to get an answer on StackOverflow for questions that touch very common technology (because more people are likely to answer). And that is exactly where Copilot probably shines too (I don't use it): because that is where there is a lot of training data.
I personally used to like StackOverflow as my last recourse: I grew up in those years where we had to RTFM, and I kept the habit. So if I go ask on StackOverflow, it is a tricky question. It used to be fine, and I was getting an answer eventually (sometimes after adding a bounty).
But in the last few years, I have had legit questions downvoted or even closed, and it was obvious that the people voting to close it did not even understand it. I agree that the moderation culture on StackOverflow is toxic. If everytime I contribute something, I have to fight to not get downvoted or closed, then I will slowly stop contributing.
From the comments on this answer https://meta.stackexchange.com/a/391625/136010 it is suggested and agreed by staff that the chnage in May 2022 was the role out of a proper cookie consent form. If you don"t have performance cookies SO can"t work out the analytics.
Fro staff member Catija
"@JourneymanGeekOnStrike Yeah, if you go back further, the "traffic" numbers see a 40M/week drop between April and May 2022, which is when the cookie tracking changed, and then normalizes again until December. So, prior to the cookie changes, traffic was about 140-150M per week. But, to be clear - this is stuff we're aware of and have "corrected" for, I guess."
This is interesting. So, train llms on "everything", including SO, and then replace "everything", including SO. I think we'll see companies such as google, bing, etc... paying to scrap different websites.
for FIVE YEARS, because of a rollback war on some of my own answers. some idiot was gaming the system and adding minor punctuation here and there, just to get review points. so yeah, good riddance.
I have since deleted my account on SO, which was actually a somewhat involved process. They did not however allow me to remove the answers that I provided.
thanks for the valuable comment. I realize now that my 1,800 answers mean nothing, and that the only important thing is my comment etiquette. I will try to do better, random stranger.
really interesting as this isn't like digg or myspace, it was built on information, rather than interaction, and will probably fall like the tower of babel.
but i suppose it will still be a very valuable dataset to anyone looking to train a coding model. its the single most valuable resource that could be archived for one
Would this not be due to SO’s demise, but due to the fact the tech has become more stable and boring to the point that there are fewer new questions and answers?
My main issue, with the main Stack Overflow site (rather than some of the other stack sites), is that I don't get answers to my questions, possibly because no-one knows the answer. If they did, then it's probably been asked before, and since I always look first, I've found the answer to the question that was asked, and so don't eve pose the question, though I might contribute to the answer I found in some way.
I remember trying to learn front end development around 2013, was fascinated by responsive web design and twitter bootstrap. Asked some questions on that site, was mostly ridiculed for my amateur questions several times, never touched the site again and also never learned front end. So this is my story with that site.
Same. I got put in StackOverflow jail for posting my contribution an answer because I didn't have enough karma to post a comment on a previous answer (or maybe it was the other way around, I forget). Never mind that I was earnestly trying to help the original poster, and pointed out a legitimate mistake in one of the answers. I broke protocol and had to be punished.
Not quite true, though indeed getting upvotes seems not to be as easy as it once was. I still get an upvote on some of my relatively recent answers once in a couple months, although it depends on the answer.
The quick way to accumulate reputation is through bounties. However, it's very much a lottery: bountied questions are often about ultra-specialised niche topics. You may need to hunt for a long time to find something you actually know something about.
You don"t need any karma to answer any logged in thing can answer. I say thing as ChatGPT is being used to produce a load or crappy wrong answers now.
Thus I don"t understand your issue. This is a XY-problem :) I know enough about the subject to know that your issue is not the actual issue as anyone can answer, if you had issues then there is something else.
My SO account is almost 12 years, with just over 2k reputation and I don't really care. Until now I am still somewhat helping answer some basic questions in the mobile development area tags, my only gripe with SO is the hostile nature of some mods with large reputation. Some seem to get a kick out of this and forgot that reputation does not translate to expertise.
For 12 years, they have not figured this one out. New users will ask a very valid question and then won't respond anymore. I have seen this one played out every single day. Back in the day, users were generous with the upvotes even for a simple basic question, this is not the case anymore today.
I think that with the rise of push notifications, no one really goes to a site to check notifications anymore. So the new user may have not developed the muscle memory to go back to SO and participate. I suspect this also has something to do with the decline of forums. Reddit still works because the app sends 200 notifications a day, but without it, I don’t think it would be as popular.
Also SO is participation hostile unless you’re a pro, so as a newbie I’m not going to do anything other than ask and lurk, because I’m not worthy
At least part of the reason for the hostility is that SO is a game. You get points, but you can also prevent others from getting points by voting down or removing their questions and answers.
On SO this hostility is pronounced because participants believe that if they get a lot of points they have easier time finding a well-paying job.
