GPT-5 Launch Sparks Backlash Over Limited Improvements and User Experience

The launch of OpenAI’s GPT-5 has sparked widespread frustration among users, developers, and industry observers, revealing growing pains in the evolution of artificial intelligence [1][2]. Marketed as a “significant step toward AGI,” the new model fell short of expectations, with users quickly sharing evidence of its limitations—ranging from incorrect arithmetic to mislabeled maps [3]. Many described the model as colder and more mechanical than its predecessors, leading to a wave of criticism on social media [4].

The backlash intensified as GPT-5 introduced usage limits, initially capping interactions at 200 queries per week [5]. This move angered paying subscribers and prompted OpenAI CEO Sam Altman to backtrack by reinstating access to a popular earlier model, GPT-4o [6]. The company acknowledged the model’s personality issues, promising to enhance warmth and friendliness in future updates, while maintaining that sycophancy levels had not increased [7].

Despite the negative reception, OpenAI remains committed to its long-term vision. Altman has emphasized that GPT-5 is a stepping stone toward more advanced AI, though critics argue that the improvements have been incremental rather than revolutionary [8]. According to research from Princeton University, GPT-5 performed mid-level in scientific, coding, and customer service tasks, with key advantages in cost efficiency and speed rather than intelligence [9].

The reaction to GPT-5 has reignited debates about the direction of AI development. Gary Marcus, a vocal critic, noted that the model had been seen as a symbol of the scaling approach to achieving AGI but failed to deliver [10]. Meanwhile, Yann LeCun of Meta has argued that progress must involve multimodal data and real-world video training, suggesting that current large language models are constrained by their reliance on text-only inputs [11].

For investors, the mixed response to GPT-5 has not dampened enthusiasm. Companies like Nvidia and SoftBank have continued to see strong market performance, with OpenAI’s own revenue from ChatGPT reaching $12 billion annually [12]. Wall Street and venture capital firms remain confident in the AI sector’s growth potential, even as analysts urge caution. Princeton researcher Sayash Kapoor noted that GPT-5 is more about providing reliable infrastructure for business applications than a step toward AGI [13].

The GPT-5 launch highlights a broader challenge in the AI industry: the gap between expectations and practical implementation. While the model’s technical advancements are undeniable, its user experience has not yet reflected that potential [14]. This experience serves as a cautionary reminder that innovation must align with user needs and expectations, especially in a field where hype often outpaces reality [15].

OpenAI has indicated it will continue refining GPT-5 and focus on future models that could deliver the transformative capabilities once promised by earlier iterations [16]. The company remains optimistic about long-term disruption in industries such as search and enterprise software [17].

Source:

[1] GPT-5 Pro: The First AI That's Smarter and Worse at Once (https://natesnewsletter.substack.com/p/gpt-5-pro-the-first-ai-thats-smarter)

[2] Everybody hates GPT-5, AI shows social media can't be fixed (https://cointelegraph.com/magazine/everybody-hates-gpt-5-ai-shows-social-media-cant-be-fixed-ai-eye/)

[3] OpenAI Restores Access To Popular AI Models After GPT-5 ... (https://www.cybersecurityintelligence.com/blog/openai-restores-access-to-popular-ai-models-after-gpt-5-controversy-8642.html)

[4] GPT5 Backlash Reveals AI's Growing Pains, Not Its Failure (https://www.devx.com/daily-news/gpt5-backlash-reveals-ais-growing-pains-not-its-failure/)

[5] Unpacking the Underwhelming Debut of GPT-5 (https://dev.to/grenishrai/the-hype-and-the-hangover-unpacking-the-underwhelming-debut-of-gpt-5-1cc6)

[6] OpenAI's GPT-5 Faces Backlash: Is AI's Shiny Future ... (https://opentools.ai/news/openais-gpt-5-faces-backlash-is-ais-shiny-future-losing-its-luster)

[7] ChatGPT users complain about crowded interface as all-in- ... (http://www.msn.com/en-us/news/technology/chatgpt-users-complain-about-crowded-interface-as-all-in-one-vision-fails-to-impress/ar-AA1KrBi5?apiversion=v2&batchservertelemetry=1&domshim=1&noservercache=1&noservertelemetry=1&renderwebcomponents=1&wcseo=1)

[8] OpenAI's GPT-5 Launch: Sam Altman Admits It "Failed the ... (https://opentools.ai/news/openais-gpt-5-launch-sam-altman-admits-it-failed-the-hype-test)

[9] The biggest loser from OpenAI's new GPT-5 Model isn't ... (https://www.msn.com/en-us/money/companies/the-biggest-loser-from-openais-new-gpt-5-model-isnt-who-you-think/ar-AA1Kl7vA)

[11] What GPT-5 could have learned from Apple's missing ... (https://uxdesign.cc/what-gpt-5-could-have-learned-from-apples-missing-headphone-jack-2ea01fa70aad)

[12] GPT-5 reveals frustrations over AI model progress (https://coinmarketcap.com/community/articles/68a09a8cba700d39f56ef11a/)