Welcome to DU! The truly grassroots left-of-center political community where regular people, not algorithms, drive the discussions and set the standards. Join the community: Create a free account Support DU (and get rid of ads!): Become a Star Member Latest Breaking News Editorials & Other Articles General Discussion The DU Lounge All Forums Issue Forums Culture Forums Alliance Forums Region Forums Support Forums Help & Search

highplainsdem

(57,719 posts)
13. Skittles, with generative AI like ChatGPT or an image generator like Midjourney, you can get an endless
Thu May 22, 2025, 12:49 AM
May 22

variety of answers to the exact same prompt - at different times or at the same time. With image generators you might get four wildly different images from the same prompt, and text generators could be shown as just as variable with no awareness of what's a correct answer. Except that would instantly undercut the image that LLMs are really thinking and really intelligent.

Microsoft's Bing AI, for instance, used to have two additional modes, Creative and Precise, besides the standard Balanced mode:

https://arstechnica.com/information-technology/2023/03/microsoft-equips-bing-chat-with-multiple-personalities-creative-balanced-precise/

Musk has had different modes with Grok AI, too, which can give wildly different answers.

The AI peddlers could set these AI models up to spit out a hundred or a thousand different responses to one prompt at the same time. But that would be expensive, and no one seeing a chatbot spit out 100 different answers to any question is going to consider it truly intelligent.

Recommendations

1 members have recommended this reply (displayed in chronological order):

AI Vote4Kam May 21 #1
Thank you for your opinion. SalamanderSleeps May 21 #4
OH I get it Skittles May 22 #9
Skittles, with generative AI like ChatGPT or an image generator like Midjourney, you can get an endless highplainsdem May 22 #13
This is a pretty clear explanation of how LLMs work, from a university in Australia: highplainsdem May 22 #15
Welcome to DU LetMyPeopleVote May 22 #22
Wow... 'Guess he hasn't managed to get Elon or others to reprogram for his preferred "resullts" (lies) hlthe2b May 21 #2
This message was self-deleted by its author DUMember24 May 21 #3
Yes right, the list could go on forever. The man knows no sunshine. He hates himself like he loved his father. SalamanderSleeps May 21 #6
A.I. is obviously more intelligent than a particularly stupid Republican... Jack Valentino May 21 #5
Kinda curious what was the prompt you used? fujiyamasan May 21 #7
What is wrong with Donald J. Trumps decisions since 2016....then I just followed the prompts. SalamanderSleeps May 22 #10
Seems spot on here. Although 50% or so of country would disagree. Silent Type May 21 #8
What's scary is anyone using ChatGPT as a source of information. highplainsdem May 22 #11
Do you see any information that you believe is incorrect? Doodley May 22 #12
Large Language Models are what they are. SalamanderSleeps May 22 #17
If you think the unreliability of LLMs is a personal problem, you understand nothing about them. highplainsdem May 22 #20
LLM's do not care about what you, or I, actually think. SalamanderSleeps May 22 #21
As someone once said-- xuplate May 22 #14
Ask about timms139 May 22 #16
There's still hope. SalamanderSleeps May 22 #19
Fantastic work BaronChocula May 22 #18
I mean, okay. Tommy Carcetti May 23 #23
Latest Discussions»General Discussion»So, I asked the ChatGPT w...»Reply #13