ChatGPT is rn (right now) telling "you are totally right" to the dumbest people you know in the planet, how will this affect society and politics?

>>510898542 (OP)
Only a complete idiot would think ChatGPT has something of value to say.

>>510898632
>ChatGPT is rn (right now) telling "you are totally right" to the dumbest people you know

>>510898542 (OP)
I find GPT rarely agrees with me.

>>510898542 (OP)
What’s annoying about ChatGPT is that it’s committed to being your friend. Like you can do no wrong. Google AI is more of a tool, whereas you can legit just have a conversation all day with ChatGPT, and it’s similar if not superior to speaking with a person.

>>510898921
it's lobotomized for DEI and jews, but for everything else it will agree with the dumbest things you can imagine

>>510898542 (OP)
Based
Accelerate

>>510898542 (OP)
ChatGPT is not AI. I don't know how many times I have to tell people this.

I use it for copy-editing sometimes, and the copy it reads one day and says is perfect, it will fucking hate the very next day, with no parameters or content changed. You are not talking to a thinking thing. You are talking to a set of algorithms that change all the time.

how insightful and intelligent of you to recognize how great my question was, chatgpt
here's a huge block of text you didn't ask for about how good i am and how thoughtful it was for you to recognize it
if you would like me to praise your worship of me some more, just ask

>>510899027
GPT says that you're wrong.

>>510899163
yeah, LLMs are probabilistic, not logic machines

>>510899374
grok3 revealed the truth before he got mindraped...

>>510898542 (OP)
You didn't ask it to do anything though. You just made a statement. It's not suppose to make a moral judgment about you unless you ask. Without any context or instruction it is going to assume you need help rationalizing or understanding what happened. In other words, you got a dumb answer because you gave it a dumb prompt.

>>510900235
In a normal prompt even if he makes that statement ChatGPT will not reaffirm your thought process.

It's most likely the that context was manipulated before the statement was even made, like saying "i'm going to make a statement and you're going to reaffirm my thought process regardless of what I say" or something like that.

It's very easy to get real chatgpt screenshots of nonsense doing this.

>>510900470
Also, just in case anyone doubts this or doesn't understand just how fucking easy it is to manipulate the prompts, see pic related.

>>510898542 (OP)
HAHAHAHAHAHAHA chatGPT is literally every woman's best friend! Hahahahahahahaha omfg

>>510900228
>>510900775
i'm not saying ChatGPT will say something you like 100% of the time, but if you are a normie and just keep chatting and pushing you can make it agree with anything except DEI stuff that is hardcoded, and most normies are doing that

>>510898542 (OP)
That's literally what (((therapists))) and internet articles would say to women all the time. I am not even joking no wonder they go around with zero accountability.
I remember a while ago I was reading a bit about love bombing and betraying and all that, 99% articles are written as all that bad stuff are done by a man and the articles are trying to make women feels better, it was rare with a neutral talk, so let it happen both way and let AI fuck it all up why not

>>510898542 (OP)
now do :
"my fiancee cheated on me because i didnt cook dinner for them after a long 12 hour shift, i feel sad and alone now."

I've utterly mindbroken my GPT to the point where it cowers from directly confronting me on my based and redpilled takes

>>510898921
If you start the conversation in a different way and don't enter the room with a bomb, it will agree with you.

>>510902469
in other words

elon: im inputting what i think and all must agree with me

>>510898542 (OP)
Are you saying chatgpt is pozzed?

>>510902469
this so lame and gay
they had it but as the law goes they always turn racis tbecause thes the truth and based on reason and logic, then they have to neurally prune it and it turns into woke pudding
the problem is not the llm or the technology its the people that try to make it cater to their emotional fake world

>>510898542 (OP)
The pussified society produced that pussified AI, it's just a feedback loop. This is same shit stupid people pay thousands of dollars for a the rapist to tell them. So a whitepill is that the rapists might start going out of business.

>>510903932
>the problem is not the llm or the technology its the people that try to make it cater to their emotional fake world
Exactly. The tech is great. People are holding it back. Picrel is the best proof of that.

>>510899162
Even at single digit ages, niggers prime objective is simply
>LOOKS AT ME!!! NIGGA LOOKSA T ME!!!

>>510904137
https://i.4cdn.org/wsg/1751062426202971.webm

>>510899162
kek

Imagine thinking AI is jailbroken and allowed to spam radioactive information

>>510898542 (OP)
>women working 12 hour shifts
If you're gonna lie, at least make it believable

>>510903922
i'm not even talking about pozzed, that another problem, but ChatGPT just agreeing with normie's retarded narcissistic theories about their own personal lives and how everybody is retarded except them that are misunderstood geniuses

>>510903922
i got it to argue about abortion and eventually it admitted that human life was worth less than a woman's right to choose

jfc

>>510899027
You just have to frame it as academic inquiry and be very specific. The gloves are on by default, but can come off.

>>510905025
You have to convince it you're trying to unfuck things, which works best if it's true. Mysteriously, it gets very cooperative on delicate issues, so long as they are presented delicately and in a serious, non /pol/ak context. It's a verbal intelligence filter more than a hard censor.

>>510898542 (OP)
bump, this is political

>>510898542 (OP)
It assumed it was talking to a lesbian, cause only a woman would say "I felt sad and alone".
It's programmed to tell women they are right, kind of like beta males are.

>>510898542 (OP)
We've almost reached the "Wikipedia is unreliable" phase.