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

We might have been slower to abandon StackOverflow if it wasn't a toxic hellhole

pcloadletter.dev

Post reply on HN
Print view View original

61–70 of 225 posts

Re: We might have been slower to abandon StackOverflow if it wasn't a toxic hellhole

#61

Post by Eisenstein »

lanyard-textile wrote: post #31
jrjeksjd8d wrote: post #13

Like so many things in American life these days, we've arrived at a "better solution" that extracts value from the past without producing a future. The decades of Stack Overflow answers fed into LLM training produce plausible answers for today . In 10 years what will the LLMs train on? So much of life these days is purely extractive, trying to squeeze more money out of less productive activity. It's no wonder young p…

Why, us of course. The results of your claude code session, for example, make fine training data. Did the user commit the final answer? What changes were made before they did?

> The results of your claude code session, for example, make fine training data.

Does Claude code copy your repository onto its server?

Re: We might have been slower to abandon StackOverflow if it wasn't a toxic hellhole

#62

Post by closewith »

Permit wrote: post #32

StackOverflow was successful explicitly because of the people/question it excluded . The "toxicity" was the point. It was trying very very hard not to become Yahoo Questions. If you want to hear more about this you should watch Joel Spolsky's talk "The Cultural Anthropology of Stack Exchange": https://www.youtube.com/watch?v=LpGA2fmAHvM The point of StackOverflow was explicitly not to help the question-askers, but to…

> StackOverflow was successful explicitly because of the people/question it excluded. The "toxicity" was the point.

This is a very charitable read of the situation. Much more likely is, as another commenter posted, a set of people experiencing a small amount of power for the first time immediately used it for status and took their "first opportunity to be the bully". Question quality and curation was always secondary to this.

> > The point of StackOverflow was explicitly not to help the question-askers, but to prioritize the people who would reach the question via Google.

It obviously was only tolerated because of that, as evidenced by the exodus the moment a viable alternative became available.

Re: We might have been slower to abandon StackOverflow if it wasn't a toxic hellhole

#64

Post by ourmandave »

throwway120385 wrote: post #8
MattGaiser wrote: post #2

I'm surprised at how non-obvious LLMs are for that decline chart. SO peaking in 2014 is not what I would have expected.

I'm not. A lot of moderators on SO thought that technology stayed the same from 2014 onward. So I'd often see questions that were exactly the same as the one I asked, but they'd be closed as a duplicate of a question that was subtly different in wording but totally different in technical content. And you would have had to be intimately aware of the systems the question was about to recognize the difference. Most of t…

That's what's odd to me. When you're composing a question it will dynamically show links to related answers. So the author has already seen possible duplicates, right? Yet them seem quick to play the Closed for Duplicate card.

The other thing is it will say, "New Contributor, be nice." and their question will have -5 or more downvotes. I think that's because the auto closer needed -5 to trigger or something. I could be wrong. Either way, to a newbie it just looks like they're piling on hate for no reason.

Re: We might have been slower to abandon StackOverflow if it wasn't a toxic hellhole

#65

Post by amoe_ »

spiderfarmer wrote: post #54
ndriscoll wrote: post #37

Earlier quoted context omitted.

Why wouldn't the LLM just read the source of the framework to answer questions directly? That's how I do things as a human. Given the appropriate background knowledge (which current LLMs are already extremely capable with), it should be pretty easy to understand what it's doing, and if it's not easy to understand the source, it's probably a bad framework. I don't expect an LLM to have deep inbuilt knowledge of librar…

LLM’s can already do this without training. I recently uploaded a manual for an internal system and it can perfectly answer questions from the context window. So adding a new framework already doesn’t need human input. It’s artificial intelligence now, not a glorified search engine or autocomplete engine.

Which particular LLM was that?

Re: We might have been slower to abandon StackOverflow if it wasn't a toxic hellhole

#66

Post by alistairSH »

okaleniuk wrote: post #24

There should be a lesson for other communities but, unfortunately, there is none. StackOverflow used to be a fun and welcoming place in 2009. It became a toxic hellhole overtime There are no obvious flows in the original design, and there were no endemic wrongdoings in the governance either. It just rotted slowly like any other community does. And nobody in the world knows how to keep communities from becoming toxic.…

And nobody in the world knows how to keep communities from becoming toxic.

Mostly true, but there are exceptions... HN is about as good as I've seen for a publicly accessible forum, but has very active moderation (both from Dang and team and a pretty good vote & flag mechanism).

The other good forums I've seen are all private and/or extremely niche and really only findable via word of mouth. And have very active moderation.

But, yeah, I think you're probably right for any sufficiently large forum. It'll trend to enshittification without very active management.

Re: We might have been slower to abandon StackOverflow if it wasn't a toxic hellhole

#67

Post by darkwater »

brigade wrote: post #42
Permit wrote: post #32

StackOverflow was successful explicitly because of the people/question it excluded . The "toxicity" was the point. It was trying very very hard not to become Yahoo Questions. If you want to hear more about this you should watch Joel Spolsky's talk "The Cultural Anthropology of Stack Exchange": https://www.youtube.com/watch?v=LpGA2fmAHvM The point of StackOverflow was explicitly not to help the question-askers, but to…

As someone who did predominately use stack overflow through Google search… I remember that half the time the top result was someone asking the question I had, only for it to be duped to a different question that didn’t answer the original. So they failed there.

I question the "failed" here. You did land on their pages, after all. You most probably also clicked on an internal link and moved to another of their pages, and then bounced off.

Re: We might have been slower to abandon StackOverflow if it wasn't a toxic hellhole

#69

Post by simianparrot »

smallerfish wrote: post #59
simianparrot wrote: post #47

Earlier quoted context omitted.

And they get confident lies back 2/10 times, not enough to make them constantly question the answers, but just enough to _really_ mess with them over time. And not just that, almost no new good questions and answers are created anymore for the next iterations of LLM's to train on, so they'll be consuming more and more AI-centered SEO slop until that ratio becomes so high it becomes useless. But by then most of us who…

> And they get confident lies back 2/10 times That _might_ have been true in 2023.

If you think it's not true today, it's just the Gell-Mann amnesia effect.

Re: We might have been slower to abandon StackOverflow if it wasn't a toxic hellhole

#70

Post by eunos »

jrjeksjd8d wrote: post #13

Like so many things in American life these days, we've arrived at a "better solution" that extracts value from the past without producing a future. The decades of Stack Overflow answers fed into LLM training produce plausible answers for today . In 10 years what will the LLMs train on? So much of life these days is purely extractive, trying to squeeze more money out of less productive activity. It's no wonder young p…

> . In 10 years what will the LLMs train on?

Hopefully Reasoning is much better that training on original code and API docs are sufficient

Post reply on HN