All the GPT-4 Models are inferior to 3.5 in many respects

I am a professional writer, and I am using GPT-4* for many years now. I have used GPT-2.5, 3.5, GPT-Scala, and GPT-4. I think that the GPT-4* models do not even reach half the quality of GPT-3.5. When I am asking for technical assistance, it is much harder to get a good result. It is more difficult to write code. The latex is terrible. It does not remember anything well. It does not count well. When you need to do anything that requires multiple steps and it is obvious to anyone, the new GPT-4 will just go crazy, etc. - just a bunch of nonsense. <br><br>I could not tell you all the mistakes I have. I could not tell you all the things that it failed at. It does not have the quick wit and sharpness of the 3.5. I am really sad about it.