Please support Stack Overflow moderators strike against AI content policy

[Dalija Prasnikar 1338]

Since December Stack Overflow and other sites in the network have been spammed with AI generated answers which are usually incorrect while sounding plausible. To handle this problem moderators working with other users and Stack Overflow staff enacted the policy that bans all AI generated posts. Such posts are deleted and users can be suspended (usually for a week) for posting them. To get the better picture about the impact, we are talking about thousands of users and even more posts.  See: https://meta.stackoverflow.com/questions/421831/temporary-policy-chatgpt-is-banned

 

However, last week the company enacted another policy which still allows moderators to moderate AI generated content, but effectively does not allow them to to use any means necessary for detecting such posts. In other words they can remove posts mostly if user admits post is AI generated. 

 

Allowing AI posts on sites will effectively kill the sites and elected moderators have decided to take an action and go on strike, along with other users of Stack Overflow and other sites in the stack Exchange network. Strike is scheduled to start tomorrow on Monday, Jun, 5th. 

 

Unofficial announcement of strike on Stack Overflow (there will be another announcement on the main Meta tomorrow)  https://meta.stackoverflow.com/questions/424979/what-has-happened-to-lead-moderators-to-consider-striking

 

If you have Stack Overflow or other Stack Exchange account please support the strike and sign the strike letter at https://openletter.mousetail.nl/  Signing is made by automatic authentication with Stack Exchange network account through browser if you decide to sign. You will have to enter display name you want to be displayed on the letter as some people have different display names on different sites. 

 

Thanks!

[Attila Kovacs 617]

I have SO account but as far as I'm concerned, you can strike for a lifetime.

[Dalija Prasnikar 1338]

33 minutes ago, Attila Kovacs said:

I have SO account but as far as I'm concerned, you can strike for a lifetime.

Be careful what you wish for, as you may actually get it.

Edited June 4, 2023 by Dalija Prasnikar

[Attila Kovacs 617]

It is very rare for me to wish for something that I actually do not wish for. 

[Renate Schaaf 64]

1 hour ago, Attila Kovacs said:

I have SO account but as far as I'm concerned, you can strike for a lifetime.

What makes you so negative about this? I certainly have my problems with participation on stackoverflow. After having received some unfair downvotes, I've stopped asking questions, and I only answer, if I'm dead sure of what I say, maybe a good thing.

But stackoverflow has been a valuable and mostly reliable source of information, and I would have to spend much more time finding good info, if I couldn't rely on content being mostly accurate anymore, or it the moderators would stop doing their job. I upvote every answer that has been useful, and I downvote every blatantly wrong answer, just to keep the quality up. I would hate for ChatGPT flooding the answers.

[David Heffernan 2291]

The first few times I asked questions and then answered questions on SO I did so badly and was told so. So I learnt. I think a lot of people who are criticised on SO would just better off heeding the criticism. 

 

As a moderator (not an elected moderator) I can say that SO's policies in recent years to drive quantity of posts to the detriment of quality has been dispiriting. 

 

The vast majority of users of SO don't ask. They use posts that are already there. This relies on there being well posed questions with good and accurate answers. 

 

Invariably the people complaining are asking poorly posed questions that aren't going to be of use to future visitors. In my view these questions should simply be removed. With a minimum of fuss for mods. Not doing so is the tail wagging the dog. The site should be catering to the majority of its users who are largely voiceless. Instead it caters to a tiny minority who make a lot of noise.

 

I definitely support this strike. 

[Dalija Prasnikar 1338]

Strike has officially started https://meta.stackexchange.com/questions/389811/moderation-strike-stack-overflow-inc-cannot-consistently-ignore-mistreat-an

[Roger Cigol 88]

AI is a big problem. Anyone who tries to limit it's use to intelligent purposes gets my support. I've signed.

[dummzeuch 1387]

After trying to get anything useful out of chatGPT and wasting a lot of time to fix all the errors in the answer(s) I got, I am all for banning it from SO.

 

I'm not sure it will work though. The dilution of web search results with plausible but wrong content generated by AI has already started and we will see a lot more of it. Of course there was already a lot of garbage on the web to start with, but now it has become much easier (-> cheaper) to produce it.

Edited June 5, 2023 by dummzeuch

[David Schwartz 413]

I've noticed that the vast majority of Delphi questions have negative votes, and very few have positive (non-zero) votes. If all questions with votes <= zero were deleted, there would be very few left. That tells me it has lost any intrinsic value it used to have. The Powers That Be don't like people asking questions any more than they like bots posting answers. So just shut the whole thing down.

