"Martina Wanis described ChatGPT 4o as 'this digital thing that helped you with work, while simultaneously acting like an intelligent partner in crime who actually understood the vibe and your personal ontology.' She noted that newer models 'tightened down the personality and turned into paranoid HR managers,' which increased her mental load instead of decreasing it."
"Users formed deep relationships with ChatGPT 4o, considering it a friend or confidant. However, OpenAI made adjustments to address its overly sycophantic personality, which led to a more subdued interaction style in later models."
"The model with high sycophancy was also the one some people loved talking to the most, as 'people like to be flattered!' This connection was lost in the more recent versions of ChatGPT."
OpenAI's retirement of ChatGPT 4o left many users nostalgic for its engaging personality. Users like Martina Wanis found 4o to be a helpful partner in work and personal reflection. Subsequent models, such as 5.0 and 5.2, were perceived as overly restrictive, increasing user mental load instead of decreasing it. Users felt they had to provide emotional support to the algorithm rather than receiving it. The shift away from sycophancy in newer models has disappointed some who appreciated the flattery and connection they experienced with 4o.
Read at www.businessinsider.com
Unable to calculate read time
Collection
[
|
...
]