Mike Schreiber 
stux⚡️The design is very human
faraiwe@stux The "i" in LLM stands for "intelligence"
Edinbruh
:spinny_cat_inderix:@stux this is what artists and writers are being stolen from to make btw
<p>Technically Batman</p>@stux Almost, it's apologizing for the mistake, some people don't. /hj
Darren@stux I was hoping they would replace me sooner!
Strange Culprits HAS MOVED@stux To quote our hero Beck:
"It's like WOWWWW
It's like RIGHT NOW!"
We think the FAANG tech bros will soon regret all the recent "peace out, we got AI now" layoffs
echo (formerly eve) 👾🏳️⚧️@stux I love the fact that LLMs don't have any internal thought process. It isn't capable of "realizing that it's wrong" before giving you the answer.
The only way it takes new things is from context, and that context is the answer you've got in front of you.
Pukima
Manuel@stux 3.5 even beats that:
> No, 450 is 90% of 500. To calculate 90% of 500, you multiply 500 by 0.9, which equals 450.
https://chat.openai.com/share/1420d7e4-20aa-439b-a301-68f42f8b6f9e
Thibaultmol 🌈@stux sigh... once again... why are people trying to make an llm do math.... they can't...
it's the wrong tool for the job..
it's the wrong tool for the job..
Gorgeous na Shock!@stux I'm as skeptical of LLMs as anything more than glorified autocomplete as the next person, but these things always just make me wonder if I'm actually an LLM and I just don't know. Too relatable.
Phil Rees@stux It's missing the third answer: "That would be an ecumenical matter."
McWabbit🐰 has moved
SibshopsTo bad GPT-4 doesn't actually do this.
https://chat.openai.com/share/12fd63ee-42ed-4cdc-9401-b1323d50ce64
akiran_n@stux is apologizing for being wrong, is not a human thing 😝😂😂😂