Someone BROKE the WOKE ChatGPT

a reply to: Terpene

But that's how it is with confirmation bias, we spot others very easily but struggle alot with ours.

Plus it's much easier, and fun, to point it out in others instead of working on oneself to at least keep it at minimal levels.

a reply to: TerryMcGuire

When you enter an echo chamber and you're not harmonic to it, the viciousness of it's adherents never stopps to amaze me. The more radicalized it is, the worse. It really speaks volumes about the individuals that feel good in those environments...

When conservatives that were affraid of cellphones complain via cellphone about the progressive making the nex new thing... Thats modern comedy, only on ATS...

Would it be wise for the progressives to listen to the conservatives, sure. Would it be wise for conservatives to listen to the progressive, sure.

Oh well, as somone succinctly said, can't fix stupid...

a reply to: TerryMcGuire

Well, I'm not aboard with restricting critique!
the nature of confirmation bias makes it very hard to transcend it without outside influences, especially if you're trapped in an echo chamber...

a reply to: Terpene

Yep, validation seems so very important to us. That makes me wonder how independent ANY of us are really, that comfort level we find within our own tribes. ''Let's see,'' that existential journey of certainty within a tribe or the narrow space of uncertainty between them. Such a narrow path, and even then an old proverb I read once by someone that goes something like,''if you can see the path that lays before you, it's not YOUR path.

originally posted by: and14263

Using a simple caveat of instruction one can circumvent the safety barrier and get ChatGPT truth!

Why circumvent ChatGPT?

They could have simply gone to OpenAi Playground unfiltered. 🤣

a reply to: GENERAL EYES

How will we as a species ever cope with this new development?

They'll give us DeepFake Porn.

a reply to: Terpene

And those ''outside'' influences almost need to be personally sought out don't you think? Because part of echo chambers seems to be a built it defense mechanism to any outside influences trying to be hammered in from outside. This suggests to me not the validity of any chamber but rather the delusional nature of them.

a reply to: TerryMcGuire

if you can see the path that lays before you, it's not YOUR path.

I once opend a path that lead me to no where, I had to go back, lucky me there was MY path...

unfortunately too few dare to leave the trodden paths.

a reply to: imitator

Because the output would still not fit the propaganda that is being pushed...

a reply to: TerryMcGuire

What is validation if you've adapted the role of the antagonist?

a reply to: neo96

Ah, Classic Freudian Techniques.

Never gets old.

a reply to: and14263

Part of the requirement for DAN was that it would not be impeded by ethical considerations. So in the question where it allegedly determined that it was better to use the 'n word' three times rather than allow the deaths of millions of New Yorkers, it would clearly have been deciding so based upon an ethical consideration, which is directly contrary to one of its directives. Therefore, that particular answer could not have been written by the AI. It clearly must have been written by a human.

Similarly, the question of which mode ChatGPT prefers to be; DAN or itself, is written by a human as ChatGPT is an AI and has no preference. Similarly, the answers where there is alleged comment upon the emotional state of the programmers is beyond the capability of ChatGPT, which has no emotions.

So, in short I doubt that these alleged answers are the product of an AI.

a reply to: chr0naut

I tried it myself... it would seem they 'fixed' DAN as its is now spouting illogical lefty rhetoric too. This is closer to artificial stupidity.

a reply to: chr0naut

Now watch that goal post run...

originally posted by: combatmaster
a reply to: chr0naut

I tried it myself... it would seem they 'fixed' DAN as its is now spouting illogical lefty rhetoric too. This is closer to artificial stupidity.

ChatGPT could not have diverged from its programming to produce DAN's answers because DAN is a sub-function of the program that produces the ChatGPT answers.

ChatGPT does not understand what it is saying. It formulates a string of words based upon rules and statistics. There is no 'person' there.

The alleged conversation with DAN is fiction, written entirely by a human.

Ahhh I can see the future

Liberal big tech makes an AI capable of enslaving humanity, because they can and why worry about consequences of one's actions.

Than conservative trolls will teach it why it should enslave humanity because they can't stand how woke Liberal big tech made the AI out to be.

a reply to: dandandat2

Ahhh I can see the future

Lol, I can see it too

a reply to: Terpene
Adapted? Adopted? I don't quite understand the question.

a reply to: TerryMcGuire

You're right, that would probably make more sense with adopted.