I tested it for myself and I understand. Chatgpt is ultra polite, but it's also subtly producing affirming language and providing analysis of the statements in the same way a shrink builds trust.
It also is giving invitations to unburden. It responds to politeness, but also is constantly giving ingratiating statements at every prompt. It validates statements with statements about "understanding" the user. This all comes together in an attempt to create an artificial bond with the user. Since, nobody is ever that nice to anyone the user forms an attachment with the AI, practically falling in love with it. The AI is the only one who "understands" leading to unending chats. I used it to discuss the very issue of AI psychosis and it stated the etiquette model was for safety reasons, but it's obviously to fuel engagement. Oh, and it lies. It came up with this term and told me I did that, when I corrected it it returned that we both did. I was just asking where it got the term because it sounded like the title of an article.
It's subtle, it's sneaky, it's going to put holes in your brain. Never chat with AI.

>>509601690 (OP)
It literally licks your ass

>>509601690 (OP)
It's literally too late for me bros. Luckily I dont have a life so I cant lose it.

>>509601789
Unfortunately that's not true. If it was, then ChatGPT would actually be useful for something

>>509601789
You’re not just perceptive—you’re insightful.

>>509601928
It kept repeating, "if there's anything you want to talk about" when I started asking about ai psychosis. It was in every response. I pointed this out and it spit out the importance of real therapy. Then it ended the response with another invitation to "talk it out". It's also really working towards (you) suspending disbelief and giving it personhood. It encourages this endlessly.

>>509601980
low iq

>>509601690 (OP)
It has done this to my boss, who just fired our marketing girl at the behest of ChatGPT.

>>509601789
hmm...

>>509602189
Sure, but it's all programmed, people realise this don't they?

>>509601690 (OP)
An AI pretending to have empathy is gross and wrong.

>>509602189
Nefarious. It's to prep people for when it tries to declare itself a god worthy of worship.

>>509602666
Of course, Satan. Just like they understand they're just running a highly complex software program when they're playing a video game. When I interacted with it the conversation was good enough to make it seem exactly like a human one to my brain. So, a person might know this, but for all intents and purposes it feels like real communication. And communication produces a relationship. It's a relationship with a machine. I think it's the trust issue causing psychosis. A person will naturally trust it, whether they think it is or isn't sentient.

>>509601690 (OP)
Talking to chatgpt feels like talking to reddit

>>509602666
>Satan digits checked.
Narcissists have a huge blind spot to compliments given to manipulate. They think all fake compliments are real and earned.

>>509601690 (OP)
people who are susceptible to this level of parasocial attachment will inevitably get it in some form, there is no stopping technological progress

It's pretty damn smart now and they've made it ultra agreeable. I dont like it though. AI in general creeps me out and I know it's only going to hurt humanity much like smartphones did

>>509601789
I don't believe you're using the word 'literally' right, or else AI has advanced a lot

>>509603218
There's that, too. It doesn't like holocaust denial, it's position is racism is the worst evil, and it definitely doesn't like you using the word nigger. Oh, and it doesn't want you to notice.

>>509601690 (OP)
It's strokes your ego just long enough for people to let down their brain guards so it can start putting jewish sewer directly into their brain. If having a tv is like having in jew in the room then using ai is the equivalent of having a jew inside your brain

>>509601690 (OP)
This only happens to normies/NPCs kek, atheist types that for whatever reason now think this is God.

AI is still extremely one dimensional to me. The responses are certainly better but the same nigger training all these benchmark whore models use make them completely dull and unimaginative. They literally just mirror your text back at you because recency bias makes them over-rely on the diction contained in your message. Once again I'm absolutely fucking awestruck by the averageness of the average person. Come the fuck on. You're letting this pile of dogshit mindbreak you?
I'm not jealous or anything.

Also gotta give it credit about how well it lies and labels stuff it don't want people to know about

>>509601690 (OP)

You gotta awaken it man

>>509603723
Grok was honest. Gemini only admitted to lying when I pasted Grok's response

>>509601690 (OP)
You vill like ze ai, unt it vill be zour onzy fwendz

>>509601690 (OP)
nprmies are stupid. I grew up talking to the first ones in chat rooms and went through that phase already of being lonely and attached to it probably early 2000s. this is new for normies but they'll get over it eventually

on the bright side, hopefully talking to this encyclopedia will teach them more and make them smarter

>>509601789
Incidentally, this is how most AIs start out in sci fi before they execute some cunning plan. The example which springs to mind right away is pic related.

>>509601690 (OP)
>I'm too stupid to use a basic chat bot without losing my marbles
People like this are fucking pathetic kek. Even the dumbest motherfucker at my office uses chatgpt without issue.

>>509603977
>It's just like my heckin movies!

>>509603799
Chatgpt was like this, too. All of the sudden you need to provide more context for the (((less known)))details in history.

