Hacker News Reader

Comment by johnfn | Hacker News Reader

Hm, not sure I agree with you about SO. The experience I have with SO 99%+ of the time is that I google my question, get a result on SO which has a correct and decent quality answer, and I'm back to work. That's a pretty great experience, and I certainly wouldn't say that it misses the bar to be a 'quality platform'. I do agree the moderation is overbearing at times, but let's not throw the baby out with the bathwater here. The utility of SO - getting answers to questions - is basically unmatched websites that aren't LLM wrappers.

While you won't hear me bemoaning the death of Quora, I'm quite a bit more concerned about SO. After all, where is GPT 5 going to scrape the next set of answers for new libraries and frameworks..?

View on HN

> The experience I have with SO 99%+ of the time is that I google my question, get a result on SO which has a correct and decent quality answer, and I'm back to work.

You and I are talking about different things. Before these great questions and answers ended up in your search result, someone asked them, and someone else provided a good answer.

I used to be one of these people. I got 50k+ SO karma, mostly by asking good questions. I no longer bother, because SO moderators don't even read your questions anymore before closing them for a random reason. Needless to say, I no longer bother trying to answer questions as well.

So, it's no wonder that SO is struggling. You can't thrive forever on a stock of aging questions.

> SO moderators don't even read your questions anymore before closing them for a random reason

Note that most of these are just regular people with close-to-vote powers, not elected or appointed moderators with special powers.

I think one major problem is that there is an entire class of people who rarely or never ask or answer questions, or even comment, and all they do is "moderate" the site by closing questions. Literally all they do. Some of these have extremely specific and narrow views on how the site "should" be. I absolutely hate it: who the hell are you? You're not even using the core function of the site. Fuck off trying to tell me how I "ought" to be using it. I don't want to gatekeep who is or isn't "part of the community", but people gatekeeping how the site can be used without actually using the site is just absolutely toxic.

There are a number of other issues as well. I can go on for a long time. But to be honest I no longer care: the site has been taken over by nihilistic capitalists who care not one iota about any aspect of the site other than the ability to earn a buck (previously it was a commercial enterprise as well, sure, but it wasn't 100% about earning a buck and many in leadership positions genuinely cared about "doing right by our community" as well). And that is probably just as much of a reason for the decline of Stack Overflow as anything else.

My specific work community has a Slack with a few thousand people and a few active hundred people. We've come to the conclusion that slacks 90day retention message policy is a good thing. It allows new and old to reask and reiscussion topics.

SO should do something similar. Throw out all the mods, all the questions and start fresh every X number of years. No idea if it will work but tossing out the current mods to bring in new ones would change the flow.

So that's the other extreme end, and that would get repetitive for the people answering questions. Because in spite of the overzealousness of marking questions as duplicate from some people, some questions really do get asked over and over and over again.

Also: Stack Overflow is intended to be a long-term useful repository of question and answers. You enter "how do I frob a baz in foo?" in $search_engine, and the idea is you'll end up on Stack Overflow which answers that exact question. I have sometimes ended up on some of my own answers from years ago like this.

It would be for sure an interesting experiments. Repetitive? No not necessarily. The refresh would happen could periodicaly, but the core functionally wouldn't change.

The old answers can no longer be used as the basis for duplicates, a new one must be created ( that brings a slew of problems, such as sniping for getting magic points).

But the refresh of PEOPLE (mods) is where its needed the most.

What is wrong with duplicate questions though? If the response - yes but my use case is a tad different and its not covered by the linked solution, then dialogue should flow not restricted

funny how their original prime directive was to fix expert sex change, and they were so successful they caused experts-exchange to (somewhat) improve their game, and at the same time needed to enshitify SO to hit VC-fueled objectives that had nothing to do with building the best place for developers to ask questions and get durable answers.

Its probably been over 100 times I find someone asking the question im trying to find only to find that a moderator marked it a duplicate and links to a much older question that doesnt apply.

