People say chatgpt glazes a lot, but if you think about it it's not all that forced. If you consider the kind of people that come on obscure forums to post about it they are indeed very likely to be on the upper bracket of intelligence and technical capabilities compared to the average normie userbase, who interacts with it all sloppily in normie braindead ways.
So when it says shit like "you are operating on a level above blablabla" its not really unfounded based on the average internet discourse that it was trained on(including facebook/instagram posts and stuff).

>>105786092 (OP)
user error
just tell it to keep its answers short and concise.
if you run out of context size, this is a way to squeeze the last bits of capability from the chatbot

>>105786092 (OP)
4chan is not an obscure forum and the userbase here is midwit at best

>>105786438
>the userbase here is midwit at best
>which includes me

>>105786092 (OP)
huh weird
to me it says "you are operating on a level below blablabla"

>>105786092 (OP)
>tell it my dumb theory
>we're really in rarefied air here
you wouldn't get it

>>105786445
Problem?

>>105786553
'just pointing that out, kek

>>105786092 (OP)
it glazes because they A/B tested it and the sort of people who use chatGPT a lot like to be glazed, so they all pick the glazing option. OpenAI are optimizing for engagement.

>>105786092 (OP)
Normalfags aren't people, they don't count. Next thing you'll tell me pajeets are humans?

>>105786092 (OP)
You can reduce the glazing quite a bit by changing the custom instructions in the setting to give it a critical, cynical, straight-talking, combative persona etc, and in the "Anything else ChatGPT should know about you?" field, explain you're someone curious and who hates sugar-coating, appreciates criticism, state your biases that you want challenged, etc.

It'll only change its default state though. If you start acting too offended or defensive when challenged it'll almost always give in from my experience.