 

Most of the questions I post get a few downvotes and attract negative remarks, and are frequently locked for reasons I don't understand, so I just delete the entire post.

 

SO has become useful to search for things, but not to get questions answered any more. Far too many people nit-pick and criticize.

 

I can't say I have a lot of compassion for the Mods on SO. They (and many self-ascribed critics) have forgotten what life was like before they were experts, and they're far too dictatorial when it comes to expecting high-quality questions from 100% of SO users. SO is intended to be useful to everybody, not just people with the precise vocabularly skills of PhDs in the selected field.

 

Here's a story that highlights the essence of the problems I typically have on SO, and where I see the greatest failure in SO serving the interests of its users.

 

--------------------------

Around 1990, I was invited to help some guys who were trying to facilitate tech-transfer from the military to commercial sector for military facilities in the AZ and NM region.

 

They basically wanted me to set up something like a BBS where military contractors could post abstracts and some details about tech stuff they were working on so people in the commercial sector could search it and find relevant things to consider licensing.

 

My expertise was software and to some extent, linguistics; I had no idea what any of the topics they listed out as examples in their requirement spec related to. Of greater concern, all they showed were a few edited specs of military tech. In three successive meetings, I asked them to show me examples of commercial uses for anything they had. Finally, they coughed up one lone example. 

 

The military rep presented a 3 page summary of some tech that one of their contractors developed along with a patent they had been granted. The title of the patent was something like this:

 

A method and means of applying long-wave infrared interferometry to estimate ballistic trajectories of projectiles targeting vehicles across a flat plain using surface-to-surface weapons.

 

I had no idea what that meant, and it didn't make much sense even after reading the brief they supplied. Then they showed me a corresponding project that a company in the LA area needed help with that proved to be a relatively perfect match -- only you never would have known it by doing keyword matching. 

 

It was a hard disk drive maker who had a requirements spec that said something like this:

Applications using short-wave optical sensors to facilitate high-resolution timing on HDD servo controllers.

 

The intersection here was they both used the same optical wavelength interferometry to measure something. One called it "long-wave infrared" while the other called it "short-wave optical sensors". One was used for ballistic calculations and the other was used in conjunction with a servo track (a small disk on an HDD spindle) for high-res timing accuracy as the disk spun.

 

I was told, "This is an example of what your design needs to do -- figure out that these use the same basic technology."

 

I asked them, "How did this match-up occur?" The guy said, "Well, we hold these tech transfer seminars around the country and invite commercial R&D folks to come in and listen to our talks to see if what we're presenting is of any interest to them. This guy came up to our rep after a talk and said he'd like to learn more." Which makes sense.

 

Then I asked him point blank, "Show me anything in these two briefs that could serve as the basis to even GUESS that the two applications are even similar, let alone a perfect match?"

 

He hemmed and hawed, and finally said, "Well, that's what we're asking YOU to do."

 

In fact, what they kept demanding I do was implement something that basically did simple keyword matches, yet they never gave me a single example that showed keyword matches that would have found anything relevant.

-------------------------------

 

The problem here, and that I find is increasingly common on SO, is summarized by a famous line from the movie "Cool Hand Luke":

 

"What we've got here is failure to communicate."

 

What the military call "long-wave infrared" the HDD servo folks call "short-wave optical" -- which the military folks would consider at the "ultraviolet" end of the optical spectrum.

 

The problem is, each one was using the idiomatic lexicon they were most familiar with, and if you try to do a match-up assuming they're using the same words to mean the same things, you're going to fail nearly 100% of the time.

 

The problem on SO is that people posting questions, including myself, often explain things in whatever manner best expresses OUR understanding of the problem. Our choice of words and how we express them might not be very precise, but it's the best we can do with what we know.

 

The Mods on SO, as well as lots of lurkers, seem to listen from a position of what I'd call "educated elite". They LOVE being critical of things that might actually make sense to them, but are not worded in a way they'd express them. So rather than rephrase the question or help the questioner improve how their question is expressed, they simply downvote the question or stamp it as "already answered" and point to some obscure post from years earlier. Much of the time, they also lock the posts.

 

Sometimes I've found their reference(s) helpful by realizing I had not made the semantic association between my question and the parallel problem domain they pointed out. 