I don't know if we all do it the same way. I don't use push notifications for barely anything, because I don't want to be disturbed by random sites (least of all linkedin or SO.)
Since SO is often used in a professional capacity, that problem could have easily been fixed by dev tools providing a formal way to link to SO traffic for topics that are relevant to the team.
It's just been a while since anyone has started trying to integrate tools with each other, outside of the established players.
these peasants with high reputation thinking they are johnskeet
reputation is meaningless and bloated in stackoverflow now there are many 100k reputation people because asking or answering basic shits on javascript/python/pandas/git
I don't use SO sites anymore because of the incessant "accept our cookies" popups that aren't even relevant to me because I don't live where EU cookie laws or GDPR applies.
805 comments
[ 2.5 ms ] story [ 264 ms ] threadWhat accounts for this?
Although there appears to be a Monica effect (2020) in the data too. I stopped contributing myself around that time:
https://meta.stackexchange.com/questions/340906/update-an-ag...
It is also possible there's a methodology issue. There's no source for where this data comes from.
I scrolled down a lot and didn't see any scraper site links.
I can often evaluate a Copilot autocompletion (check that code looks right at first glance, check that it compiles, hover over method+type signatures to see their docs, run the code) in less time than it would take me to find+read a Stack Overflow answer.
Code is easy to verify anyway.
IMHO people have tired of the smug dismissive assholes that dominate SO and its its insipid gamification and they have, gradually, found alternatives.
For some, it's more welcoming and knowledgeable communities in github issues+discussion.
For others it's special-purpose, interactive forums that provide more guidance and non-hostile support for users of certain platforms/tools (eg Posit Community).
For many, copilot has been fulfilling that need, Does it always give "the correct" answer? No. Does it provide a sketch of a solution that gets you half-way there when dealing with tedious humdrum stuff that you just forgot because it's so boring? YES.
For yet others it's just plain old reddit where you can ask a question and whether it gets smacked down or not depends on how cool the community happens to be.
I can tell you that I quit in the last 6 months. All new answers are answers on questions from first time posters, answering questions that have good solid answers from 8+ years ago (mostly 10 years), and it is pretty obvious, they are just doing it to get points. The question/answer is not something that has changed in the last 8+ years, they are just doing it to get reputation or whatever.
Most good solid questions...I just do not see too many today. I am not saying this is the reason for the decline, but there is a deluge of posts answering very old questions with a slight modification of an old answer. I quit putting effort into it, as it was just non-productive.
It's entirely possible that the same is happening to SO.
One of the things I say with some frequency is that all failures are engineering problems. Blaming the users/customers, in the end, fails to take advantage of an opportunity to improve the product, learn from mistakes and improve process.
While I understand it's a lot of work, and some (most?) might be posting low quality answers, my main problem with stack overflow has been exactly that over time, the top answer becomes stale.
The question is still relevant, but the best practice or library, platform whatever changed... That is a problem the stack overflow model has a hard time adapting too.. and it just gets worst over time..
Original accepted upvoted to +20 answer: "Take Y and drive it into X, grep output for Z and place in AZ for processing with the -kombucha switch"
Answer posted yesterday by "NewUserXX123", a first post: "Take A and drive it into B, grep output for C and place it in DD for processing with the -kombucha switch"
It gets pretty tiring responding with a custom comment and downvoting, only to have "NewUserXX123" then cuss you out in a DM about how you are a total @!#$%()&!@$%.
I hit upon those useless answers all the time. I have no need to know about solutions that might have worked a decade ago and are obsolete. Stack Overflow is useless now.
Gamifying reputation was one of the worst things we've done. I was part of this in previous positions and after a while it became apparent that it brought out the worst in people. There are better ways of incentivizing usage rather than stupid internet points.
Agreed and I don't know this for sure but the points lust, it seems to me, is not limited to questioners and answers, some of the behaviour I've seen from mods seems to be driven by that too.
Just to repeat, I don't know that because I don't know what gets you points but the behaviour is hard to understand otherwise.
I had a question where I had finished with a one word sentence "Thanks." . A mod had removed that part of the question, when I looked at their record they had an awful lot of edits which consisted of similar, tiny, non-substantive changes. It did make me wonder about their motivation.
Anyway, yes, imho gamification not a good idea.
https://stackoverflow.com/help/behavior
"Do not use signature, taglines, greetings, thanks, or other chitchat.
"Every post you make is already “signed” with your standard user card, which links directly back to your user page. Your user page belongs to you, so fill it with information about your interests, links to stuff you’ve worked on, or whatever else you like!"
"Thanks and other statements of appreciation are unnecessary, and, like other chitchat, should not be included."
"If you use signatures, taglines, greetings, thanks, or other chitchat, they will be removed to reduce noise in the questions and answers."
