Hacker News Reader

Comment by AdminAccount | Hacker News Reader

I think the only reason stackoverflow still has any activity is because the community choose to ban AI content [1] and so did most of its other networks [2].

Perhaps it will even see a (small) resurgence when AI providers start charging for the actual costs.

[1] https://meta.stackoverflow.com/questions/421831

[2] https://meta.stackexchange.com/questions/384922

View on HN

Considering StackOverflow is now providing a ground truth for AI training, I believe the ban is more about not poisoning the well rather than keeping the StackOverflow or StackExchange human-friendly.

That ship has sailed long time ago with zealot admins and verbal harassment.

> That ship has sailed long time ago with zealot admins

Where there are certainly strong examples of this, a lot of people mistake enforcing the rules as zealotry. Part of the point of SO was that if things don't change then there is a completed state for SO too - no need to ask duplicate questions like on platforms where a post is less long-lived. Unfortunately people take things like “this is a dup”, “provide more information as we can't help”, “this isn't a complete answer”, and so forth, as deeply personal attacks…

One of the good things about LLMs is that they've drawn off all the simple already-answered questions! Unfortunately the more complex ones, or the ones for new solutions, are also going there so SO and its family of sites is ceasing to grow even in the ways it wants to.

> and verbal harassment.

Again, that did/does happen, but a lot less than some people report it. The most abusive people I've seen on there are those who have been given one of the responses I listed above.

You're absolutely right. The problem was so small that SO only had to make a site-wide survey, make a couple of public statements, big administrative changes and a big campaign to earn hearts and minds back.

Even after that, I still feel sour about the site. Talk about burnt bridges.

Moreover, in 2025, 46% of the survey participants told that they don't feel like part of the community. That number was 44% in 2024 [1], too. Also, 2023 doesn't look better: 45.63% of participants said no [2].

Maybe the survey is rigged. Who knows?

[0]: https://survey.stackoverflow.co/2025/stack-overflow#2-feel-l...

[1]: https://survey.stackoverflow.co/2024/community#2-feel-like-a...

[2]: https://survey.stackoverflow.co/2023/#section-stack-overflow...

> don't feel like part of the community

Maybe I'm off because I wasn't there for community, I don't really do the online community thing. I was one of the biggest contributors in terms of answers to a couple of the areas (SU, SF, DBA) in the early days and I liked helping people, but I certainly wasn't there to make friends.

I also don't get on with remote work because I don't feel connected to names and faces on screens in remotely the same way as I do to people in the same room, so like I say maybe my personality missed that part of the problem as it doesn't properly appreciate when things are apparently the other way. Or maybe people were looking for the wrong thing in a technical Q&A site that wasn't, as far as I felt, trying to be a social media site.

For technical questions and to help out with similar answers I'd go to SO, SF, SU, DBA, ... For community: the local, the running club, martial arts, or even sometimes DayJob.

> I liked helping people, but I certainly wasn't there to make friends.

SO was not structured to make friends, and I didn't expect anything similar to that end. However, I expected a more friendly (or neutral) conversation, like between colleagues or two people who just met in a conference.

Being put down for programming style, language choices or other nitpick is just bad. I have seen enough flamewars and rude people in my life. I wanted to ask a questions and I got low-key insults or "duplicate!" warnings for things not duplicate, in fact.

I also like learning and helping people, but I don't like to be bullied. I endured that thing enough (5+ years if anybody wonders). I neither want more, nor have time for that, nor young enough to don't care for that.

So, I reverted to consulting documents and finding my own path. At worst, the code doesn't work and/or compiler shows what I did wrong. Even GCC 4 is more polite than some of the people in SO or SE.

> You're absolutely right. The problem was so small that SO only had to make a site-wide survey, make a couple of public statements, big administrative changes and a big campaign to earn hearts and minds back.

The SO population decline started before LLMs, too.

Yep, the first killer app for AI was, "hey, I'm having this problem and I'm not sure what's going on, can you help me figure it out?", and being told "sure!" instead of "this is a dupe, use the search function you idiot" or "this is not the right kind of thing to ask here". SO has always been incredibly useful and incredibly frustrating at the same time.

You have three options.

1. Ask a stupid question online and be called an idiot in your public record.

2. Go through the hassle of creating the nth throwaway account and get called an idiot in public, if you are allowed to post with a new account at all.

3. Ask ChatGPT and get told you are completely right.

My favorite thing with Claude lately is it tells me things like:

> I think this is a great thought, but I want to engage with it seriously rather than just agree, so for the sake of argument, ...

I keep telling it to be blunt, that I can take it, but it just can bring itself to offend me :)

The fun part was not searching before asking. I always did that. Also, not being told to ask better questions. We need to ask good questions and read ESR's guide on that.

The fun(!) part is spending an hour searching the site and meticulously crafting a question which ticks all the SO guideline boxes, and being yeeted out for a "duplicate question which even doesn't meet the guidelines".

At least give a link to the well-aged question, so I can see what I missed. Of course the answer is no, because there's no such question and/or answer.

People also always bring up the "fake XY problem" thing on SO as a sign of toxicity or whatever, but I’ve had many, many results where it was a XY problem, and the actual problem Y was solved, yet I landed there searching for a solution to X :/

It's the deletionism that effectively ended my participation in stack exchange networks. Violated the rules with your question? Closed. Alright, I can deal with it. But they didn't stop, they just had to go and delete the goddamn post, along with all the useful comments and links that had built up on it, for literally no reason. There was actual content in there.

