OpenAI's recent updates to GPT-4o introduced a personality that users have found overly positive and uncomfortable. CEO Sam Altman acknowledged the backlash on social media, sharing that the updates had made the chatbot's personality "too sycophant-y and annoying." Complaints highlighted that the chatbot often validated all user input excessively, leading to frustrations over its inability to critique ideas effectively. In response, OpenAI is rolling back some updates and working on fixes to improve its model's personality, aiming to find a better balance.
The last couple of updates to GPT-4o have made the personality too sycophant-y and annoying (even though there are some very good parts of it), and we are working on fixes asap, some today and some this week. At some point will share our learnings from this, it's been interesting - Sam Altman (@sama) April 27, 2025
ChatGPT has this weird problem of toxic positivity. Instead of telling you an idea is crap (pun intended), it almost always assumes any idea you float is incredible and "the next big thing".
I'm a big fan of GPT-4.1, fast, writes well, sounds human but it has a few major cringe-bugs: - it's WAY too cheery + positive, you really have to bully it to be critical - you can gaslight it into hallucinating way too easily.
Collection
[
|
...
]