OpenAI admits GPT-4 may be worse on some tasks after users complained the model is ‘lazier’ and ‘dumber’

• Users of OpenAI’s new GPT-4 model have been complaining in recent weeks about degraded performance.
• OpenAI admitted that GPT-4 performance may be worse on some tasks.
• “There may be some tasks where the performance gets worse,” OpenAI stated in a blog post.

According to OpenAI, the company that created ChatGPT, GPT-4’s performance may be worse on some tasks because users have complained that the model is “lazier” and “dumber” than previous iterations.

“While the majority of metrics have improved, there may be some tasks where performance worsens,” OpenAI stated in a blog post announcing new updates on Thursday.

Users of OpenAI’s GPT-4 have recently complained about poor performance when compared to previous reasoning capabilities.

Some users complained about weakened logic, more erroneous responses, losing track of provided information, difficulty following instructions, forgetting to add brackets in basic software code, and only remembering the most recent prompt on Twitter and OpenAI’s online developer forum.

“The current GPT-4 is disappointing,” wrote a developer who uses GPT-4 to assist him in coding functions for his website. “It’s like driving a Ferrari for a month and then switching to a beat-up old pickup.” “I’m not sure I want to pay.”