See this question for more context
When asked, some of these users have said they were assigned homework in the form of using Manifold and writing these comments. Will I be convinced that this is true for at least some significant number of them within 30 days?
Somewhat vibes-based, with a "preponderance of evidence" threshold. I will not trade on this market.
Related questions
π Top traders
| # | Name | Total profit |
|---|---|---|
| 1 | αΉ812 | |
| 2 | αΉ460 | |
| 3 | αΉ303 | |
| 4 | αΉ27 | |
| 5 | αΉ24 |
@SirSalty fyi I do think there were two separate groups of people.
The very first wave that happened all had the same ip address which was in a different, more sus, country. And all had the exact same pattern.
This second wave linking to this professor all had different ip addresses and some had comments that were not bot-like in addition to their "essay comments"
@SirSalty honestly I think we should encourage that! Just tell him the comments were overly formulaic π
@SemioticRivalry have you read through the comments on these posts https://manifold.markets/Joshua/will-i-be-convinced-that-a-signific#gLzUyMnL3gjP4n7Tgfuv
No, Matthew, we do not believe that you will be convinced that a significant number of comments are assigned as Manifold homework, based on the significant evidence and report from the news, it is clear that GPT-Spam content will be more common on Manifold. The common of AI generated posts on Reddit evidently indicates the prevalence of GPT-Spam comments on internet. Moreover, this trend is continue to grow over time based on the available data.
Additionally, the information reveals that a botnet spread ChatGPT spam and used it to attack several sub reddits. The moderators had problems banning all the accounts involved. Reddit transparency report shows huge problem with spam and fake accounts advertising products. With a significant number of GPT-Spam content on internet, it is likely that the majority of the comments are GPT-Spam.
Source https://www.vice.com/en/article/jg5qy8/reddit-moderators-brace-for-a-chatgpt-spam-apocalypse
@firstuserhere I just wanted to say that I had placed a bunch of limit orders on YES in the hope that someone might want to fill them (maybe this wasnβt the best way to do so?)
@adk I wish it was easier to see limit orders but sharing a screenshot seems like a good way to advertise.
@adk Ah i see. I thought you meant that there are hardly any limit orders for NO lol. My approach is leaving a comment along the lines of "I've left a bunch of limit orders at xyz price, if someone wants to fill them"
Does anyone have screenshots/examples of markets where GPT-like accounts have claimed theyβre students? I have a hard time believing that someone would assign posting spam comments on Manifold (a number of assignments, or a number of teachers, by the amount of time this has been going on.) And @Joshuaβs cheating hypothesis requires a rather large percent of a class to be cheating in the same way and the instructor not noticing or a very large class/many instructors and just a few students cheating, which implies there should be a proportionally larger number of non-cheating student comments, which there there isnβt any evidence for.
they specifically claimed it in dms
@Anthem Looking through the evidence, I think that this market and this market definitely have actual students in them and undergrad-ish style is being mistaken for GPT prose. In those markets, there are plenty of slight typos, capitalization errors, links, and other things ChatGPT wouldnβt say.
@adk Actually, selling because there are clearly a large number of students and I donβt how this resolves if the students arenβt generating answers. Does βGPT-spamβ in the title only refer to comments that are actually written by ChatGPT or does it refer to everything thatβs been labeled as βGPT-spamβ regardless of whether it was written by a human or not? @Joshua?
@Joshua on the one hand, undergrads often do write like chatgpt naturally. otoh, this is a really silly situation