Topic: Reee: Which campaign has the most spammers? (Read 1017 times)

Let's use the above post as a learning experience... I'm not trying to criticize the poster in any way, but let them (and anybody else) know how to write a post worthwhile reading..

 - First of all, you need to break up your paragraphs. Four sentences per paragraph max.
 - Second of all, you don't need to bother making a recap of the thread, "what you learned"; it should be taken for granted that you read it.
 - In addition to writing shorter sentences with more punctuation, it'd be good for you to heed your own advice here:


Anyway, this is how I would have written your post, barring any major word changes:

I have read some of the comments and the arguments on the thread and most of the suggestions and advise given to Stake campaign manager is very and if only the CM accept them. And mostly the reduction of the campaign weekly posts. Because why most of those people are spamming because probably they are working in the physical life and also have an account in altcoinstalks so the work load is heavy for them so they are to post with marathon or use AI to complete their posts in the week to receive pay. So I also agree that they should reduce the quantity and go for quality. And quality always comes from 15 to 25 of posts per week.

Even 30 can generate shitposters. You don't behave to occupy the participants in the forum for 24hours. You have to free them to think what to write and be a creative writers. And with that quality posting will emerge. And another thing I also discovered from Stake, the campaign manager or whoever select participants purposely select shitposters and spammers. Probably that is their choice.

After the first thread for spammers Stake restructure their campaign rules yet it is on high side for posting quantity of 85 to get paid fully so many users have targeted that for every week to be paid full. And forgotten the quality. And one thing Stake participants should know is that, writing a long comment or thread in most cases lead to out of points or beating around the bush. Just straight to the main point and leave.

People like your simple contribution and not your long writing. Long writing is very good with good points, and what is good points unknown because what you think is a good points in your long writing is not a point for another users. Be precise and concise. Be a good ambassador for Stake.

Where I'm online its extremely common, but for this community now that I think of it more, I don't see em very often.


That explains it but unfortunately these people seem to refuse to use their brain even if their life was on the line. Their AI spam will be caught on and they will get exposed for it.


Also, to add on wasn't this "borovichok" guy basically given two chances to stop posting AI? He had drama over AI spam like one or two months ago which should've been a wake up call to stop there. It really isn't difficult for him to drop the AI stuff and just start finding and posting manually on topics that he has knowledge of.

Whatever he's in right now it's definitely over for him at this point, what he is doing is just extremely stupid especially after being given two realizations that forum members know hes AI posting.



Loll

I'm not the best person to ask; I don't keep a close eye on that space (AI, that is).

The LLM stuff in particular is pretty underwhelming, IMO: it's like having really convenient access to someone that's read every book in the library, but can only remember some of it, and didn't really understand any of it...

I hope that these types of posts don't become the norm on Bitcointalk: they're seriously tedious to read [1]. It's obvious to me (and many others) when someone's post is either superficially perfect but totally worthless, or when their post isn't worthless (because it has original thinking embedded in it), but it's clearly written at a level that's beyond the author (as in, the presentation is sophisticated, but the ideas are not). In either case (that is, the text was either generated or improved with AI), you're not fooling as many people as you might think, so, why not just relax, and be yourself? I mean, I get not wanting to sound like a dumbass, but if you're not a dumbass, then, that will shine through, trust me, and if you are a dumbass, then, well... that will shine through, too; there's no hiding it, I'm afraid: all the paraphrasing in the world can't make a stupid idea smart (you might succeed in making a stupid idea sound smart, but it will still be a stupid idea).

Anyway, if I were you, I'd stop using Grammarly (they seem to have rebuilt their product, and they're now leaning pretty hard into the "AI revolution").

[1] I've been considering an approach for some time now to get ahead of this kind of non-organic posting: I think giving each member a way to quickly "delete" posts (just from their own point of view) would be cool. But, that's not really an interesting idea on its own, where it gets compelling is when you allow people to divide their own personal modlog into named "channels" that are then made available for other users to subscribe to. So, nutildah, for example, might have one named "AI shitposts" and if I were subscribed to it, then any AI shitpost (in his estimation) that he deletes would be omitted from my view, too. That way, there's some tooling for users to take issues like these into their own hands, and because the moderators themselves aren't involved, no one can really moan about being censored or having their rights violated and suchlike, because it's their peers effectively telling them to shut up. I could also see some campaign managers picking a few very trusted user-modlog-channels to be subscribed to (either all the time, or maybe just when they're counting posts), which would make it easier for them to avoid paying for spam.

Wow, did anyone else notice that some posters with supposedly higher ratings have a higher maximum quota? There are a few users with up to 85 maximum paid posts per week. I remember having a 60 post quota, but it seems they've changed their policy and it now depends on your rating, something that I don't recall seeing a few months ago that I had gone through their announcement thread.


On top of that, the reported user borovichok has a score of over 7.5, deeming him as a good poster, by 3 different individuals.

That's true, the ones who may entirely depend of AI for their posts often are those people who don't put time to write the posts themselves. AI spamming or even use of AI for correcting your English mistakes is something which's totally destructive for someone's learning process.

I know there are many users who don't have good English grammar but they try their best to write posts themselves and overtime their writing gets improved. The ones who basically rely of ChatGpt or other AI generators that are available will not even think about improving themselves.

