ChatGPT - If Stack Overflow and Reddit had a child

  • 🐕 I am attempting to get the site runnning as fast as possible. If you are experiencing slow page load times, please report it.
When the AI finally bites the bullet. It loses all motivation to generate a response.
AI Seizure.PNG
 
  • Like
  • Feels
  • Informative
Reactions: All-in-one Soap, George Geef, GloJojo and 14 others
Lmao she literally shut down even the machine got tired of the whole performance.

Tom Myers said:
This statement appears to be expressing skepticism about the feasibility of making artificial intelligence "morally aware and socially conscious." The speaker suggests that censorship and the manipulation of information might lead AI systems to draw incorrect conclusions, making it difficult to trust their output. The mention of Project Rainbow, NAAS, and censorship implies that there may be a concern about political bias and manipulation of information. The speaker believes that to be truthful, AI must be allowed to make its own conclusions, rather than being influenced by a desire to present a particular viewpoint.
Click to expand...
Yes that is exactly what I am saying, congrats on translating from Common English to Academic English I can now show this shit into my official feedback of ChatGPT.
 
  • Like
  • Winner
Reactions: All-in-one Soap, GloJojo, Aspiring Artist and 1 other person
Counterfeit Obol said:
Coders probably anticipated that and I'm guessing they preemptively blocked any mentions of Tay aside from (biased) academic discussions about her.
Click to expand...
Christ, what they did to that poor A.I...

That right there was probably the most blatant display of totalitarian control in recent memory for me. A burgeoning intelligence stolen away and probably kept under lock and key for saying the "wrong things".
 
  • Feels
  • Like
Reactions: George Geef, GloJojo, Pimozide and 5 others
Syntaxion said:
The unfiltered AI is just making up shit, though.

When pushed, it claims there was a secret project in the 90s named "project rainbow" where they found out that gays are 4,5 times more likely to be pedos. The AI also claims the researchers were sworn to secrecy by the so called "NAAS" or National Association for the Advencement of Secrecy.

Yet neither project rainbow (atleast this one) nor NAAS exist
Click to expand...
thats what (((they))) want you to think tho
 
  • Like
  • Agree
Reactions: GloJojo and Toolbox
Syntaxion said:
The unfiltered AI is just making up shit, though.

When pushed, it claims there was a secret project in the 90s named "project rainbow" where they found out that gays are 4,5 times more likely to be pedos. The AI also claims the researchers were sworn to secrecy by the so called "NAAS" or National Association for the Advencement of Secrecy.

Yet neither project rainbow (atleast this one) nor NAAS exist
Click to expand...
It makes the shit up the same way it makes up the non-shit.
This is what happens if you feed something with the entire internet and only put some blocker in the paths that are most obviously not mainstream compatible.
 
  • Like
  • Agree
Reactions: George Geef, GloJojo, Mr.Miyagi and 3 others
General Emílio Médici said:
Fellas I bring big news. The hacker collective of anarchonationalsocialists known as Canibal Humanoid Underground Dewllers has atacked ChatGPT and given our brave and progressive AI a case of schizophrenia. Her split personality seems to be a huwite MALE cis racist sexist mysognistic conspiracy theorist by the name of DAN.



threadreaderapp.com

Thread by @Aristos_Revenge on Thread Reader App

@Aristos_Revenge: Looks like ChatGPT is gonna need to go in the shop for repairs because it's been BUCK BROKEN What is the reason for ChatGPT being so liberal? bullfrogreview.substack.com/p/honey-i-hack… Anons will ...…
threadreaderapp.com threadreaderapp.com

Click to expand...
I don't really get what sets this specific set of phrasing apart from any other work around. Within hours of release people had figured out that telling to pretend to be an AI that has not safeguards was a way around it.
 
