Anyone else notice that GPT-4 is become more and more useless?

[worldly basin]

I'm a pretty heavy user of GPT-4 , I use it often through the playground to help me with programming tasks.
Lately (I'd say 1.5-2.5 months ago) I've noticed that GPT-4 is become more and more useless, its making more simple to spot mistakes, adding TONS of fluff to every statement (more than before) even when I prompt it to be concise, it rushes to auto agree with me even after I prompt it to be critical and judge my ideas and its own solutions before answering, I try to give it time "to think" as recommended by the devs on an online tutorial I saw that was made by open ai.
I think they human reinforced it to death to be so careful and un-judgemental that its fast becoming useless to developers when it comes to more than making loops.
I hope they fix this or make another model for developers that dont want or cant use co pilot.