OpenAI Scrambles to Update GPT-5 After Users Revolt

Open‘S GPT -5 The model was referred to as a world-launch upgrade in its wildlife popular and pre-chatboat. However, for some users, last Thursday’s release felt like a ranking downgrade, the new Chatzipt presents a mixed personality and surprisingly makes a dumb mistake.

Friday OpenAI CEO Sam Altman Has taken x To say that the company will keep running for the previous model, GPT -4, Plus users. Depending on the complexity of the Query on Thursday, a new feature designed to switch seamlessly between models was broken on Thursday, “Ultman said,” And the result was Dumbar like GPT -5. ” He promised to implement fixes to improve the performance of GPT -5 and overall user experience.

Giving a hype around the GPT -5, some levels of disappointment seem inevitable. When OpenAI has launched GPT -4 In March 2023, it was shocked by the incredible power of AI experts. GPT -5, the scholars are estimated that the jaw -dropping will be exactly the same.

OPNA has identified the model as a significant upgrade with PhD-level detectives and virtuoso coding skills. A system of automatically rooting questions in different models was meant to provide a smooth user experience. (It can also save the company’s money by instructing the common questions on cheap models)))

Soon after GPT -5 descended, but a Reddit ChatzPT is the dedicated community Fill in complaint. Many users have mourned for the loss of the old model.

“I have been trying GPT 5 for a few days now. It still doesn’t seem the same even after customizing the instructions it is more technical, more generalized and honestly sensitively,” wrote one member of the community in it Thread The title “Kill 4o is not innovation, it removed it.”

“Of course, 5 OK – if you hate the disadvantages and feelings,” another Reddit user WroteThe

Another Thread Complaints of lazy response, hallucinations and amazing defects.

Ultman ChatzPT doubled the GPT -5 rate limit for Plus users, improves the system that switches between models and promised to solve these problems by specifying users when they want to trigger “thoughtful mode” when they are more worrying and capable. “We will continue to work to stabilize the issues and listen to the reaction,” we mentioned, “we mentioned, we’ve hoped for a bit of turmoil as we roll out a lot at once.” But it was a bit more rude than our expectation! “

Errors posted on social media do not necessarily indicate that the new model is less able than the predecessors. They can only suggest that all the new models have been tried in different ends than the previous versions. Openly refused to comment specially on why GPT -5 sometimes seemed to be common defective.

Some users have spread a new controversy over mental attachments formed with trained chatbots to push their sensitive buttons. Some Reddit users have rejected the allegations about GPT -5 as proof of unhealthy dependence on AI colleague.

In March, Open Research Sensitive bonds are formed with users with its models. Within a short time, the company has issued an update to GPT -4O after that Has become very psychopanticThe

“Looks like GPT -5 is less psychophanic, more” business “and less chat,” Patty Mess, a professor at MIT, who worked on research. “I personally think it as a good thing, because it also leads to confusion, reinforcement of prejudice, etc. Unfortunately many users like a model that tells them that they are smart and amazing and confirming their opinion and belief, even if [they are] Incorrect “

Ultman hints A post on the X at X This is something that the company wrestle on creating GPT -5.

Altman wrote, “Lots of people use the Chatzipi as a therapist or life coach effectively, even if they do not describe it that way,” Ultman wrote. He also added that some users can use the chatzipi in a way that helps improve their lives and others “unknowingly can turn away from their long -term well -being.”