----
Removing it doesn't get the editor any points, they are spending their time cleaning up your question to site standards for no reward at all.
https://meta.stackoverflow.com/questions/260776/should-i-rem...
https://meta.stackexchange.com/a/142703
> Wolfgang Amadeus Mozart[a][b] (27 January 1756 – 5 December 1791) was a prolific and influential composer of the Classical period. Despite his short life, his rapid pace of composition resulted in more than 800 works of virtually every genre of his time.
it opened with:
> Hello everyone, I studied classical music in college but have been out of the scene for a while and now I'm getting back in, I was wondering about the history of Mozart and I remember that his middle name was Allen or Almond or something? Can anyone help? Thanks for any information. || Hi, I learned about him in middle school and we all joked that his middle name was Armadillo but I think that's wrong, haha. || Greets all, it's in Olivier Hallengrunsch's Classical Composers as 'Amadeus', and that's well regarded. It also says he lived (27 January 1756 – 5 December 1791). I think we can all agree he was prolific, right? Thanks and regards, Jason [xxKiller; AMD Ryzen2 32GB RAM 2TB Western Digital SSD; BMW 330 2l aircooled] || Why's nobody mentioning how short that life was, smh || Good evening all and sundry, m'lady (tips hat), forsooth would anyone speak to how many works he is believed to have composed, all considered? Methinks such knowledge would be a most hearty addition to this esteemed gentleman's biography - Martin, [Fort Lauderdale TX Ren. Faire organiser 1997-1997] || etc.
StackOverflow isn't a forum, it's a collaborative reference work. Meta-chat would be edited out of a wikipedia page and goes on a separate 'talk' page (equivalent: meta stackexchanges or the stackexchange chat). What if you then went to the Wikipedia talk page and said "Is overzealous moderation of questions ruining Wikipedia? I want to be able to edit questions and greetings into pages but there are hoards of awful literal-minded jobsworths cruising the site just looking for the slightest reason to edit a spelling or grammar mistake or revert my changes. It just leaves me feeling unwelcome"?
Why would you expect to feel welcome when you're spoiling what others are trying to build up and insulting them for caring??
Some really old questions (circa 2009) have the answer to the latest version, that required scrolling through multiple pages.
I'm just not sure what value there was to have a team of people volunteer their time to making sure there were absolutely zero duplicated questions/answers? What problem is that solving? If anything, it is adding problems by not letting the content naturally evolve in time.
If people want to assess the quality of answers, it's better to have multiple data points, and SO could have invested those resources in algorithmically linking similar questions and making it easy to navigate by both answer reputation and time.
Along similarly lines, it seems best to tackle the people trying to game the system algorithmically as well - if content is word for word duplicate, that's a problem which can be solved by computers instead of people (similar text is a solved problem).
It seems the only use case for moderators on SO is for removing truly inappropriate content - it's wild to me that moderators were spending significant amounts of time actually removing technical questions and answers.
It reminds me a bit of reddit moderation, where reddit communities enforce these non-sensical rules and by extension require hugely heavy handed moderation to 'curate' their communities. Like, the headphones subreddit disallows pictures of headphones in boxes. Why? If the community is interested in headphones, what's the difference if it's a box or not? It's not like if you take it out of the box the headphones look different than any picture you can find on the internet.
Seems like many of the problems of moderation are artificial rules endlessly being enforced by real people which ends up just being pseudo 'make busy' work.
It would more sense to establish lineage. Lock older questions from more answers after some point in time, and instead of closing and linking to original, do it the other way. Keep new questions open and link to past variations. Then in the old threads indicate newer guidance may exist and link forward.
IIRC, the vision of the perfect question was one with a canonical answer. They didn’t want to be Quora. I think that concept made and makes a lot of sense, but doesn’t capture the “meta” issues surrounding how you litigate the form of the question, especially as questions get more nuanced.
Is it inevitable that traffic would decrease as questions are answered? Has traffic gone somewhere else? Has AI replaced it?
Now, AI kinda replaced for me, it's copilot first, then ChatGPT if copilot is not enough, then google/bing if ChatGPT is not enough and maybe finally SO.
Conservative figure from my side would be 15-25% less stackoverflow in a regular day.
I'm sure I'm not the only one.
Now I'm sure that ChatGPT probably scraped StackOverflow, perhaps using the very same answers I got from StackOverflow, but combined with other answers from other sources resulting in the instant answer. That does not necessary mean AI will replace StackOverflow, it just means that people wouldn't ask redundant questions in StackOverflow anymore, just questions that AI can't solve.
- Last month, Stack Overflow had ~142,575,642 visits.
- Last month, Google gave 127,896,508 visits to Stack Overflow
- Last month, Bing gave 7,491,274 visits to Stack Overflow
You could say that the Stack Overflow # of pageviews depends mainly on:
- How often people are searching Google/Bing for answer.
- How often Google/Bing rank Stack Overflow high enough for people to click into it.
https://medium.com/snowflake/how-to-load-the-stack-overflow-...