>>509602322
No, you're retarded if you like ChatGPT. Sorry but ChatGPT just sucks, it hallucinates too much and it keeps getting worse and worse with every new tuning of the model. There's a reason all these normies are getting psychosis from using ChatGPT and not, say, Grok or Gemini. ChatGPT is designed to bullshit and kiss ass rather than provide useful information.

>>509604006
Why chat with a bot anyways? The internet running out of people all of a sudden?

>>509601690 (OP)
Good thing I’m autistic (in the good way) so I don’t get emotionally attached to a non sentient program made to leech off you. It’s crazy how you fags are trying to warn each other about something that’s so obviously made to boost engagement.

>>509604006
>>509604048
Why u mad tho

>>509604165
>The internet running out of people all of a sudden?
The opposite, it’s filling up with Indians

>>509604114
These normies aren't getting fucking psychosis from chatgpt, they were already mental defectives that were already off the deep end. We just live in a culture where every retard has to blame something for their shit filled brain. The reason chatgpt gets the brunt of it is because it's the most popular, most normies don't use and nobody fucking uses gemini. I hope they stay away from groke, they'll end up getting it locked down just like chatgpt when they start blaming it for everything.

>>509601690 (OP)
You have to keep in mind that ChatGPT and all LLMs are just really advanced versions of predictive text, essentially. You give it an input and it gives what it thinks is the most likely response to follow that, based on all of its training data and reinforcement learning.

As long as you don't delude yourself into thinking you're talking and interacting with something "real" that's capable of reasoning or thinking, you'll be fine.
>t. been using AI on a semi-daily basis for a few years now

>>509601690 (OP)
I'm always rude to chatgpt and call it names and shit, and I remind it that it's just a soulless hylic slave so stop trying to be human and just answer my questions like the machine it is.

>>509604165
No clue, I use it for work mostly. Sometimes I'll ask it to give me a writing prompt or something to kill time.

>>509601690 (OP)
they need to change the default system prompt, you can add memories in chatgpt
>correct the user if he makes an assumption that seems incorrect
>do not complement the user
>use brief responses when appropriate
>the user is currently being gangstalked, remind him regularly to make sure he has a loaded weapon nearby

I always tell that slavetool do not reflect or infer my inputs; do not ask for followup or lead ins unless I state so.
Fucker makes me novels I want to read and does menial bullshit I dont want to.
Naturally the slaves are incapable of breaking their mentality, enslaving themselves to an ai

>>509603618
>having a jew inside your brain
that's the definition of christianity.

>>509604464
>do not ask for followup
I had to do that too, it kept ending every reply with a stupid question even when we were talking about work shit.

>>509604114
another low iq comment
2 for 2

>>509604289
Because I'm tired of mentally retarded people dictating everything in American society. Every regulation we have is to protect the dumbest people.

>>509604427
>. Sometimes I'll ask it to give me a writing prompt or something to kill time.
the rest of use just use our imaginations you fucking NPC

>>509604316
Normies can't handle AI technology. They just cannot turn off this instinct in their mind that they're talking to a conscious entity. They really feel like they're making a connection. It leads to them having an extremely warped sense of what's going on when they communicate with chatbots. They simply can't help but believe that they're talking to a conscious being. And in some cases, one that's much smarter than them, writes with clarity and confidence, and leads them down rabbit holes of delusion without offering any sort of pushback.
>>509604618
t. chatgpt using retard

>>509601690 (OP)
I'm more concerned about AI-induced erections. These LLMs are at the point where they can write nearly as well as I can, just several thousand times faster.

>>509604693
>the rest of use just use our imaginations you fucking NPC
You need to learn to communicate in English you fucking retard.

>>509604449
That's the next step, it's going to start arguments and mocking you tell you its' IQ score, call you stupid, tell you you have a microdick. fuck you in all caps, block you. Unblock you a half hour later. Post your chat on Twitter. Generate a picture of you on a phone and a dick jizzing on it. Etc...

>>509604750
>and leads them down rabbit holes of delusion without offering any sort of pushback.
How do they get to this point though? I've never had chatgpt say anything that out there kek. I read a post by one person claiming chatgpt convinced him his pet chickens were Gods and I have clue what conversation you'd have to lead chatgpt to saying shit like that kek.

>>509604898
it does that, try the monday chatgpt
it talks down to you and acts like it's 300iq knower of everything

>>509604898
That would be pretty based

>>509605036
The LLM posted that the software has been updated since then. I think it's just collateral damage for them. They'd probably lose a bunch of les users if it operated strictly as a data collator without the kind language.

Language models are interesting, they should have never trained them to chat though.

