OpenAI said it had received feedback from GPT-4 that it was "getting lazy", but it had not been fixed for almost a month.

Thanks to CTOnews.com netizens Baa Yang, Alejandro86 for the delivery of clues! CTOnews.com, December 9, OpenAI's GPT-4 big language model has been complained by some users, some users said that during this period of time when using ChatGPT or GPT-4 API, they will encounter a series of questions such as very slow peak speed, perfunctory answers, refusing to answer, interrupting sessions, and so on.

According to reports by foreign media such as The Independent, the above problems are usually reflected in this form: if a user requests GPT-4 to enter a code, it may only provide some information, and then instruct the user to fill in the rest. Sometimes, GPT-4 even tells people, "you can do this on your own."

At noon on Friday, Beijing time, ChatGPT officially informed users via the X platform, "We have heard your feedback that GPT-4 is getting lazier and lazier! we haven't updated the model since November 11, of course not on purpose."

CTOnews.com note: although the official personally admitted the above facts, but so far OpenAI has not completely solved this problem, the specific reason has not been disclosed, as to when it can be resolved is also temporarily unknown.

The behavior of the AI big prophecy model may be unpredictable during its use, and CTOnews.com will continue to monitor the progress of GPT-4 's "fix".