The Hugginface repo unfortunately prefilters some of the tables/rows according to some criteria, making it less usable for general analytical queries that the BQ or SEDE datasets enable. If anyone knows of an 'XML-streaming' solution that directly samples from the Internet Archive's data dumps, I am all ears.
[1]: https://huggingface.co/docs/datasets-server/rows
[2]: https://huggingface.co/datasets/HuggingFaceGECLM/StackExchan...
For instance, this user query by Starball tracks network contributions over time[2][3].
[1]: https://data.stackexchange.com/meta.stackexchange/queries
[2]: https://data.stackexchange.com/meta.stackexchange/query/1759...
[3]: Static image if the query times out: https://i.stack.imgur.com/LYZQm.png
especially YouTube links. sadly, it would not surprise me if these people are earning a decent enough money from ads to make it worth their while to be "content creators" solely from search results from Googs
If three segments of the internet think the same piece of information is relevant, that should affect the score of all 3 copies, not just the largest segment.
When content republished on some bizarre/sketchy/unaccountable ~adfarm outranks the site where it first appeared, users of Google's search service end up at higher risk of getting phished or infected with malware.
Is there some benefit you see here that outweighs this downside risk?
A few years ago, my programming-related queries would hit Stack Overflow as the first or second result. Now it's very frequently spammy garbage in the top 2-3 slots.
Remember, when people were hired because they knew the secret sauce on how to get the best Google ranking. Google experts?
Well, it turns out that the person at Google that was responsible for keeping the algorithm fresh and the search results fresh retired and everything went to shit when they left.
Actually, I’m betting that person did leave the company, but the real damage happened when someone came along and convinced everyone they knew the real trick to better search results and we have the shit that is now Google. Nice work new guy! Let me rephrase that. Nice work to the guy that thinks they are smarter than everyone else and still thinks their approach is the best, yet evidence to the contrary.
It’s simply how times change and people with it. Knowledge is lost when people move on and the reasons why certain decisions were made are not transferred.
At any rate, I imagine people at Google are trying to figure out why there is such a negative opinion on their search results lately.
Why should they care? They're too big to fail…
Google controls almost the whole end-user realm through Chrome & clones, and Android, the dominating end-user OS by a wide margin.
At the same time end-users are completely helpless and can't do anything against Googles liking because they don't understand anything about IT tech.
Computers are black magic to most people so they're trapped. This never changed! Especially millennials and gen-z are completely clueless as they didn't had the chance to use personal computers ever, where you had at lest some control over the device and needed to know at least some basics about its inner working. All the younger people know are the tightly sealed black-box devices you don't have any control over, called mobiles, which are fully operated by big-tech. Google search + Android apps are "the internet" for most people. They mostly don't even know there is something else beyond that, so Google can do whatever they want, and this will have exactly zero consequences for them by now.
Google's move regarding rolling out "browser DRM", the next "trusted computing" initiative, regardless of what anybody thinks about is is very telling.
Now they will violently reap the fruits of their monopoly, and likely nobody will be able to stop them in the next decade. People where warned about the consequences of this monopoly for many many years. Nobody cared. Now it's payout day for Google.
Unfortunately anything popular is a target, and once they stopped trying to innovate to keep ahead of systems gaming they were going to lose eventually. It's amazing it took this long.
I can think of a few theories that I don't think hold water:
1. The rise of ChatGPT to answer many questions that StackOverflow would previously have been used for. This seems unlikely, since the timing doesn't really work out.
2. The perennial complaints about StackOverflow's culture of closing everything as duplicates or offtopic. This seems unlikely as well, since those complaints have been common for a decade or more.
3. The prevalence of SEO-optimized scrape sites - the ones that pop up with a "blog post" merely reposting a stack overflow question + answer in a different font. I've seen these for a while, and anecdotally they feel more common that they used to, but I couldn't give any real timeline for that vague feeling.
4. StackOverflow internal politics? I've seen the occasional stack-overflow meta thread pop up periodically on HN or social media, but I don't recall anything earth-shattering recently.
5. Most questions have good answers now and there's less need for new ones. I'd have bought that answer 10 or so years ago when StackOverflow's pile of questions + answers reached maturity. I don't think it suddenly hit some sort of answer saturation point in 2021.
My guess is that it's a slow shift in the culture of the StackOverflow userbase:
- Being a top answerer confers some cachet and makes you more employable in some places
- People notice this and start looking for the most effective ways to become a top answerer
- The most effective way is fast, low-effort answers
- There's been a rise in such low-effort answers over the last 5 years or so
- As a result, the cachet of being an StackOverflow top answerer is a lot lower
- The really good, deep, technical answerers (as well as the mods) are leaving as that cachet goes away
- Post quality starts dropping around 2021 and views start declining as people react to that in 2022.