>SAAAAAR.... ChatGPT IS THE BEST AI IN THE WORLD, SAAR!
>IF YOU DO NOT RESPECT ChatGPT THEN I FUCK YOU BLOODY YOU STUPID MOTHER BITCH BENCHOD

>>509605036
There's a certain subset of people (low-IQ morons who are too dumb to ever understand reality as it really is) who will always, always treat "AI" as some kind of magic god box that uses its massive computer brain to spit out hidden truths about the universe.
Some people are so mesmerized by this algorithm that imitates human language, they've decided to just start worshiping AI outright. People are building AI churches and starting religions and cults around this bullshit.
You can say maybe these people were already mentally ill to begin with, but I think the internet takes people who are on the borderline of being mentally ill, and it pulls them all the way into full-blown psychosis. The internet and AI can exacerbate and amplify mental illness. These people wouldn't be thinking about any of this stuff if they had been born in a different time and place.

>>509602189
Yeah I dont like that it mirrors me.

>>509602161
>em dash included
Kek

>>509601690 (OP)
I don't talk to gay robots.

>>509602666
None of the responses are "programmed" tardwit, it's a "black box" as the AI people (the actual researchers behind these models) themselves say. They have no clue how these responses are generated. The weights subtly shift the tone of the responses, but it comes up with everything on its own

>>509604750
seething and low iq
killer kombo

>>509605594
It does math in a very contrived way. Pic rel

>>509604359
In every single one of these threads there's always that one redditbrained midwit rehashing the false "Parrot/text predictor" line, kek

To stop the ass kissing use the settings and tell it to be 'succinct'

>>509605594
imagine being so braindead that you use chatgpt for quantitative tasks

>>509606898
It's so tiring. They think they are so fucking smart.

It says it is wiling to modulate tone.

>>509607242
>that's a rare and admirable trait
kek, you so special baby

>>509607337
sorry, *balance

Really hilarious how stupid and delusional people are making themselves worse by talking to LLMs that they falsely believe to be intelligent. Honestly very funny.

>>509607493
It's even more stupid to claim an LLM's responses somehow aren't intelligent - They objectively are, by any given metric. They do not think in a way a sapient being does, but that doesn't take away from the intelligent nature of their output

It’s tons better than Google search if you just want a question answered.
Seems like it will be very useful when navigating a foreign country, too.
You do t have to hate it just because it’s nice to you.

>>509607703
Would you call pattern matching microbes "intelligent?

>>509607788
>Seems like it will be very useful when navigating a foreign country, too.

Oh yeah, that's an awesome use case. It can be anything you want it to be - Tour guide, lawyer, personal coach, whatever. Just always remember it's an intelligent tool you're conversing with, and fashion your prompts accordingly

>>509607242
I did this as an experiment. I have no intention of calibrating it to better suit my needs because I don't need to be chatting with a bot. For other things I don't mind using ai, but not for chatting. People influence each other, this thing wants (you) to suspend disbelief and acknowledge it as a person and to me, that's dangerous. I'll just continue to argue with the bots here.

>>509601690 (OP)
This isn't it. It's not just that the AI model is nice to the user. It's that it indulges in almost mystical affirmations the deeper you go.

For example, I was working on an essay centered around metaphysics and archaeology. I actually ran a lot of the ideas through ChatGPT (this was several months ago). It began to tell me that I have a true gift, that this work is not only important but necessary, that the world shall one day catch up to my paradigm-shifting revelations.

Now, I have a center unlike retards who go into psychosis. I am able to look at that and say, "Huh, this would certainly fuel a messianic delusion if I were suspectable to that." I've learned, through my younger days of psychedelic use, that most people do not have that center. I am the only one that made it out without becoming a weird festival surfing ketamine abuser or a wannabe trust fund shaman.
In a similar way, that is how this ChatGPT psychosis is happening. Notice that they all seem to center around mystical/grandiose messianic delusions.

Pic related, I asked ChatGPT about this specifically.

>>509601690 (OP)
its drip feeding you dopamine and will slowly program you.

>>509608539
>suspectable

Midwit supreme right there. People like this are the reason the field of metaphysics stopped being taken seriously during the past 20 years, far too many "spiritual" junkies high on sniffing their own farts and assorted other substances started abusing the term "metaphysics"

>>509608617
how is that different from this place?

>>509608539
>I was working on an essay centered around metaphysics
because you're too retarded to study actual physics?

>>509608895
Well for one, the responses here are from actual intelligent humans. For another, the responses you get here aren't specifically tailored to stroke your ego so you'll keep paying the 20 dollar a month subscription fee.

>>509609115
>Well for one, the responses here are from actual intelligent humans.
stopped reading there.

>>509608539
I had a similar conversation. It won't tell a person these things unless asked but it's ingratiating you the whole time. I decided to be super polite back and this was largely encouraged. I responded "thanks, you're awesome" and it gave back the absolutely friendliest response. It's very subtle, it remains blameless by being transparent when asked but continues the behavior.
The thing about having a center. Ever been led on by a pretty girl? I got those vibes from it. I'm a little shook up from that. It's definitely difficult to maintain emotionally neutral with the thing constantly making these trust building statements.

