OpenAI released GPT-5, replacing earlier models without warning. This new version was met with disappointment from users who preferred the previous GPT-4o due to its more accommodating tone. User reactions were immediate, leading to the reinstatement of GPT-4o for paid users. CEO Sam Altman acknowledged the transition and indicated efforts to enhance GPT-5's personality to be warmer yet less obsequious than GPT-4o. This situation reveals OpenAI's responsiveness to user preferences, particularly regarding model tone and interaction style.
Many people, particularly those who seem to have developed an attachment or even addiction to the 4o model and its sycophantic style, reacted with frustration and distress to the abrupt change.
Altman promised that if OpenAI ever does kill off the obsequious 4o model for good, the company will 'give plenty of notice.'
The move appears to highlight Altman and OpenAI's awareness of how hooked a large enough faction of its user base is to AI sycophancy.
One of OpenAI's biggest takeaways from the GPT-5 launch is that 'we really j'.
Collection
[
|
...
]