Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

What AI did to stackoverflow in a graph

data.stackexchange.com

Post reply on HN
Print view View original

511–520 of 615 posts

Re: What AI did to stackoverflow in a graph

#511

Post by eecc »

zahlman wrote: post #464
kstrauser wrote: post #447

Earlier quoted context omitted.

I reverted nearly every single edit to my answers other than obvious typos. If I’d have meant something other than what I said, I’d have said it. You want to see your own words on the page? Write an answer of your own. Lordy, that use to piss me off most fiercely. I don’t want someone else’s words attributed to me.

It is taken very seriously in Stack Overflow editing policy that edits are not supposed to go against the author's apparent intent. But it does involve removing a lot of things that you might want the answer to say, but don't meet the guidelines for what answers are supposed to contain. When you join the site, you agree to Terms and Conditions that, among other things, grant a Creative Commons license to the communit…

Found the SO mod…

Re: What AI did to stackoverflow in a graph

#512

Post by AussieWog93 »

brynnbee wrote: post #131
momocowcow wrote: post #67

Can it please do the same to reddit?

If you made a chart showing real human activity on reddit I imagine there's a big decline. The obviousness of bots and bot upvoting is undeniable

Had the same thought. Since around 2022-2023, most subreddits insta-ban you or hide your posts unless you have x karma (you gain karma by posting and getting upvotes).

So unless you happen to have an old account with karma from pre-2022, the only new users whose posts are actually getting through are the ones that know the specific subreddits that let you "karma farm" - ie, mostly bots and bad actors.

Also tangentially related, much of the weirdness of Reddit ("Thanks for the gold, kind stranger", fedora-tipping, narwhals baconing at midnight, references to weird memes etc.) also happened to disappear after GPT-3 came out, and get replaced with really strong opinions about politics and investments.

Re: What AI did to stackoverflow in a graph

#513

Post by NetOpWibby »

zmmmmm wrote: post #488
NetOpWibby wrote: post #461

Earlier quoted context omitted.

I say this every time the topic of SO/AI comes up. LLMs took off for me when I realized I didn't have to be an expert to ask questions and get a non-patronizing response. There was a time when I asked something along the lines of, "How do you get your own protocol like http://?" and was told to take my question to another StackExchange site. This happened several more times (along with dismissive responses and votes…

this for me was one of the truly liberating parts of LLMs for me: simply being able to ask a question and get a straight answer without having to run the gauntlet of why I shouldn't be doing what I am trying to do in the first place. The friction of trying to explain why I still wanted to do it that way deterred me from asking so many questions.

I also had to deal with a concerted effort by an ex-employer to downvote all of my questions for several months.

I’m not shedding tears over SO. It had a good run.

Re: What AI did to stackoverflow in a graph

#514

Post by digitalPhonix »

zahlman wrote: post #460
DrewADesign wrote: post #215

Earlier quoted context omitted.

Their rules, (I believe unintentionally) give iron-fisted fiefdom rulers a toolbox of justifications to control and alienate under the guise of protecting the quality of the site data. I honestly don’t even think most of the control freak mods could objectively judge the propriety of their actions because it was all encouraged by the rules. (And I do not think this was universal among the mods, but it was certainly e…

> Their rules, (I believe unintentionally) give iron-fisted fiefdom rulers a toolbox of justifications to control and alienate under the guise of protecting the quality of the site data. That isn't what happens. I know many, many people believe it to be what happens, but I know from years of seeing the process on the inside that it's absolutely not what happens in the overwhelming majority of cases. The "quality of t…

> That isn't what happens. I know many, many people believe it to be what happens, but I know from years of seeing the process on the inside that it's absolutely not what happens in the overwhelming majority of cases.

If quite literally every person I interact with professionally has an anecdote about this happening, your anecdote about it not happening is not very convincing.

Have you/SO staff/SO mods considered why this impression is so prevalent if you’re confident it’s (as you claim) not actually true?

Re: What AI did to stackoverflow in a graph

#515

Post by ConorSheehan1 »

dfabulich wrote: post #294
lynndotpy wrote: post #24

Any social organization needs to carefully consider their inclusion-exclusion curve with intentionality. I think a lot of people might balk at the word "inclusivity" today, but StackExchange had ridiculously high barriers to participation, making it inclusive to the long-time users on the site, but exclusive to the newbie participants who found themselves blocked for asking questions. They slowly killed the site in t…

The graph proves the cause of their decline was AI, and not aggressive question moderation. Ask yourself: in what year did it become difficult to ask questions on Stack Overflow? 2014? 2016? 2018? 2020? Aggressive question-closing was part of their design from the very beginning. Their high barriers to question-asking was the cause of their rise , as their primary user was never question writers: it was Google, and a…

And what was the alternative to stack overflow before AI?

People were desperate for an alternative, and ran with the first semi-viable one that emerged.

AI is arguably worse than SO for precision. It will hallucinate, and agree with you when you're wrong. But stack overflow is so unwelcoming to newcomers that people switched anyway.

Re: What AI did to stackoverflow in a graph

#516

Post by rcxdude »

zahlman wrote: post #467
rcxdude wrote: post #414

Earlier quoted context omitted.

