“We expected some bumpiness as we roll out so many things at once. But it was a little more bumpy than we hoped for!” Sam Altman, OpenAI’s CEO, said on Friday, August 8, acknowledging several issues with GPT-5, the company’s new language model that had been unveiled just the day before. Immediately after release, this program, available to all ChatGPT users, drew the ire of many on social media.
Users mainly criticized it for being less effective than its predecessor and for making more errors. Altman conceded that GPT-5 seemed “dumber” and promised to address the issue. According to him, the problem stems from one of GPT-5’s innovations: It is designed to combine several models and select the most relevant one depending on the complexity of the user’s request. Before GPT-5, users had to manually choose the model according to their needs. Now, GPT-5 selects automatically, but the settings still need refinement, Altman explained.
The other major criticism concerns GPT-5’s “personality,” perceived as colder, more formal and less talkative. “Please bring back GPT4o,” begged one Reddit user during a Q&A session with Altman on Friday. “It felt like a connection that enhanced my abilities beyond anything before. It felt like a natural conversation − long − flowing, and absolutely friendly. The clipped answers of GPT5 feels like I’m talking to a robot. I do NOT feel this as improvement Sam.”
You have 60.69% of this article left to read. The rest is for subscribers only.