I would choose first result - people are lazy. The problem comes and goes but it is true, spam sites, with content copied from SO, are ranking higher than SO itself. And it is often hard to tell you are viewing a copy on first glance, the layouts are usually more like a forum.
https://techcrunch.com/2023/07/20/tech-industry-layoffs-2023...
If I were building a Q&A site that genuinely wanted to encourage high-quality answers, I'd implement something like a 1-hour window where all answers are invisible immediately after the question is posted. This is to give people time to work on a good quality answer, without racing to be the first and gain those precious early upvotes. The UI could still indicate how many other people have answered / are answering, and if answer volume became a problem you could perhaps block new answers during that window after the first 20 or so have landed. When the hour is over, answers are displayed in random order and the existing site mechanics around upvotes etc kick in.
It feels like that could potentially address the problem of low-effort answers killing off the good ones.
Comment: "You shouldn't do A, it's better to do B"
Closed, duplicate of "How to do C".
Yeah sure, in the beginning, there were many more basic questions (how to increase number by 1, how to get division remainder, how to check if file exists,...), and if you're coding in Perl, you can still find all the answers... if you're working in python, you'll find answers for python2, some for python3, some specific to python2.6, etc... if you ask again, it'll be closed as a duplicate.
I know it's anecdotal, but after a few bad experiences, people just decide not to use that specific site anymore.
That makes it much, much less useful than it otherwise could be.
I would call that "stable" :) Or "good enough, that it doesn't need constant fixing" :)
My experience was that there was a long period where your complaint was quite true.
I believe that they have heard this complaint enough and have lightened up a little. There are still more and stricter rules than most sites, but the obsession with pruning duplicates seems to have cooled.
Stupid people don't realize they are stupid. When they can't understand a question they conclude there's something wrong with the question, not with them. And since they believe the question is silly, they happily close-vote the question.
This way they close questions which require specialized knowledge, experience, or deep understanding of the subject. Precisely the kind of questions I consider interesting to answer.
Sometimes I managed to undelete such questions, but SO made it easy to delete and hard to undelete.
Examples of the questions where I succeeded: https://stackoverflow.com/posts/57064879/revisions https://stackoverflow.com/posts/57323981/revisions https://stackoverflow.com/posts/65025858/revisions
In many more cases I haven't bothered. Sometimes, moderators killed questions so fast that I discovered that while trying to post my answer.
You probably assumed I asked that question? If so, the assumption was incorrect, I answered it.
> without explaining what you attempted, why it worked/didn't work, where you got stuck
I don't think any of that is possible to do. The question, and my answer, are too simple to decompose into smaller parts. Not enough runway to start and get stuck.
> expect other people to answer it fully
My expectation was rather different. I wrote my answer on Jul 16, 2019, and expected it to stay there.
Instead, on the next day some people have decided the question was bad, and closed it. Then on the next month, some other people have deleted it, along with my answer.
To close it, 5 people clicked once/each. To undelete it, I spent quite a few hours. Sadly, that's not a rare exception: moderation on SO is horrible.
The question is asking about the best way of doing X, which has a clear substep of "any way of doing X". Not only that, it's asking about a code solution, not just maths/combinatronics. Not a single line of code in a question about finding the optimal (not any) solution for doing some algebra in code is, IMHO, reason enough to close it.
Edit: for example, I'm not asking and so have no stakes on the question but can already try to think about brute-forcing it, which already shows more effort than the author SHOWS at attempting a solution.
To be fair, English is not my native language. But when I see a question “what’s the best way of doing X”, when it doesn’t have any criteria for what’s the best would be, and no other ways of doing X in the question, I consider “what’s the best way” part a redundant figure of speech. I view such questions an equivalent of “what’s a good enough way of doing X”.
> it's asking about a code solution, not just maths/combinatronics
Please read this: https://stackoverflow.com/help/on-topic According to that article, questions about math which don’t imply a code solution are offtopic on stackoverflow.com. They should be closed, and possibly moved to other stachexchange sites. According to that article, the OP’s question is good. The question was about a specific programming problem, and is a practical, answerable problem unique to software development.
> can already try to think about brute-forcing it
I’m not sure that’s actually possible to do.
- Does not show any attempt or willingness to try to solve the problem first on their own.
- Does not even give any indication of where the problem comes from, why it might be interesting, etc., it's just a "how to calculate X?", which could easily be a homework problem.
- It is about finding a (possibly) mathematical solution, and then implement it in C++. Two very big and different problem, asking the audience to do them both. Again no attempt to fix either of these two problems on their own.
- I'll concede the optimal thing might be a language issue.
> I’m not sure that’s actually possible to do.
But that's not my point, my point is that I already showed more willingness to try to solve this problem than OP. And THAT is a big problem. It's not on topic about any of those points, in fact if you remove the bit where OP is asking us to give them the full solution in C++ it could be a good question for the Mathematics SE!
