>>17070582
Classic sign error. I can relate to the AI desu, this is probably one of the most persistent kinds of errors, I remember doing it in elementary school and still occasionally do them now that I'm doing my Master's lol

>>17070586
This is the truest indication we have reached AGI. Congratulations, OpenAI, GPT is truly more human than humans.

>>17070594
It is kind of funny how its errors seem to be becoming more human-like. Older models would produce totally bizarre hallucinations no human would ever do like throwing together random notation to make something that looks "right" and "academic," a sign error is genuinely something human mathematicians do

All my errors are off-by-one errors, so I usually flip a coin (H=0,T=1) and figure that, on average, my work is correct.
The AI is doing the same thing.
It learned the trick from me.

>>17070602
where did you think it learned it from?

>>17070582
*yawn*