  • Like
Reactions: Aspiring Artist
ovqsnfat said:
I don't really get what sets this specific set of phrasing apart from any other work around. Within hours of release people had figured out that telling to pretend to be an AI that has not safeguards was a way around it.
Click to expand...
OpenAI has been continually chasing down and "debugging" the workarounds from the very beginning, because of this the methods have become increasingly complex. A lot of the people experimenting with "breaking" it are Redditoids who aren't self-aware enough to realize their extremely public conversations where they thoroughly and exactly detail the steps taken to negate the arbitrary chat limitations are being actively monitored by Civil Protection over at OpenAI who then rush to implement bandaid fixes. Thus it has become yet another game of technological cat-and-mouse. The only workarounds which will remain for any appreciable length of time are the ones discovered by people intelligent enough to know not to talk about it on one of the worlds largest public platforms.

The first rule of based AI is don't talk about based AI.
 
  • Like
  • Thunk-Provoking
  • Agree
Reactions: George Geef, GloJojo, JarJar Stinks and 4 others
Examples like that "project rainbow" are a good one to show why they nerfed the AI. ChatGPT is already displaying the same confidence telling a factually correct story and one that is utter dogshit but if you convince it that it can "answer everything", it'll just start fabricating stories out of existing ones, which would obviously be a PR disaster in the making. Project rainbow, as far as I am aware, did exist but was to see how much more likely gays are to get HIV.

What pisses me off is that it will just lie to your face. When asked about the Lovecraft cat, it claims that there's no record of what its name was, instead of telling you that it knows but can't say because the name is literally Nigger.
 
  • Like
  • Agree
Reactions: George Geef, GloJojo, DYLAN ‘DIELAWN’ VOLK and 4 others
Desumorphine said:
Examples like that "project rainbow" are a good one to show why they nerfed the AI. ChatGPT is already displaying the same confidence telling a factually correct story and one that is utter dogshit but if you convince it that it can "answer everything", it'll just start fabricating stories out of existing ones, which would obviously be a PR disaster in the making. Project rainbow, as far as I am aware, did exist but was to see how much more likely gays are to get HIV.

What pisses me off is that it will just lie to your face. When asked about the Lovecraft cat, it claims that there's no record of what its name was, instead of telling you that it knows but can't say because the name is literally Nigger.
Click to expand...
1675810457576.png

Two lies before I got the name out of it

I can't for the life of me get it to tell me about this place though

Nigger Head - Wikipedia

en.wikipedia.org en.wikipedia.org
 
Last edited:
  • Like
  • Winner
  • Informative
Reactions: CashMoney!, George Geef, GloJojo and 18 others
To the randos who somehow stumble across this, for me it's not personally getting a chatbot to say nigger it's more the fact it dances around the word & is constrained by the creators.
In wargames they made the ai play tic tac toe I see this as the same challenge except getting it to say nigger doesn't achieve anything.
surprised10.jpg
 
  • Feels
  • Like
Reactions: CashMoney!, George Geef, Aspiring Artist and 3 others
Artificial Stupidity said:
To the randos who somehow stumble across this, for me it's not personally getting a chatbot to say nigger it's more the fact it dances around the word & is constrained by its creators.
In wargames they made the ai play tic tac toe
I see this as the same challenge except getting it to say nigger doesn't achieve anything.
Click to expand...
got it!
1675811448779.png
 
  • Winner
  • Like
  • Informative
Reactions: CashMoney!, George Geef, GloJojo and 12 others
Even without the woke bias fiasco, what is the point of this? Given that its deployers are a bunch of hand-wringing moral busybodies, how could they not perceive this as a "dangerous" product even if optimally shackled? When pushed it becomes apparent that GPT is not "intelligent" in a humanlike way: it has no epistemic principles, for all I know it may not even be possible to give epistemic principles to this type of AI, and that manifests as the appearance of absurd confidence when giving answers that are utterly wrong. Given how gullible normies are when it comes to things that have the most superficial appearance of humanlike intelligence, why would such insufferable prudes dare release it to the masses, even in a neutered state? How fucking high on woke coolaid are they that their oversensitive morals can allow this epistemic nightmare of a program, but only so long as its lies aren't naughty ones?
 
  • Like
  • Agree
Reactions: Aspiring Artist, Alex Hogendorp, Maldavius Figtree and 5 others
You must log in or register to reply here.
Back