@parthknowsai: Chatgpt, DeepSeek, Llama when put the test perform alot better with simple prompts vs complex once. #ai #Tech #chatgpt #LearnOnTikTok #EduTok

Hating Your Life :

my advice has ALWAYS been to talk to your AI like a hardworking intern. They're dumb, but they're a workhorse.

Janani Subramanian :

thank god i only ask chatgpt dumb questions

Landlord decapitator :

simpler words = more trained data = more robust inferences, that's why I've been giving it simple instructions and tell it to be a rigorous researcher

💻 /dev/10x :

Makes sense, is an statistical machine, there is more “simple” vocabulary in the training set

wr :

me try caveman prompts nwo😏

Abdulrahman Sa :

Noa Ranta :

And we got mfs calling themselves prompt engineers✌️😭

Matt W. | ideasocial.pl :

welp I'm only using polished prompts as I am Polish 😭

Fed :

Caveman talk win again. Like da reddit slam at end

Joel Daniel Jimenez :

It's just freedom of choice, the simpler prompts give the AI more options to choose from the more fancy prompt this limits the probability paths the IA can take.

Rob2079🇨🇦 :

I thought intuitively that something like this might be happening. I've found with more directive prompts I get worse output than if I just use casual language. Couldn't prove it but had a feeling. cool.

XXIVandMe :

💯 😂😂😂