Timeline for GPT on the platform: Data, actions, and outcomes
Current License: CC BY-SA 4.0
10 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jun 22, 2023 at 22:44 | comment | added | Era | @jcs I don't understand why your speculation is relevant. | |
| Jun 8, 2023 at 18:43 | comment | added | jscs | As I said, "regardless of whether they're correct in their perception that they are so competing". It doesn't matter what the GPT posters are actually doing, it matters what the leavers think. | |
| Jun 8, 2023 at 16:12 | comment | added | This_is_NOT_a_forum | @jscs: That is an interesting theory, but, like regular plagiarism, most ChatGPT answers are not on new questions, but on very old questions (mostly the popular ones with many existing answers) or relatively old (bountied questions). They want to avoid the scrutiny on the new questions (where most of the attention is)—read instant downvotes. They hope for the stray upvotes. And in some cases a virtuous circle (the first upvote can start the snowball rolling). | |
| Jun 7, 2023 at 23:11 | comment | added | jscs | Absolutely this. The OP even says quite clearly that the attrition inflection timing matches the introduction of GPT, not the ban policy: "In total, the rate at which frequent answerers leave the site quadrupled since GPT’s release." The simplest explanation that occurs to me is that these answerers don't want to compete with GPT posters (regardless of whether they're correct in their perception that they are so competing). I don't have any evidence for that, but neither does the OP disprove it. | |
| Jun 7, 2023 at 22:28 | comment | added | TylerH | @user1937198 FWIW I agree with the argument that some bad data analysis is being made here. Unfortunately Stack Overflow doesn't seem to employ data scientists anymore' by the OP's own admission, they are engineers and community managers trying to perform data science analysis of statistical data. My comment(s) above are simply trying to address Thomas' question of 'what are you trying to solve here', based on my interpretation of the question/announcement. | |
| Jun 7, 2023 at 22:27 | comment | added | TylerH | @ThomasOwens Yes, it is a long-term trend, but the post above mentions that, and focuses on how, since ChatGPT, and coincidently the enactment of insta-banning chatGPT users, the negative growth rate has become significantly worse. That's presumably why they are running around like their hair is on fire trying to fix it. | |
| Jun 7, 2023 at 22:17 | comment | added | user1937198 | @TylerH What they haven't shown, which seems a fairly obvious hypothesis to me, is that both decrease in user growth and suspensions, are not co-morbid symptoms. Ie, they correlate, because both increase in suspensions and decrease in user growth are directly caused by the growth in use of ChatGPT. | |
| Jun 7, 2023 at 21:38 | comment | added | Thomas Owens | @TylerH If the problem statement is something more like "the network is experiencing slow or negative growth", then I have company-provided analytics that show that the problem has existed for much longer than ChatGPT. That doesn't mean that ChatGPT isn't part of the trend, but it's a very recent contributing factor. | |
| Jun 7, 2023 at 21:31 | comment | added | TylerH | The clear message I got in terms of "the problem being solved" here is they are seeing negative user growth at an alarming rate and it coincides with ChatGPT suspensions becoming a standard practice. Therefore they are attributing that alarming negative growth (whether it is correct or not) to the ChatGPT suspension practice. | |
| Jun 7, 2023 at 20:39 | history | answered | Thomas Owens | CC BY-SA 4.0 |