ChatGPT Updates Following Public Criticism Regarding GPT-5 Developments by Sam Altman
In a bid to address widespread user dissatisfaction with ChatGPT, OpenAI's CEO, Sam Altman, has announced a series of changes to the AI model. The updates aim to improve the user experience and repair OpenAI's relationship with its core user base.
Following the rollout of GPT-5, which was met with criticism over its cold tone and perceived shortcomings, Altman admitted that OpenAI "totally screwed up some things" and acknowledged the controversy. In response, the company has made GPT-5's default personality warmer and more familiar to better engage users.
Key recent updates include improvements in ChatGPT's search capabilities, providing smarter, more comprehensive answers with better instruction following and the ability to run multiple searches automatically for complex questions. OpenAI has also expanded support for custom GPT models, offering more flexibility for enterprise and education users. Ongoing fixes aim to improve response behavior, addressing unexpected "chain of thought" reasoning appearing in simple queries.
Looking ahead, Altman has pivoted publicly to build excitement about GPT-6, promising it will introduce a much better memory system to remember user preferences and habits across chats. This new feature aims to make interactions feel more personalized and continuous over time, unlike GPT-5 which forgets details after each session. However, this raises questions about privacy and data security, which OpenAI will need to address to gain user trust.
The changes also include improvements to the free tier's quality, though concrete details have not yet been provided. The rollout of GPT-5 initially replaced all previous models, leaving paid subscribers restricted to a single AI option. However, GPT-4o has been restored for Plus subscribers, and new response modes, including Auto, Fast, and Thinking, have been introduced.
Critics argue that GPT-5 responses still lack emotional depth and longer answer consistency. Some users demand permanent access to legacy models alongside GPT-5. The company's shift to a unified system raises questions about transparency and control, with the decision to route queries automatically instead of letting users choose their model viewed by some as a step toward reduced transparency.
The competitive landscape, with pressure from Anthropic, Google DeepMind, and Mistral, is likely to influence OpenAI's roadmap. Enhancing the free tier's quality could attract new users and convert them into paying customers. The company must continue demonstrating that these revisions are not temporary appeasements but part of a sustainable approach, balancing innovation with user autonomy.
The past week has shown that user trust is as important as technological advancement for OpenAI. The company's quick recalibration in the face of customer backlash serves as a notable example for other tech leaders.
- To combat criticisms about the cold, impersonal tone of GPT-5, OpenAI is working on implementing a more warmer and familiar personality for the AI model, as announced by CEO Sam Altman.
- In the future, GPT-6 is expected to introduce a superior memory system that will remember user preferences and habits across chats, creating a more personalized interaction compared to GPT-5.
