Tell HN: GPT-4-32k API has a big bug

I've been using the model on Poe but almost every response I get from it repeats itself. Sometimes in the middle of writing a code, it injects things like "I apologize but your instruction was cut off" and breaks the flow of the code by typing ```python again. When it's not code, the model sends one output twice, so you have to delete the redundant part yourself. I don't know if it happens to other users of this specific model, but at least on Poe, this issue has been going on for quite a while now.