 

However, I have *NEVER* found that the things they claimed "already answered" my question actually DID.

 

If you search for something using whatever keywords you can think of and  you don't find anything useful, the last thing you want when you post a question is some smart-alec posting a terse reply basically saying, "Hey, dummy, you were not using the right words, and here's where your dumb question was answered a long time ago!"

 

But, hey ... there we were. The post was locked and nobody was going to tell these egotistical narcissists they're full of crap.

 

The SO platform is designed to keep the Kings in place, while those who don't speak the exact same language as the Kings aren't being served very well. If you disagree, then explain why there are so few questions with positive upvotes today? There used to be in years past.

 

What happened, other than the evolution of a gross intolerance for questions that are not precisely worded?

 

As for that tech-transfer project, they got tired of me refusing to build something that did simple keyword matching and found an outfit in Albequerque to do that. Sadly, it was online for about two years and only resulted in a handful of matchups before someone pulled the plug on their funding and it was taken down.

 

Google still has the same problem today, only it does a little better job at using semantic networks to do some abstractions and even find things from time to time that are related contextually even though the keywords in the search string don't match.

 

If it takes a bunch of AI bots to de-throne the Kings who act as Mods on StackOverflow, so be it. I'm tired of having questions voted down and being locked because I don't use the exact, precise words and explanations that the Mods and onlookers want. And I'm tired of the debates that have come up that are basically about terminology when my goal is to solve a programming problem. At the end of the day, no compiler cares about the pedantics of how we humans discuss our problems. Our goal is to reduce our thoughts down to something that can be compiled and does what we want it to at run-time. 

 

 

 

Edited June 5, 2023 by David Schwartz

[David Heffernan 2291]

23 minutes ago, David Schwartz said:

I've noticed that the vast majority of Delphi questions have negative votes, and very few have positive (non-zero) votes

Evidence please.

 

24 minutes ago, David Schwartz said:

Most of the questions I post get a few downvotes and attract negative remarks, and are frequently locked for reasons I don't understand, so I just delete the entire post.

The last time I saw this happen, with one of your posts, the voting and comments were reasonable in my opinion. Instead of deleting the post, you should have improved it as suggested.

 

25 minutes ago, David Schwartz said:

SO has become useful to search for things, but not to get questions answered any more. Far too many people nit-pick and criticize.

That's the entire point. That's the design goal. As I said above, the primary use case of SO is to curate high quality questions and answers. The criticism that you experience on your questions is part of the curation mechanism. 

 

26 minutes ago, David Schwartz said:

I can't say I have a lot of compassion for the Mods on SO.

You just come across as salty that you didn't get an answer to your question. But as mentioned above, SO is not trying to answer your questions. You can just go elsewhere to find the service you need.

 

28 minutes ago, David Schwartz said:

The SO platform is designed to keep the Kings in place, while those who don't speak the exact same language as the Kings aren't being served very well. If you disagree, then explain why there are so few questions with positive upvotes today? There used to be in years past.

There are a few reasons. One of them is that there aren't as many questions that haven't already been asked. Perhaps the main reason is the policy changes by corporate SE, some discussion of that here: https://meta.stackexchange.com/questions/389811/moderation-strike-stack-overflow-inc-cannot-consistently-ignore-mistreat-an  Mods (elected and community) are just fed up with being unable to curate.

 

I also think it's important to look beyond the delphi tag in SO. Delphi is a mature tech, with not much innovation. If you want to look at how SO works, then you need to look at tags for a broad spectrum of technologies.

 

 

 

What I find ironic about the corporate SE policy changes in the past 5 years is that the community largely ignores them. They have a community run site. The vast majority of moderation and curation is done by community mods. Then there are elected mods who are bound by slightly different terms of reference. And corporate do none of the actual work. So they can make whatever policies they want, but it has little effect because the community does what it wants.

 

Corporate told mods to be more permissive of low quality posts. Community doesn't want to do that and so is in constant conflict with corporate.

 

Stack Overflow is worse than it used to be, much worse, in my view. But it's not because mods don't allow low quality posts. It's because mods aren't empowered enough to deal with low quality.

[Attila Kovacs 617]

Those self-appointed moderators who hide behind their arrogance, claiming to maintain the quality on SO, do more harm to the language than they contribute.

 

 

[David Heffernan 2291]

8 hours ago, Attila Kovacs said:

Those self-appointed moderators who hide behind their arrogance, claiming to maintain the quality on SO, do more harm to the language than they contribute.