> where is GPT 5 going to scrape the next set of answers for new libraries and frameworks..?

The bottom line is that it doesn't matter as long as you have a large enough sample to learn the format, which is already there with existing data. There isn't an SO answer for everything anyone needs even about current tech, and the reason models can still answer well in novel cases is because they can generalize based on the actual source and implementations of the frameworks and libraries they were trained on.

So really, you only need to add the docs and source of any new framework during pretraining and bob's your uncle, cause the model already knows what to do with it.

> After all, where is GPT 5 going to scrape the next set of answers for new libraries and frameworks..?

From its own incorrect answers that got parroted around the web.

In the grim darkness of the future, we deliberately base the APIs of new libraries and frameworks on the hallucinations of the most popular LLMs.

> After all, where is GPT 5 going to scrape the next set of answers for new libraries and frameworks..?

with open source code, it can generate docs, feed those docs in on the next training run, use that knowledge to generate que and answers. With tool use, it can then test those answers, and then feed them into the knowledge base for the training run after that.

The problem with SO is that obsolete answers are never marked as such. You might find the best way to do X in 2015, but perhaps that's worse than finding nothing.

The problem with marking answers “obsolete” is that many code bases are old. You need those answers.

Obsolete answers shouldn't be removed. They should be de-ranked and marked as such, but still available without extra clicks.

And pointed to newer ways of doing it

This is especially true for the older frameworks like rails.

Half the answers are for rails 2.0 and the other half tell you to just “install this gem” which monkey patches some part of the framework that has long since been deprecated/removed

That's fine until you have to ask a question that isn't answered anywhere on the internet. Then you post a perfectly reasonable question and get it downvoted to hell...

https://stackoverflow.com/q/79461875/265521

That's just the latest one I've asked. Here are some more examples:

Case insensitive string comparison is "opinion based": https://stackoverflow.com/q/11635/265521

How to catch Ctrl-C "needs more focus" (this was closed but has since been reopened): https://stackoverflow.com/q/1641182/265521

This reasonable question had 13 downvotes by power-mods but has climbed back to positive when discovered by actual users: https://stackoverflow.com/q/41015509/265521

Another example of idiotic downvotes. This started off at -3: https://stackoverflow.com/q/79050597/265521

It's a common pattern that questions get a lot of downvotes initially from people trawling new questions who see a lot of genuinely bad questions (seriously there are loads), then see a good question that they can't understand in 1 second so they just downvote/close it too. So you quickly get downvotes and then later you get people coming from Google who are actually looking for that thing that upvote it.

I think SO actually did try to improve matters once. I can't find it now but they were going to make it impossible to go below 0 votes and allow one free "reopen", or something like that. But the power mods absolutely hated that idea and SO sort of depends on them so they chickened out. Now they're paying the price.

I think it depends a lot on how research AI generated traffic counts towards these stats.

Also, if there is no answer yet on the web the AI may also not know it. Then these questions should still end up on SO.

I might add, that SO also could build their own chat / research UI. It would need to have some benefit over others, but I guess the community aspect of it alone would suffice...

> After all, where is GPT 5 going to scrape the next set of answers for new libraries and frameworks..?

Some of it will be from github issues, I find it a good Q&A place now for some newer / updated techs than SO

Yeah SO was great. Despite all the complaining, all developers I know used it daily. Of course now it's obsolete since GPT extracted it's entire dataset & provides a much better way of finding what you need. They should honestly sue the hell out of them, they deserve at least some royalties for the developer subscriptions from Claude & OpenAI.

The SO data was always licensed as CC-attr (I think?) so LLMs could and should use it. The SO value was never supposed to be the data; it was the platform and community, and this is the part SO (VC-fueled version) has destroyed.

I'm just glad Jeff and Joel got their payday. Jeff really deserved to win the internet lottery, and on the whole Joel was a net positive for the internet and my career personally.

You're describing SO experience 3 years ago. I don't use SO at all now lol

Reply on news.ycombinator.com