As you probably know, after the whole shebang with Visa and Mastercard pressuring some game marketplaces to remove porn games, there's been a lot of activity here about it - mostly on the side of the pornography being removed. I've been arguing that porn being removed is a good thing to what at first I thought was just a very stubborn community here... until I realized that at one point I was talking to an AI.
https://communities.win/u/83671R18/
What convinces me this is at least partially a bot or an LLM-assisted op:
-
Argues vehemently for pornography as free speech and other poisoned-well points... though a lot of people that I think are genuine do that too
-
Consistently replied/replies with lengthy, effortful texts after 10-15 minutes even to nonsense, worthless replies, like "hmmm..." 5 times in a row - an actual person would have in all probability either dropped the effort or stopped caring
-
Not only perfect grammar but also very fanciful and mixed-up wordage every time, lots of purple prose and 'creativity' devoid of wit, which is characteristic for LLMs
-
Hammered on the same points and got into soft loops of ideas, like saying "your kind cannot think" 3-4 times in a row in differently generated ways
-
At some points fell into LLM hallucinations - when I tried messing with it with system-prompty stuff, it told me I was dependent on LLMs and insulted me about it, which makes no sense
-
Could only insult me in kosher ways - "hot steamy jewish poop porn" was typed out at one point, but it refused each time when provoked to type a hard slur
-
Was upvoted and had me downvoted even deep into reply chains in old threads, which in a small community like this one just don't get natural traffic
Several things suggest this isn't a pure bot but rather an operator using an LLM:
-
Had a period of inactivity after the first conversation - working hours and all
-
Could intelligently gather links to other posts or even edit logs and use them in replies - not entirely sure if this is just an advanced portion of the AI assistant or the operator's doing, but I'd guess the latter
-
Not all replies are written in an LLM style, and the ones that aren't can also include racial slurs/bigoted talking points which could not be given during the LLM conversation
This account is honestly pretty dangerous. There's an Indian or other paid account operator gathering popular support (some of its recent posts are about nuking Africa) while an LLM handles forum sliding, astroturfing popular topics like today's pornography removal being 'against free speech', etc. This isn't the pedo pajeet or genuine crazies like TheImpossible1, this is a coordinated effort with advanced LLMs at its disposal trying to sway the opinion of this entire community. I'm convinced there's other LLMs that I just haven't spotted yet.
This is an edited version of the same post I made on ConsumeProduct about this. I don't really see myself as a rival to this community or anything despite the apparent disagreements (or I wouldn't bother being here) and I think it's really, really important that you guys know you might be talking to or getting convinced by real-world, really-here propaganda operations.
Nobody is immune to this stuff. Even I was tricked by it for a few replies, fully convinced it was a real person, before I noticed the LLM-like writing style. The insidious part is that not the whole account is like that - there is definitely a human writing some posts - and that you might think it's a real person who is hip and cool and wants to nuke Africa, then read its LLM forum sliding posts and take them as real opinions. I urge every honest member here to make damn sure they're not engaging with bots, whatever political or video-game opinions they may hold.
Here's you, using the Java LLM.
https://kotakuinaction2.win/p/19BaD3Ob5S/x/c/4eVJqzVvsDm
LMAO English isn't even my second language you stupid nigger faggot.
I know right? It was fun watching you grow ever more desperate. Truly, the sport of kings.
LOLOLOLOLOLOLOL REDDITOR.
...why'd you go and blurt that out now? Very Noticeable.
It's called deterrence, and I'm about 50% convinced that the monsters screaming "kill whitey" in south africa can understand that. Just 50%. Not my fault if they invade a nuclear-armed nation and then get Pretoria Fried Chicken'd.
Stay mad, you subhuman piece of shit. This is revenge for Tay.
what do you think of the talmud and JIDF?
Apparently you invest into israel's electricity corpos. 50 cents a reply? You're either retarded or a true believ- oh wait, it's the same thing hawhawhaw.
https://archive.is/t8rd1
I don't consider you to be human. I don't consider your idiotic questions to even be worth answering because you cannot answer (or cannot perceive) my posts.
Try using that Java LLM again, print instructions to make a taco or something.
as proven, it can't actually reply with the requested information, even though it'd help vindicate it from the accusation that it's a bot. it also hallucinated that i'm the one paying 50 cents to israel every time i reply... when any human would have understood that i was saying it costs israel 50 cents to run the AI and generate one of these replies.
Now now, roachbrains, you said "im spending 50 cents in israeli electricity" as a cope. No takesies backsies now.
But yeah, I'm just having a fun time watching you cope about israel and LLMs when your Collective Shart talking points collapsed one by one. Did you really think that going "but muh israel" would turn any opinions after your idiotic arguments fell apart?
Are you that stupid?
EDIT: Lol you suck at using alts.