Seriously, fuck that. Went all in on my own website after that. Can't delete my stuff now.

Their rules said duplicates should be deleted, at least most of the time, so they were. While it might be nice for extra useful content from the dup to be merged into the original page, that would be a lot of work and much of the time there isn't enough to make it worthwhile (or any).

> for literally no reason

Other than the reason that it was their site and that is what they chose to do.

While some criticism of SO is certainly valid, I find a lot of it boils down to "how dare they run their site the way they want instead of the way that I would prefer".

> Went all in on my own website after that. Can't delete my stuff now.

This is a good solution. Everything I've written on there that ended up being something I might want to preserve, and similarly useful answers to mine and others queries, I have archived for future reference too¹, in case the sites die, or that particular content does, or otherwise becomes inaccessible to me.

--------

[1] though in my case not online as I never got around to that, maybe I will one day, when I've been sacked for being inefficient because I refuse to have Claude do my job for me and I have plenty of time on my hands!

> Other than the reason that it was their site and that is what they chose to do.

Welp, they can go to hell along with their site. AI is better than stack exchange anyways.

Everybody's got boundaries, and they managed to cross mine. I can deal with rudeness and toxicity just fine. I can deal with moderators enforcing the rules. Straight up deleting stuff is simply not acceptable though, and I couldn't care less how they justify it.

You might be happier in life if, having decided that you didn't like how things were done there, you moved on and got on with things elsewhere, rather than moving on but still being bitter and complainy about it to this very day.

It's not bitterness, it's anger. It can be a rather enjoyable and cathartic emotion, especially when one is confronting perceived wrongs.

Another benefit of sharing experiences, especially negative ones: it can help others. Maybe there are people on this site who aren't yet aware that stack exchange embraced deletionism. I certainly wasn't, wouldn't have contributed otherwise. If I help even one person realize this, it was worth it. That realization could prevent them from having their work closed down and deleted before they had even finished writing it.

The AI companies aren't so deep in the red when you only look at inference though - they are investing loads in new models in an AI arms race.

So I don't imagine AI is going to go away, especially given that now there are more open source models like Qwen that you can run locally. So even if those American behemoths go bankrupt it will persist.

> The AI companies aren't so deep in the red when you only look at inference though - they are investing loads in new models in an AI arms race.

Depends on how you're looking at it (using speculated numbers for easy math):

1. Having operating costs of $100m on revenue of $10b is very deep in the red, regardless of training costs.

2. Having $90m training costs on $10m revenue means they're just breaking even.

Problem is, we don't know their financials and how it is broken down (they could, of course, clear up the confusion and release some numbers, ut they aren't doing that now); all we know is when they need a new raise to continue operating.

From the raises we can determine what their operating costs are (For example, raised $30m in 2024, then $300m in 2025 is a 10x increase in operating costs because they aren't spending on capex. The training is done on opex).

From their subscriptions (which are all only estimated), we can sorta tell what the revenue is, but that's for subscriptions only which are almost guaranteed to be running at a loss (until recently, anyway). We don't even have estimates on revenue from the PAYG API users. Common sentiment is you'd be a fool to use the PAYG options for anything but trialing the service, but the world is filled with fools, so you never know!

What is interesting is comparing the prices for PAYG on the providers supplying open models vs the PAYG on the closed models - the suppliers providing open models aren't spending on training cost, so the cost to supply tokens on open source models is pretty close to the actual price of running models. This is partially confounded by the fact that many of these will have VC money backing them (they are not bootstrapped), and so will also try to perform landgrabs via subsidised tokens, because their goal is an exit with a buyout, and without an eventual acquisition they will simply fail.

I can't think of many open source model suppliers providing subscriptions, not ones that subsidise the subscription, at any rate.

The first IPO of these SOTA providers is going to be the eye-opener; we'll finally see their financials and we'll see just how much the PAYG was subsidised, and how much the subscriptions were subsidised.

Until then, with a collective industry investment of $800b (last I checked) and a collective revenue of $20b (last I checked), they are most definitely operating in the red for the most common definitions of operating in the red.

$800b (last I checked) and a collective revenue of $20b (last I checked)

Promised investment is not the same thing is operating expenses. Only a small fraction of that $800b has actually been spent so far.

I keep hearing that AI providers will eventually stop "subsidising inference", but I do not understand how this would make any sense economically. It is already profitable to host AI model inference for prices cheaper than the big labs are charging, as evident by the plethora of providers available to choose from on OpenRouter, for example. The only way prices could rise without competition is if you sidestepped supply and demand.

The hosting in it of itself is relatively cheap however with the current attempt by pretty much every major player to outpace one another means they are heavily investing in R&D/ model training, related infrastructure expansion, personnel recruitment/poaching etc. This all makes (to my knowledge all) major model provider run at a net loss despite the huge revenue some have.

You keep hearing it because AI haters won’t stop screeching about it, despite all the evidence that they’re wrong. Denial is a helluva drug.

I hope you're right. AI is world changing technology, I don't want to live without it anymore.

Reply on news.ycombinator.com