What AI did to stackoverflow in a graph

> It's just not that important to me for a dead site. I get it, that means "just trust me bro" is in play.

See, this is the part where I'm actually getting insulted by the invocation of the "bro" language. No; what's happening here is that my lived experience as a curator of the site (which was quite stressful) is being denied, and it is supposed that evidence to counter me is not necessary.

This isn't about "winning"; this is about me wanting you to understand something that you apparently don't understand. Again, it's hurtful to have that motivation ascribed to me.

> SO was filled with this sort of "technically this is a duplicate and you are just a nasty rule breaker" style comments with litigation that ultimately goes nowhere.… Others experienced and are reporting here and elsewhere experiencing the harsh moderation of SO.

First off, I want to make this extremely clear: posting a question that gets closed is not, and was never, considered "breaking the rules". When people close a question, they are not moderating the site, but curating it; which is exactly why the overwhelming majority of such actions are taken by non-moderators. People are wrong and ignorant to call them mods, and it's hurtful that they don't take the time to understand this. I understand perfectly well that people felt that these actions were harsh, or even cruel. Many of them were not executed properly. But nobody ever takes the time to understand my perspective; they assume many false and hurtful things about me (and others) and refuse to be corrected on those assumptions. Worse, the site software, and the company, actively interfered with the intended curation of the site in many different ways over the years (changing according to whatever the company was currently trying in order to be profitable).

There is nothing "technical" about things being marked as duplicates. There is a clear standard for it which was reasoned out according to a very specific purpose: people who search the site should find the best possible version of the Q&A, so that they can directly get the answer to their question.

The only time that "litigation goes nowhere" is when people come to the meta site with the assumption that the rules do and ought to work the way they expected them to, and refuse to accept that they actually work differently. Stack Overflow was (and still is, to the extent that anyone remains) a community, and communities are supposed to have a right to self-determination. It's not the fault of the curators that the site presented a misleading interface, to the extent that that happened; they had no power to modify the interface beyond making more Q&A posts on the meta site.

If thousands of Americans travelled to a tiny country where cars drive on the left, and all tried to drive on the right as they do at home, would that make the country in the wrong? Would it be a "serious endemic problem" with the country?

> That's ridiculous and you're wrong. I grant them a CC license to my content to fold, spindle and mutilate it for their own purposes, but not to leave my name attached to the folded, spindled, mutilated version.

The choice of license is made clear at https://stackoverflow.com/legal/terms-of-service/public#lice..., and the consequences elaborated at https://meta.stackexchange.com/questions/347758/creative-com... . The license itself is at https://creativecommons.org/licenses/by-sa/4.0/ . It is one of the BY licenses, so the site is legally required to leave your name on the work. The interface explicitly marks both the original author and the most recent editor on the main facing page; and it offers the ability to click in and see the entire editing history.

> But I've had people add extra sentences or paragraphs, and oh hell no.

This is overwhelmingly done in good faith, and it's explicitly policy not to change your meaning. In fact, there are many posts on meta where people complained about attempting a good-faith edit and having been judged as falling afoul of that for no reason. See e.g. https://meta.stackoverflow.com/questions/349950 .) But it's also explicitly policy that they can "edit your words, attributed to you" (see e.g. https://meta.stackexchange.com/questions/403176 ). The people who make these changes, generally, sincerely believe that they are either saying things you intended to say, or at least adding information that you would have added if you'd come back to keep things up to date. There has been a ton of discussion trying to clarify and/or figure out where the line is: see e.g. https://meta.stackoverflow.com/search?q=%5Bediting%5D+intent... .

If you saw someone change something that fell afoul of this, there was a meta site where you could raise the objection. But we were constantly inundated with people who thought it was wrong that we took "thank you in advance, my_username" off the end of the their questions or similar (see e.g. https://meta.stackoverflow.com/questions/260776 ), as if they thought they had a right to that.

Your claim is that "quite literally every person you interact with professionally has an anecdote about" the internal mental state of other people.

Even if everyone else on the Internet thinks a person has a certain belief or motivation or ideology, that doesn't mean that person actually has that belief or motivation or ideology.

I'm speaking to my own mental state, as well as to my general impression of what other curators told me, and the frustrations they espoused. There were chat rooms where people could have easily spent all day talking about how cool it is to power-trip, if they'd actually been power-trippers. Instead, they talked about how helpless* they felt as new people kept coming in and making the same innocent mistakes; about how they had no tools to deal with this except the ones that left offsite users with a bad impression; about how annoying it was to be constantly aware of that offsite bad impression when they were trying to do things necessary to the site's objective; etc.

What you also seem to be missing is that since the Prosus acquisition, "SO mods" (and the rest of the community, who are much greater in number and take the large majority of the actions incorrectly deemed "moderation" by outsiders) and "SO staff" were in active conflict most of the time. The LLM-content ban policy was against the staff's (really the company, who employ a handful of staff as liaisons) wishes; the fighting over this got so bad that there was an actual moderation strike and an extended period where the site got the amount of spam you'd expect for a site that size instead of the almost miraculously small amount it would normally get. (The spam filtering also involved community-operated automated systems that were suspended in solidarity.) Then we got hit with proposal after proposal for AI-powered features that were obviously terrible half-baked nonsense with no real value ad; we constantly tore into them on the meta site; the company showed zero sign whatsoever of caring in the slightest, while Prashanth Chandrasekhar (the new CEO) would give TedX talks and such demonstrating a complete ignorance of how the community felt about AI in general.

Atwood and Spolsky were quite explicit, on multiple occasions, about what they wanted the site to be. That vision was not compatible with having a site with millions of users and millions of questions, because the masses simply aren't interested in the kind of interaction that the vision required. They may not have understood this at the time. But it's clear that the early community attempted a lot of gatekeeping, much of it misguided in the details (you want lots of questions that address beginner-level problems, because most people in a field are beginners at any given time for any given field, so that's what's most useful) but directionally correct (most of the time, you don't want the question that a beginner asks, because the beginner is not adept at asking useful questions and lacks the perspective needed for a proper Q that allows for presenting a useful A).

The intent was very much to build a library. It's literally on the tour page at the top:

> Stack Overflow is a question and answer site for professional and enthusiast programmers. It's built and run by you as part of the Stack Exchange network of Q&A sites. With your help, we're working together to build a library of detailed, high-quality answers to every question about programming.

That is not something inserted by people on the meta site. We literally did not have access to make that kind of change. Believe me, there are so many other things we would have fixed if it had been made possible.

Most of the important policy was established around 2013 or 2014 and those are the dates of the most important posts on meta that get brought up again and again (and used as duplicate targets, and cited in arguments). That was also right around when site activity was peaking. Atwood and Spolsky didn't leave until the acquisition in 2019.