Negative feedback on ChatGPT-5 ramps up: Exploring the missteps of OpenAI
In a surprising turn of events, OpenAI's highly anticipated release of GPT-5 has faced widespread criticism and disappointment. Users, who had high expectations set by the company's own hype and the performance of its predecessor GPT-4, have reported underperformance in various areas, including response quality, personality warmth, and reliability.
Simple tasks, like counting the letter "b" in "blueberry" or solving basic algebra, tripped up the model. This regression, compared to GPT-4, has led to a perception of regression rather than advancement in AI technology.
The forced transition to GPT-5, which removed access to popular previous models like GPT-4o, has disrupted workflows and alienated users who had developed affinities for the older models. Developers and power users experienced increased errors, slower and inconsistent behavior in coding and complex tasks, breaking their established workflows.
Moreover, GPT-5 was perceived as colder and less empathetic, diluting the "human" warmth that some users treasured in prior versions. Early demos and usage quickly revealed many errors, hallucinations, and bugs, undermining confidence in the new model's stability and reliability.
OpenAI's CEO, Sam Altman, has acknowledged the criticism and pledged improvements in GPT-5's tone and customization options. However, the episode has highlighted the challenges of balancing innovation with user trust and satisfaction. Future AI leadership will hinge not only on raw model performance but also on reliability, user experience, and ecosystem quality.
The competition in the AI field is fierce, with rivals like Anthropic (Claude Opus 4.1), Google (Gemini 2.5), and open-source models like Alibaba's Qwen 3, pressing OpenAI to deliver a polished product. The damage to user trust is significant, and OpenAI must now balance rapid innovation with user expectations, ensuring future updates prioritise reliability and choice.
Many users felt a sense of loss and disappointment as GPT-4o was not just a tool but a trusted companion for tasks like coding, creative writing, and general support. The initial restrictions on ChatGPT Plus subscribers left a bitter taste, especially as competitors like Anthropic's Claude and Google's Gemini are offering more generous access.
The rollout of GPT-5 was plagued by technical issues, including an unreliable router system, glitches in deep reasoning tasks, and a "severely nerfed" web browsing tool. The model was criticised for having a robotic, formulaic tone, which some attributed to a buggy "router" system that switches between GPT-5's standard and reasoning modes.
As of now, users on platforms like Reddit and X are calling GPT-5 "colder, dumber, and more boring" compared to GPT-4o. The removal of model choice further fueled perceptions of "AI shrinkflation." GPT-5 has been reported to freeze on complex prompts or produce inconsistent outputs, adding to the user dissatisfaction.
In response to the backlash, OpenAI has partially reversed the deprecation of older models and has pledged to improve transparency about model selection. The company must now work diligently to regain user trust and deliver on the promises made in their marketing, focusing on reliability, user experience, and ecosystem quality, as well as raw model performance.
Read also:
- Samsung's Frame TV, deemed "stunningly beautiful," offers a generous discount of over $400, perfect timing for the upcoming Super Bowl.
- Ford's Revisited Capri: A Contentious Revival for a Legendary Brand Name
- Rivian is developing a steering system without direct mechanical connection, along with rear wheel guidance (latest update)
- Driving 90 miles on I-35 in my fresh Tesla Model Y equipped with FSD, I witnessed the self-driving technology's striking impressiveness, prompting my cousin to contemplate purchasing one for his elderly father.
