The ChatGPT answers are only noticiable because they are higher quality grammar that the human generated low quality content, which has not been a concern for over a decade. I think ChatGPT bots answering questions with the API do not get served ads is the actual problem.
I was once in the top 0.05% of the users in the Java tag and top 1% of the users overall with a very low number account from about 8 months after stackoverflow launched.
I requested my account be disconnected from all my content and deleted my account around 2017/2018 because my highly upvoted and more importantly CORRECT answers were being modified to be incorrect by people just gaming the system for points for editing posts and the management made it specifically clear that this was not a concern for them.
I and a vocal minority of OG users had been complaining about the lack of concern for quality and the focus on quantity for about 4 years already. When they changed the tag line from "quality answers" to focus on "quantity" of questions and answers.
There was a slow by perceptible mass quitting over about 18 months and the site quality rapidly reached what it is now, because all the members with enough reputation points to actually moderate the site for free either refused to do it anymore or did like I did.
I took my volunteer time and applied it to another community that respected my efforts and time, and did not risk compromising my professional reputation.
That they now are banning answers that are only delectable as ChatGPT because they are grammatically correct and arguably higher quality in that sense is laughable.
Top comments (7)
I think the big problem with ChatGPT in SO is answers like this one. It's so detached from the question it makes absolutely no sense. A new user who might not know anything might waste their time reading that drivel.
StackOverflow is far from perfect and I get that. The politics is hard. Unfortunately, the alternatives seem worse. Google search for even basic stuff was completely gamed by SEO. It produces horrible results from companies that can afford the aggressive SEO.
that is better than most human answers. at least it is comprehensible english. majority of answers in the last decade are wrong or have nothing to do with question AND lazy incomprehensible english to the point that you can not tell if the are wrong or unrelated to the question. at least ChatGPT is not that lazy.
This hasn't been my experience but maybe I hang too much in the
codenameonetag which is pretty clean. I do see a lot of answers when I search for stuff, they aren't always great but I the top answers are usually decent in my personal unscientific experience. I do recall answers that sent me on a wild goose chase. But I would have probably been worse off if stackoverflow hadn't existed.
Can you point out a couple of bad examples so I can get a sense of what you consider a bad human answer?
I think a big reason why the internet is much less fun than before is that the venture capitalists have successfully their ideology of "go big or go home". There were plenty of ideas that were good at a small or medium scale, but that was not enough. You had to try everything to make them yuuugee. Breaking tons of good things in the process is how you play the game. killedbygoogle.com/ and its 275 and counting products are an example of that.
Slack started out as a better marketed Google Wave. Wave "failed" because it was not well defined on what it solved and why it was worth your time adopting it. The team that created it was also kind of "if you do not get it immediately, then it is not for you" as well.
Honestly, I appreciate the ChatGPT because so easy to understand and it give a direct to the point answer that make your time save, however ChatGPT have a limitations also that is the time I am going to browse from great resources like stack overflow, especially if I need more explanations. They both great but if I want some fun I'll go with ChatGPT 😁