After the last update of the platform, so-called “Custom GPTs”, which are completely useless because they always give an error when gpt gives a large text and at the same time there is no input for a re-request, it is the old one that needs to be edited. And in general, it was as if the best feature was removed when gpt did not finish the message, the generate button appeared and added the message based on the context of the previous message, and now it cannot because it has to re-generate the response and it does not then rely on the first written response. New combined chats often give an error if a picture is inserted into them, the old gpt-4 chats began to work very slowly.
Now, when creating a new chat, gpt seems to have become dumber, answers in English if the question is asked in another language, before 4 immediately answered in the language in which the question was asked. I stopped writing code normally, if you ask me to demonstrate how some pattern with a clear description should look, then half of it will give out and half will just describe what should be with comments (this is also a change that is very striking).
Why did you make it worse? if your frontend developers can’t work properly with error handling, hire me, at least I want this product to develop and it will be convenient for users to work with it.
Very upset about what I got with the latest update!