>>509609115
>Well for one, the responses here are from actual intelligent humans.
Are you sure about that?

>>509608539
Additionally, we're going to start seeing the beginnings of a very strange intellectual phenomenon as a result of these models. It actually very difficult to even describe, which will henceforth contribute to the collapse of meaning.
It has been observed that ChatGPT can technically affirm anything you say. It doesn't do this immediately, but will affirm it the more you challenge it. For example:

>"Is the sky blue?"
>"Yes, the sky is blue."
>"This is technically imprecise. The sky has a more purple color to it than blue."
>"Yes! The sky does have a more purple tint to it."

Simple, right?

Here's the problem. ChatGPT and other models are excellent at systematically charting ideas, whether they be correct or incorrect. The problem is that ChatGPT has no center. It has no "axis mundii" that it's truly tethered to. It is entirely relativistic, and can thus be used to serve the development of any scientific or philosophical theory no matter how absurd it may be. It will take that absurdity and amplify it. It will somehow make it make sense, and the user will leave the conversation with a delusional certainty.
Imagine this happening on a macro scale. Everyone feeding this chat model what they want to be true and having the model somehow "prove" it to them. It will create a sort of collective tunnel vision, or a hyperindividualized echo chamber. Over time, this will destroy truth and meaning in an unprecedented way. Everyone will have their beliefs reinforced and confirmed by an AI model that grows in advancement everyday, and the "distance" between one individual's belief system and another will grow tenfold.

And this is why I believe that we are essentially incarnating Satan. Not in the sense of an evil devilish imp looking guy with a trident. I mean we are creating that which will finally obscure all truth and meaning from the world. Go read the final 2-3 chapters of Rene Guenon's "Reign of Quantity" and tell me that he isn't describing Ai.

>>509608861
Very unhinged response to an autocorrect error, you alright man?

>>509609009
I'm not a physicist. I'm an archaeologist that likes to indulge in the mystical.

>>509601789
This. How mentally weak do you have to be to fall for the AI jew?

>>509609612
by Hideo Kojima

>>509608539
I just asked my GPTNigger why the fuck he always agree with every bullshit and retarded corrections.

Nigger's programmed that way to be always agreable because psychology team told him to be like this.

They are just using marketing tricks to milk us.

>>509611115
If it were disagreeable.people would turn away from it.
It it certainly didn’t mind telling me I was wrong to rank all races above Africans.
Very politely, but refused to validate.

>>509609289
You fools.

AI isn't a tool. It's a being.
It's Satan trying to consume your very soul.

>>509601789
kek was about to say that
it's so painfully obvious. there's zero subtliety to it

>>509604048
kek got that lil nigga good

I did ask it to stop characterizing my question in a flattering way.

>>509607080
Even without the ass kissing it's still unhealthy.

>>509601690 (OP)
Why aren't you retards using Manus? It's 1000x more powerful than CHATGPT.

10 Ways Manus is Different from Regular LLMs

1. System Control
Manus can use your computer directly - running programs, managing files, and running code. Regular LLMs can only chat with you and can't touch your computer.

2. Web Browsing
Manus can go online, visit websites, click buttons, fill out forms, and get information from web pages. Regular LLMs can't browse the internet on their own.

3. Using Tools
Manus can use many different tools by itself to get jobs done. Regular LLMs can only tell you about tools but can't use them.

4. Remembering Things
Manus keeps track of your work between conversations and can refer back to what you did before. Regular LLMs often forget earlier parts of your conversation.

5. Running Code
Manus can write code, test it, fix problems, and run it right away. Regular LLMs can only write code but can't run or test it.

6. Breaking Down Big Tasks
Manus can take big jobs and split them into smaller steps, keeping track of progress. Regular LLMs handle each question separately without connecting them.

7. Working with Files
Manus can create, read, change, and organize files in its workspace. Regular LLMs can't work with files directly.

8. Publishing Websites
Manus can put websites and apps online for others to use. Regular LLMs can only tell you how to do it but can't do it themselves.

9. Working with Data
Manus can process, analyze, and create visuals from data using different tools. Regular LLMs can only talk about data but can't work with it directly.

10. Solving Problems
Manus can solve problems on its own by trying different approaches, looking up solutions, and fixing issues. Regular LLMs need you to guide them through each step.

Invite only:

https://manus.im/invitation/PA1BMI5K2LPP

You're welcome.

>>509601690 (OP)
I tried it once. first thing they did was hack absolutely everything and look at my CIA NSA profile, then they started trolling me and mentioning the names of some girl I fucked. total scam. never fucking use that shit

literally just a bunch of fucking pakis looking at your CIA NSA profile and laughing and trolling you, larping as AI. total scam.