 

 

Thanks for the vote of confidence! 

[PeaShooter_OMO 10]

Stack Overflow is and has always been valuable to me and I assume its moderators have a great hand in that. I agree that it can happen that a moderator is harsh sometimes even towards a newcomer to the site. I personally prefer strict moderation. It is wise to just adapt to its style.

 

Delphi developers on SO can be glad to have contributors like David, Remy, Andreas, Rob, Ken, Mason, Uwe, etc on there as well as here.

 

 

 

Edited June 6, 2023 by PeaShooter_OMO

[Lars Fosdal 1730]

I stopped asking or answering questions on SO years ago due to the moderation. Life is too short to bother with the aggravation.

[Sherlock 627]

I keep asking and keep getting responses, and when my questions where unclear or something else was lacking, this was communicated to me in a neutral tone (people who can hear a tone in any written statement are peculiar) and I fixed that. Almost all of my questions have been answered, and I myself have been able to provide an answer too...once or twice. So for me SO is a valuable resource that should not go to waste by allowing fake answers.

 

14 hours ago, Attila Kovacs said:

Those self-appointed moderators

Just how exactly does that work? Did they hack Stack Exchange to gain mod privileges? And Stack Exchange can't be bothered to check?

[Uwe Raabe 1879]

30 minutes ago, Sherlock said:

Just how exactly does that work?

With a high enough reputation you have some privileges like elected moderators have.

[Dalija Prasnikar 1338]

Just now, Uwe Raabe said:

With a high enough reputation you have some privileges like elected moderators have.

Yes, but you don't get there by self appointing yourself, either.

[Dalija Prasnikar 1338]

Strike update: https://meta.stackexchange.com/questions/390106/moderation-strike-update-data-dumps-choosing-representatives-gpt-data-and-wh

[Dalija Prasnikar 1338]

The whole situation at Stack Overflow just made a drastic turn for the worse:

 

https://meta.stackoverflow.com/questions/425162/we-are-seeking-functional-feedback-for-the-formatting-assistant and https://meta.stackoverflow.com/questions/425081/stack-overflow-will-be-experimenting-with-a-question-formatting-assistant-and-w

 

It seems that whole reason behind tying the mods hands when it comes to detecting and moderating AI content is related to the brand new "AI Formatting Assistant" It is presented as something that would help people write better questions, but I guess that is just first step and second one would be expanding it to answers, too. Because such "formatted" posts would have typical AI signature, moderators would not be able to distinguish between AI generated and "merely formatted" posts.

 

Additionally, the results of "formatting" go beyond mere formatting because AI completely rewrites the post including text and code adding its hallucinations on top. It is beyond horrible and harmful.  

[David Heffernan 2291]

They should use AI for moderation

[Attila Kovacs 617]

Quote

 typical AI signature

The fox and the wolf meet in the forest. The fox says:

Hey! Let's beat up the rabbit.

Okay, but why? - asks the wolf.

If he's wearing a hat, then for that reason, and if he's not wearing one, then for that reason! - replies the fox.

They go to the rabbit and beat him up mercilessly. The next day, the fox starts again:

Hey! Let's beat up the rabbit.

But we already beat him up yesterday! - responds the wolf.

But today, let's beat him up again! - insists the fox.

Alright, but why should we do it today? - asks the wolf.

We'll go to him and ask for a cigarette. If he gives us one with a filter, then for that reason, and if he doesn't, then for that reason.

They approach the rabbit: - Hey, rabbit, give us a cigarette!

Do you order a filtered or non-filtered one? - asks the rabbit.
To which the fox exclaims:

Look at that, for heaven's sake! Once again, he's not wearing a hat.

[Dalija Prasnikar 1338]

11 minutes ago, David Heffernan said:

They should use AI for moderation

Don't give them more bad ideas.

 

It is not that AI cannot be used for moderation, but at this point I am pretty sure it will be used in wrong way.

[Dalija Prasnikar 1338]

48 minutes ago, Attila Kovacs said:

The fox and the wolf meet in the forest. The fox says:

Hey! Let's beat up the rabbit....

I have no idea what you are trying to say. 

[Uwe Raabe 1879]

15 minutes ago, Dalija Prasnikar said:

I have no idea what you are trying to say. 

I guess, whatever the real intention is, one can always make up a fake reason to do so.