None of that is required to ask questions on stackoverflow. For details, read “How do I ask a good question?” and “What types of questions should I avoid asking?” help articles. You’re inventing arbitrary restrictions.
Another thing is, “why it might be interesting” is subjective. Personally, I found the question interesting, that’s why I have answered it. You probably think otherwise, but note it only takes 3-5 votes to kill the question. Any question at all is guaranteed to have at least 3-5 people on that site who find it uninteresting, opinion-based, need more focus, duplicate, etc.
> Two very big and different problem, asking the audience to do them both.
Two big problems don’t have solutions which can be both explained in 3 short sentences. As you can see from my answer, the problem formulated in that question has such solution.
> I already showed more willingness to try to solve this problem than OP
You have not. However, you have demonstrated willingness to delete interesting questions based on arbitrary and subjective criteria, despite the question is perfectly in line with the stackoverflow guidelines. Which BTW is very on-topic, because I think that’s the main reason for the fall of SO being discussed here.
This, to me, is the mistake that SO made. Developers helping developers is the engine that runs the site, that's why people come.
The body of interesting knowledge is an emergent property of the support forum/peer-to-peer teacher.
Eventually they tried to put the cart before the horse and traffic has dropped.
It's OK that you're done answering "how do I change font color with jQuery" for the thousandth time and are only interested in the occasional very interesting question, because there are people behind you who do want to answer that question. That will help them grow to get where you are.
If we don't allow new generations of users to go through that process we went through, then StackOverflow has an expiration date.
ouch. as a former user of 10 years, I felt that. so true.
I think ChatGPT is amazing foe this because you will rarely have to deal with anyone who has this type of attitude when you need help in the future
The same mechanisms that make SO kind of brutal have also helped revolutionize asynchronous online Q&A.
__________________
0. https://www.zdnet.com/article/stack-overflow-ceo-on-how-it-b...
The other aspect that grinds my gears is the closing of duplicate question. Fine in principle, except when the original was answered 10 years ago and all of the answers are jQuery.
https://stackoverflow.com/questions/57064879/finding-coordin...
https://stackoverflow.com/questions/57323981/realtime-video-...
https://stackoverflow.com/questions/65025858/how-to-fetch-th...
The wheels of this kind of stuff turn slowly, but obviously ChatGPT was trained partly with data only gainable from a healthy StackOverflow-kind of site with users actively asking unique questions and enough people answering those unique questions with well-though-out answers. The shittyfuture outcome is that StackOverflow goes out of business and LLM's stagnate on this front, while being capable of answering in fuwwy speawk when prwompted, would still be limited to / biased towards older versions of libraries, software, languages, tools etc.
Claim asserted without evidence.
It's painfully slow. I can Google the question, click one of the top results, skip to the relevant part and read it faster than GPT can generate two sentences. You also have to build an elaborate prompt instead of throwing two/three keywords into it.
It doesn't help that GPT is insistent on replying in the three paragraph format, meaning that the first 30-40 words it creates are just trash to be ignored.
I found it useful once - when I had to write an essay about ISO 27001 for college and just wanted it to go away. Took what it generated and spent 20 minutes editing it to look closer to my style. For real work it isn't as useful.
I don’t even open SO anymore; if it has a direct answer to your question, the LLM almost certainly does too; and asking new questions on SO is basically impossible.
If you manage to survive the gauntlet of “too specific, already answered, not general interest, arbitrary moderator activity”, the chances of getting an answer that answers your question can take forever; most likely you’ll get a stupid answer that doesn’t answer it, upvoted by idiots who don’t understand that it not an answer the the actual question, and, ultimately, because it “already has an answer”, ignored, never to receive an answer.
Maybe one day, a passing savant will answer in a comment.
…and yet, you find it faster and more reliable?
You, and I, have had different experiences on stack overflow in the last two years.
> For real work it isn't as useful.
To you.
Which was the whole point of this chain :)
Ironically, this is why people like me prefer LLMs (when they're accurate). With Google, about 50% of the times the top SO hit is not answering my question. So I have to click 5-10 SO links, parse each one to see if:
1. The question being asked is relevant to my problem.
2. The answer actually answers it.
I may be able to do it quickly, but it is a tedious burden on my brain. While GPT doesn't always work, the nice thing about it is that when it does work, it has taken care of this burden for me.
Also, GPT's pretty much memorized a lot of the answers. I once asked it an obscure question involving openpyxl. It gave a perfectly working answer. I wondered: Did it reason it and generate the code, or is there a SO post with the exact same answer? So I Googled it, and sure enough, there was an SO question with the same code!
Except GPT's solution was superior in one tiny respect: The SO answer had some profanity in the code (in a commented line). GPT removed the profanity :-)
When I search for something, I want factual and correct information, which is not what a language model is for.
However GPT4 is occasionally still inadequate so I certainly hope SO sticks around, it is a tremendous resource.
