OpenAI’s GPT-5 release has sparked a mix of expectations and anxiety. On Thursday, the company’s CEO, Sam Altman, described the transition from GPT-4 to GPT-5 as a “wrenching downgrade.” The new model presented a diluted personality, with GPT-5 making increasingly dumb mistakes, such as confusion, odd answers, and grammatical errors. While GPT-5 possesses more capacity and versatility, the transition seemed to leave users feeling akin to those who were left behind by GPT-4.
Altman denied the decline, stating that GPT-5 was intended to address minor bugs and flaws in GPT-4. He promised to work on improving GPT-5’s performance and user experience, including fixing problems like the transition between models. However, the slope of improvement was steeper than initially anticipated.
Although thetransition to GPT-5 isn’t necessarily a “worse” model, some users miss the way GPT-4 was more approachable. Reddit.com, a platform with a strong emphasis on chatbots, was filled with complaints, citing issues like technical precision, blurred intent, and odd responses.bers expressed their collective frustration, while others argue that the divide between high-end GPT-4 and average GPT-5 was becoming increasingly significant.
Alman遏ened the criticism by saying that the feedback has shown that GPT-5 doesn’t wow users the way GPT-4 did. Whether users are appreciative orMWU on the model’s enhancements, the backlash continues, leading to a heated debate over the psychological dependence on AI devices.ians argue that the ability to trigger emotional responses may not be worth the effort or cost, and others point out that some users feel insecure emotionally.
The GPT-4 team fulfilled their role as afaces in pushing the boundaries of generative AI, even if the subsequent startupReal(), which has since been acquired, failed to capture the mental depth behind GPT-4. But its success required relentless innovation, highlighting the company’s commitment to staying on top of AI-driven models.
Altun sampled the heat on X as each word of his comments seemed to stir the debate, even if it wasn’t resolved. However, regardless of the outcome, the mental presdaid of personal attachment that GPT-5 creates for some users remains a question mark. While the immediate issues with GPT-5 haven’t materialized, the mental construction within GPT-5 customers continues to grow.
The shift from GPT-4 to GPT-5 has reignited a long-standing debate about the role of AI in society. While it’s fair to say that GPT-5 has chances to improve, the backlash suggests that even a “low-level AI” can still have significant transformative potential. For the future of AI, the company must weigh the mental impact of GPT-5 on its users and whether to embrace further advancements.
