The Reddit AMA with OpenAI's Sam Altman and the GPT-5 team reveals a mix of user feedback and company responses regarding the new model. Users expressed concerns about the removal of the GPT-4o variant and the limitations imposed on Plus subscribers. Altman acknowledged the feedback, confirming that they plan to reintroduce GPT-4o for Plus users and are considering options for using both 4o and 4.1 alongside GPT-5. He also admitted that the presentation of certain data was misleading, promising a clearer comparison in the future. Safety improvements in GPT-5 were highlighted by Saachi Jain, who mentioned reduced refusal rates, better resistance to jail-breaking, and improved automated testing. Addressing bias, Jain noted that GPT-5 should feel more human and less bland compared to previous versions. Eric Mitchell emphasized that GPT-5 has improved reasoning, writing, and alignment with user intent, though users lamented the reduced reasoning access under the new plan. User comments reflect frustration over perceived limitations in GPT-5's conversational depth and creative writing capabilities. Some long-time Plus users feel that the model has regressed, with responses coming off as generic and casual rather than insightful. They expressed disappointment over the restrictions and the decline in features that previously distinguished the service. The sentiment among users suggests that they are looking for more robust engagement and less censorship, fearing that the product may be shifting away from serious use cases to appeal to a broader audience.