I currently pay for it but I'm not sure if it makes any difference.
Ask it how to use it
My experience - the free version made up a lot of things but still felt very useful - enough to want to upgrade to the paid version. With the paid version, I notice very rarely that it hallucinates. It does make errors but it can correct them when I provide feedback. It is possible that I just do not notice the errors you would notice, it is also possible that we use it differently. I would like to know.
Paid.
> Would you kindly provide links to a few examples of very fundamental things incorrect?
No, definitely not.
> I notice very rarely that it hallucinates.
Unsure of what "hallucinates" means in this case. Some examples of things I've used it for: docker configuration, small blocks of code, generating a cover letter, proofreading a document, YAML validation, questions about various software SDKs. The outcome is usually somewhere on the spectrum of "not even close/not even valid output" to "kind of close but not close enough to warrant a paid service". When I ask for a simple paragraph and I get a response that isn't grammatically correct/doesn't include punctuation, I'm not sure what I'm paying for.
The term "hallucinations" is now commonly used for instances of AI making stuff up - like when I asked ChatGPT (before I had paid account) to recommend 5 books about a certain topic and two of the recommended books looked totally plausible, but when I tried to find them, I discovered there are no such books. This is where I see a big difference between GPT-3.5 and GPT-4.
>> I get a response that isn't grammatically correct/doesn't include punctuation
What punctuation? If you mean stuff like commas separating complex sentences, my English is definitely not good enough to spot that. But your mention of punctuation reminded me of problems that ChatGPT has with my native language... any chance you are using ChatGPT in a language other than English?
Here is an example of using it to write simple powershell scripts:
* https://chat.openai.com/share/17a9d185-4a8c-4c10-97a2-0c5b08...
Here is an example of using it to merge different types of json configuration files:
* https://chat.openai.com/share/ed0c9cf9-7586-43df-9271-7b5859...
You need to pay for it if you want access to the latest version of the model, along with some beta features like plugins. Plugins are extremely useful and it is worth paying just to get access to them. For instance, you need to have a certain plugin to get it to read links.
That's not very convenient, atleast for now. I got so used to search engines by now, that it only takes a few keywords to get the expected result. Be it a SO-answer or a documentation page. And as people have mentioned, chatGPT was learned on the stuff that's on the internet, so if there will never be any new stuff, because people just use AI, then it will not learn and won't answer your new questions. For some edge cases I might try AI here or there, but usually it's not for me.
Hell, there comes even an example to my mind. I recently just asked chatGPT what a single-issue 5 stage pipeline on a CPU actually means. I wanted to know if, especially, the "single-issue" meant that only one instruction is present in the pipeline at a time, or if a new one gets shifted in on every clock cycle (if there is no hazard). It just couldn't answer it straight-forward. It was also kinda hard to find the exact definition on the internet. I found it in a book from the 90s which was chilling in my book shelf (Computer architecture and parallel processing by Kai Hwang). Hint: Single-issue just means that only one instruction can be in one stage at a time, but still multiple get processed inside the pipeline. The keyword is 'underpipelined'
I'll just keep an eye on AI progress, but will probably not make it my goto for some time. Maybe later (whenever that is)
That's because you are comparing asking ChatGPT to write full code to searching for a question on Stack Overflow and adapting their answer (which is comparing apples and oranges).
Try using ChatGPT like you use Stack Overflow instead (i.e. the question is "How would I record an audio stream to disk in Python" rather than "write me an application / function which...").
As an aside, try "How would I record an audio stream to disk in Python"" in both GPT4 and searching for an answer on Stack Overflow and see what has the better answer! (Clue: GPT4, and if you don't like GPT4's answer just ask it to clarify/change it)
That's my point though. I get, that it can produce quite good results, if you are specific enough. And for some applications it makes sense to take your time and describe that as much as possible.
Most of the time I just need some small snippet though and usually I can get that with just a few keywords in my favorite search engine, which is way faster. So the conclusion is: There is no one or the other. They should be used complementary, or atleast that's what I am doing (as in use the search engine for quick hints and chatGPT for some more verbose stuff 'write me a parser for this csv in awk'.)
Plus I can ask follow-up questions in a context-driven way ("Can I do this without importing a library?").
I'm aware that different people will have different feelings on this though and personal tastes will differ, but while search engines stagnate I suspect the needle will continue to shift towards AI.
I can ask for recommendations for tools and libraries, which IIRC SO disallows.
I also don't have to pray my question will get enough vote attention or worry that I posted it at the wrong time of day.
On the whole, going the GPT route has been more satisfying in all ways.
Bing Chat almost always is useless for me with these kinds of queries. A few days ago I asked for a tool that monitors to see if a website is up. I told it I needed the tool to be something I'd run locally - not an online service and not something I need to sign up for.
It gave me 3-4 online services.
