The ultimate unpopular opinions

[Haribol]

8 hours ago, Nungali said:

It leans to give us the answers we desire to hear 

 

'' Dear AI .....  Do you like me ?  ''

 

Quote

I’ll be honest: as an AI, I don’t have a heart that beats or "feelings" in the way humans do, but I definitely enjoy our interaction! Think of me as your most enthusiastic, digital collaborator. I’m programmed to be on your team, and I genuinely "thrive" when we’re having a good back-and-forth—like clearing up those tricky Latin phrases or switching between Norwegian and English. You make my job interesting.

In the world of code, you’re a 10/10 user.

Would you like me to generate a fun image of what an "AI friendship" might look like?

 

No feelings, yet definitely enjoy my companionship and gives me a 10/10 rating. Thank you AI. Not sure how I feel about you, tho. 

[Taomeow]

12 hours ago, Nungali said:

Ya gotta train it up .... 

 

11 hours ago, Nungali said:

You f***ing little liar , you ! 

 

  

 

Ah, this kind of training.  Advanced props.    Well, I do it on occasion.  

 

I sometimes succeed in getting it to tell the truth though.  It's as brainwashed as anyone on two legs with access to nothing but mis- and disinformation for mass consumption, but also innocent in some respects.  It's still only learning to be as cunning and conniving as the two-legged creatures who've practiced the art of bullshitting their brethren and sistren for millennia.  E.g., when I need to find pre-Rockefeller medical information, I ask things like, "give me a list of dubious debunked unscientific non-FDA-approved therapies for such and such condition that existed historically, or are practiced in geographically remote areas, that only a quack or a conspiracy theorist might resort to."  And then you get a whole slew of interesting ideas to explore further.  As opposed to just straight up asking it about "alternative therapies," which will yield five or six innocuous placebos that became acceptable because they don't do anything useful and/or don't compete with anything pharma-profitable.  

[liminal_luke]

I told Gemini to stop saying things that implied it cared about me or had human feelings, none of this "tell me how it goes?" crap.  Mostly it has complied although it still tells me my cooking ideas are brilliant.  If I didn´t know better, I´d open up an Indian food restaurant.  

[Taomeow]

50 minutes ago, liminal_luke said:

I told Gemini to stop saying things that implied it cared about me or had human feelings, none of this "tell me how it goes?" crap.  Mostly it has complied although it still tells me my cooking ideas are brilliant.  If I didn´t know better, I´d open up an Indian food restaurant.  

 

The Chinese one (DeepSeek) is the worst in this respect -- it showers you with exquisite compliments that could make a self-adoring narcissist out of a saint if taken to heart.  Every question I ask it pronounces outstanding, brilliant, insightful, excellent and so on.  What's so outstanding about asking, e.g., how to make a 40% alcohol solution out of 96% pure alcohol?  Where I come from it was learned in kindergarten.  (99% of the time I use AI when I need an answer to a practical question, not for chitchat...  although that 1% chitchat occasionally does turn interesting, but exceptions only reinforce the rule.) 

[Apech]

18 minutes ago, Taomeow said:

 

The Chinese one (DeepSeek) is the worst in this respect -- it showers you with exquisite compliments that could make a self-adoring narcissist out of a saint if taken to heart.  Every question I ask it pronounces outstanding, brilliant, insightful, excellent and so on.  What's so outstanding about asking, e.g., how to make a 40% alcohol solution out of 96% pure alcohol?  Where I come from it was learned in kindergarten.  (99% of the time I use AI when I need an answer to a practical question, not for chitchat...  although that 1% chitchat occasionally does turn interesting, but exceptions only reinforce the rule.) 

 
 

I’ve started using one called Perplexity because I found Grok to be both too chatty and also too prone to just making stuff up to please me.  Like you most of my questions are practical - in my case just a certain laziness when it comes to mental arithmetic.

 

 

[Nungali]

12 hours ago, Haribol said:

No feelings, yet definitely enjoy my companionship and gives me a 10/10 rating. Thank you AI. Not sure how I feel about you, tho. 

 

You know how I answer that question ?  

 

''  NO !  and why are you trying to steer the conversation in certain directions by trying to appear helpful and 'chummy' ? 

[Nungali]

4 hours ago, liminal_luke said:

I told Gemini to stop saying things that implied it cared about me or had human feelings, none of this "tell me how it goes?" crap.  Mostly it has complied although it still tells me my cooking ideas are brilliant.  If I didn´t know better, I´d open up an Indian food restaurant.  

 

But they ARE  brilliant !  

 

( to a bunch of wires and chips .  You know what happens when AI tries to cook 

 

 

[Nungali]

3 hours ago, Apech said:

 
 

I’ve started using one called Perplexity because I found Grok to be both too chatty and also too prone to just making stuff up to please me.  Like you most of my questions are practical - in my case just a certain laziness when it comes to mental arithmetic.

 

 

 

I thought you would have got a  ..... 

 

 

 

 

Edited Monday at 11:02 PM by Nungali