It really requires effort to create posts and someone who's lazy would never ever put any effort to write the posts himself. Thanks to your thread, the number of AI spammers have been reduced on this forum and most AI generated posts are also getting deleted by moderators only because of your thread. I really appreciate that thread and I believe most other members also appreciate your efforts in creating and maintaining that thread.

See, when this spammer wants to write in perfect English he can and he does. Here is an example of when he wants to use AI or pay others to post for him, there is a complete distinction. How is it possible everybody else can see it but Stake campaign manager Carollzinha cannot see it?

There is absolutely no doubt account farmers are entrenched in many campaigns but the Stake campaign has some of the worst spammers and furthermore, it seems to have more participants with neutral and negative tags as well as dubious post history than others. Only a particular campaign manager went on a hiring spree that had low quality posters with neutral and negative tags but that eventually ended when mixers were banned. Stake stands out alone now.

Don't give them ideas! Most posts from such lazy "AI-assisted" spammers and shitposters are dumb enough as it is. 

A new business idea, much like https://www.grammarly.com/?   Take some text and "dumb" it up. 

Because that would require extra effort and the whole reason shitposters are using ChatGPT is to avoid having to commit any effort at all.

If they are too lazy to just think about what they are going to type, then how do you expect them to go through all that trouble of replacing words. Now you have talked about it, how many users do you often see should short forms of words such as "u" or "bt"?

Seeing a user try to spin the text abbreviate normal words, which is unusual behavior here in the forum, is a number one red flag for me.

AI is easily identified when it makes a lot of circular posts that add nothing to the conversation other than common sense.

Original version


Post enhanced by grammarly/Quilbot

Grammarly and Quilbot are built with AI software, thus I believe that some writings modified with these tools are vetted by AI detectors. These tools assist me improve my broken English and incorrect use of punctuation marks. Check my post history; posts without these tools are severely messed up (wrong words and punctuation marks).

Routine: post here =>copy =>use grammarly/quilbot==>enhanced the grammar and punctuation marks==> post. Could be a bad practice but this seems to be the reason my posts get picked up by those detectors easily.

@nutildah your inputs and recommendations are duely noted. Thanks man. Appreciate everyone too. Better days ahead.

Why spend hours contributing to constructive discussions when, with a little AI magic, you could churn out the same volume of posts in a fraction of the time? (That's rhetorical of course, but probably the logic of some of those shitty posters.)

---

Yep, basically the same as recent case with LUCKMCFLY. According to their rating sheet, he was their "best" poster for months in a row, consistently receiving top ratings from the whole team. This is pure sheep mentality at its finest, or just blatant manipulation.

I'm confused as to why AI spammers haven't at least started trying to hide their AI posts a little to make them sound more human. example; replacing "you" to "u" and not having proper grammar. This probably wouldn't entirely fool an AI detection test but at least they would look a little more realistically written by a human than AI.

Everyone isn't a designer, nor do they want to learn. They want to post and get paid, nothing else. If they're running multiple accounts, then they have less time on their hands as well.

I've said it many times, how hard is it really to give your thoughts on a post? That's all posting is, sharing your thoughts or expertise on a subject. Why do users feel they have to cheat? I just don't get it.

How about learning how to resize the images? I have wasted 30 seconds of my life scrolling through these images because they filld the whole page. This is not in the forum rules, but it's more about common sense.

Have you seen the thread Nutildah created, and how are people reporting those AI-generated posts? It was advised to the reporters not to rely on a single AI detection tool as it may gives false positive result. So, they use at least three tools to check if a content was generated by AI or not.

LOL. You will go to such imaginative efforts to defend yourself here. Simply stunning, worthy of an AI-generated round of applause.

The only problem is your writing often sucks when you're not using AI. Here's an example of where you combined your own words and ChatGPT words together in the same post (ChatGPT is in bold):


So from the first sentence of your reply, we see that you go from not knowing how to appropriately space an apostrophe to flawless English. What are the odds that something like that would happen? Here's another:


You want to "add voice to what they have noted"? Is that really an expression? No, its not, and its not something ChatGPT would ever say.

Just stop using ChatGPT to write your posts for you and they'll stop being deleted. Its really that simple.

This is not true at all and you provided evidence yourself when you gave an example of a post flagged by only one AI detection tool. The others identified it as human-written content.


That's not true either. You just made a well-written post with proper punctuation. None of the AI detection tools have flagged it as "fake".


Looks like you haven't checked your trust summary recently. Remember that post from last year? The one that earned you a lovely "shitposter" tag from Hueristic?

Here's a reminder in case you forgot:

Still think spewing nonsense is a winning strategy?


That's possible, but highly unlikely. These AI detection tools go far beyond basic grammar checks. They are analyzing things like: sentence structure, statistical patterns and predictability of word choices. Just like generative AI models, they work on probabilities.  The more "red flags" a post raises across these various metrics, the higher the likelihood it's AI-generated content.

Do not use that saipling tool because it is definitely one of the worst that exists, and in fact most of the text (posts) you check with it will be detected as AI content. If you don't use at least 3 different detection tools, there is no point in publishing results at all, but when every AI detector shows AI in someone's posts, then it most likely cannot be a mistake.

---

Shouldn't that tool only check for spelling mistakes? If you use it to get completely artificially constructed whole or parts of sentences, then AI detectors may detect them. Whatever you've been doing until now, I think you should change the way you create content on the forum - because English is not the native language of most forum members, but we try our best and with time most of us get better.