OpenAI through their official channel now informs that they are reviewing and investigating several complaints, which state that the GPT-4 model is getting "lazier" than before.
Among the complaints that were present was that GPT-4 did not provide full feedback as before, or responded indifferently, and did not implement what the user requested.
There are also some speculations stating that OpenAI made changes in order to save more processing power and speed of response. However, now OpenAI through the ChatGPT account informs that they are reviewing the complaint saying that the GPT-4 model is getting "lazy".
OpenAI says they have not updated the GPT-4 model since November 11, 2023. And the changes that happened to this model were also not intentional. OpenAI says the behavior of a model is difficult to predict, and they are looking into fixing the problem.
For now, there is no date given for when this matter will be resolved. Also for information, OpenAI has previously temporarily closed ChatGPT Plus account registration, and still has not given a date when it will be reopened to the public.