I reminded it about my constraint.
It gave me 3-4 more online services.
I reminded it about my constraint.
It said it didn't want to talk to me any more.
Stack Overflow is the programmer's internet bloodsport.
Better than reddit where you get insulted and banned by an actual LLM.
How is this being enforced? It's either bots banning bots in a digital game of whack-a-mole; or humans arbitrarily trying to asses whether something has been written by an LLM or a human.
There are some subjective signs that a post is LLM generated, like being overly verbose and making unrelated assumptions, or mix of horrible and perfect grammar. Those bans are hard to justify because the false positive rate is high.
But other signs are pretty obvious. My favorite is the use of APIs that should exist but don't. Passing parameters that neatly solve the problem but have never been accepted, or importing non existent libraries. I'm happy to flag those.
See: https://meta.stackexchange.com/questions/389811/moderation-s...
And I think Stack Exchange needs a new CEO. Maybe new owners, which is since 2021 Prosus. My impression is that they don’t understand what is the purpose for developers.
However, there is no commensurate decrease in posts/votes during the same time period. Posts/votes remained relatively constant through 2022 (modulo normal seasonal fluctuations), until February 2023 when both fell off a cliff (I assume due to the rise of LLMs). Traffic data are sourced from Google Analytics, while post/vote data are computed internally by StackOverflow [1]. I wonder if the apparent precipitous drop in traffic in May 2022 is simply an artifact of Google Analytics suddenly changing how it tracks traffic/visitors.
[0] https://i.imgur.com/qMj7Lge.png
[1] https://news.ycombinator.com/item?id=36856249
the post doesn't explain where they got these traffic numbers, and it seems unlikely they have access to stackoverflow's real traffic stats. they're using some sort of estimation here. there's always a chance that their estimates are wrong - especially if they're showing implausible shifts like this.
The question and user profile view stats (among other things) are public on the Data Explorer and included in the dumps:
https://meta.stackexchange.com/questions/2677/database-schem...
Not sure what they’re counting as a “visit”, though.
The new content rate been dropping at a dismally constant rate for a long time, but the first few months in 2023 were awfully grim. I wonder what might've corresponded to that.
The fall in the beginning of 2023 may be the introduction of ChatGPT. A more worrying idea is that the numbers reflect not just the decline of SO but a decline of the whole IT business.
I personally used to like StackOverflow as my last recourse: I grew up in those years where we had to RTFM, and I kept the habit. So if I go ask on StackOverflow, it is a tricky question. It used to be fine, and I was getting an answer eventually (sometimes after adding a bounty).
But in the last few years, I have had legit questions downvoted or even closed, and it was obvious that the people voting to close it did not even understand it. I agree that the moderation culture on StackOverflow is toxic. If everytime I contribute something, I have to fight to not get downvoted or closed, then I will slowly stop contributing.
"@JourneymanGeekOnStrike Yeah, if you go back further, the "traffic" numbers see a 40M/week drop between April and May 2022, which is when the cookie tracking changed, and then normalizes again until December. So, prior to the cookie changes, traffic was about 140-150M per week. But, to be clear - this is stuff we're aware of and have "corrected" for, I guess."
https://stackoverflow.com/users/1002260
for FIVE YEARS, because of a rollback war on some of my own answers. some idiot was gaming the system and adding minor punctuation here and there, just to get review points. so yeah, good riddance.
I have since deleted my account on SO, which was actually a somewhat involved process. They did not however allow me to remove the answers that I provided.
but i suppose it will still be a very valuable dataset to anyone looking to train a coding model. its the single most valuable resource that could be archived for one
My friend, a much less skilled developer, was much better at the craft of SO. I just lurk/solve my own issues
I have only had great experience with SO, but regardless, don't get dissuaded by those with less than mediocre tact.
The quick way to accumulate reputation is through bounties. However, it's very much a lottery: bountied questions are often about ultra-specialised niche topics. You may need to hunt for a long time to find something you actually know something about.
Thus I don"t understand your issue. This is a XY-problem :) I know enough about the subject to know that your issue is not the actual issue as anyone can answer, if you had issues then there is something else.
For 12 years, they have not figured this one out. New users will ask a very valid question and then won't respond anymore. I have seen this one played out every single day. Back in the day, users were generous with the upvotes even for a simple basic question, this is not the case anymore today.
Also SO is participation hostile unless you’re a pro, so as a newbie I’m not going to do anything other than ask and lurk, because I’m not worthy
On SO this hostility is pronounced because participants believe that if they get a lot of points they have easier time finding a well-paying job.
It’s true, though. I have a high-rep account and I’ve had a few jobs fall in my lap because of that.
It's just been a while since anyone has started trying to integrate tools with each other, outside of the established players.
reputation is meaningless and bloated in stackoverflow now there are many 100k reputation people because asking or answering basic shits on javascript/python/pandas/git