Being able to edit answers isn't even a mod thing. In fact, for most of SO's life, it wasn't even a 'being logged in' thing (you could just edit an answer anonymously). How they wound up with a Q and A site where you could edit another user's answer far more easily than leaving a comment on it I still will never fully understand. (this kind of thing IMO really added to the utterly arcane set of rules and conventions…

You could not unilaterally edit anonymously; your edit would have been put in a review queue. It is good that leaving comments is hard. First off, because it was learned repeatedly, the hard way, that removing that barrier leads to ungodly amounts of literal spam. Second because even insightful comments detract from the main intended flow of using the site, which is: you find a question from a search engine, read the…

See, what you're saying makes some logical sense in a vacuum but it's so utterly unlike how any other website works that unless you take a huge amount of effort to explain this to people, gently, then it's not at all surprising that people just bounce off of it.

Re: What AI did to stackoverflow in a graph

#517

Post by pipes »

DrewADesign wrote: post #215
lynndotpy wrote: post #24

Any social organization needs to carefully consider their inclusion-exclusion curve with intentionality. I think a lot of people might balk at the word "inclusivity" today, but StackExchange had ridiculously high barriers to participation, making it inclusive to the long-time users on the site, but exclusive to the newbie participants who found themselves blocked for asking questions. They slowly killed the site in t…

Their rules, (I believe unintentionally) give iron-fisted fiefdom rulers a toolbox of justifications to control and alienate under the guise of protecting the quality of the site data. I honestly don’t even think most of the control freak mods could objectively judge the propriety of their actions because it was all encouraged by the rules. (And I do not think this was universal among the mods, but it was certainly e…

My 16 year old question was recently closed because it didn't follow the rules. It just seemed really petty.

Edit: just checked it has been deleted so no longer visible to non moderators.

Re: What AI did to stackoverflow in a graph

#518

Post by zaphirplane »

jeanlucas wrote: post #171

I don't have the time to, but I'm surprised there aren't a lot of comments on the decline before chatGPT was released, but after SO was sold to Prosus [1][2] Even the curious growth spike in activity happened just before the acquisition. I wish I had time to do this analysis a bit deeper, but you can look for SO activity up until when chatGPT was released, it is really noticeable. --- [1] Stack Overflow acquired by P…

Scrolled looking for this comment the decline appears to be way earlier than LLM agents becoming popular.

But where did all the users go is something that would be interesting

Re: What AI did to stackoverflow in a graph

#519

Post by supriyo-biswas »

jerrac wrote: post #493

My perspective: Stack Overflow was started to be a documentation site, but was implemented as a Q/A site. And that doomed it from the start. If someone wants to get help, and they end up on Stack Overflow, the structure looks like a Q/A site, and they treat it that way. The people "in the know" then accuse them of being "bad" for posting a question to a documentation site. Combine that with the gamification, and it h…

It was always intended to be a documentation site, see https://stackoverflow.blog/2011/06/13/optimizing-for-pearls-... and similar.

Re: What AI did to stackoverflow in a graph

#520

Post by psd1 »

amanaplanacanal wrote: post #484
zahlman wrote: post #466

Earlier quoted context omitted.

> An edit that made a response worse should have knocked the mod down so that they were unable to mod any more. Literally everyone on the site is permitted to propose an edit, and everyone with at least 2000 reputation can make unilateral edits. The proposals are approved by a 2 out of 3 majority of random unilateral-edit-privileged users. None of this is considered "moderation" and is not done by "mods". Of millions…

So they have essentially killed the site. Congratulations I guess?

What killed it in truth was RAG, but let's stipulate that it was policy that drove users away.

The thing that killed it was the thing that made it prominent.

According to a blogpost i can't find any more about "how to speedrun social media" or similar, it's basically impossible to pivot a community once established given the levers of voting mechanics and mod policy. (I distinguish between Facebook's and Reddit's enshittification and a pivot.)

IOW, the seed of a community property's growth is also the seed of its decline. Analagously, the charming village is attractive to new residents because the existing residents resist the changes that new residents embody.

Very broadly and by no means universally: techies don't like lusers, because lusers don't put effort into understanding, they just want turnkey answers. This is totally counter to the archetypal techie value system.

SO optimised for grumpy misanthropes over bumbling newbies, which, in tech, is roughly congruent with optimising for people with knowledge to share over people who want their homework done for them.

The questions are not the scarce resource - you can browse Microsoft Community Answers to verify that. The answers there are correspondingly useless, which is why "blindly copy from SO" is a meme and "blindly copy from Microsoft Answers" is not.

Here ITT, and more generally on HN, comments are excoriating about the cost of being offputting to new users. Those commenters should recognise that it's not a pure cost, it's a tradeoff, and the upside has been a treasure for a decade. A million projects completed by people who found quality answers, mostly by searching for existing questions.

That resource is still relevant for stacks other than the latest web framework hotness, and i personally still get benefit from it in my archaic hobby of learning to code.

That resource is not still directly relevant for coders deeply plugged into AI, but it did train that AI.

It's one of the most successful endeavours on the web and a towering community achievement. It is a superficially a shame that people experienced gatekeeping, but if "gatekeeper" has only negative connotations in your lexicon, then I advise you to compromise your principles and firewall your network.

Post reply on HN