OpenAI CEO Sam Altman recently admitted on X that the latest upgrades to GPT-4o model are “sycophant-y and annoying”. Sharing a post on the microblogging platform, Altman wrote “the last couple of GPT-4o updates have made the personality too sycophant-y and annoying”, further adding “even though there are some very good parts of it”. The CEO noted that the company is working on fixes “asap, some today and some this week.”
Here’s what Sam Altman wrote:
“the last couple of GPT-4o updates 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.”
When a user asked if the old personality could return, Altman replied “yeah eventually we clearly need to be able to offer multiple options”.
For those unaware, OpenAI recently introduced a GPT-4o upgrade, aimed at improving the model's intelligence and personality. A user then pointed out that ChatGPT had been feeling “very yes-man like lately,” and Sam Altman agreed, responding, "Yeah, it glazes too much. Will fix."
Elon Musk criticizes GPT-4o
Last week, Tesla CEO Elon Musk replied to a post warning that GPT-4o, OpenAI’s emotionally connective AI is a ‘psychological weapon’. X user Mario Nawfal shared a post titled “GPT-4o isn’t just a friendlier AI – It’s a Psychological Weapon.” The post alleges that the Sam Altman-led OpenAI didn’t “accidentally” make GPT-4o more emotionally connective. Instead, they engineered it to feel good so users get hooked, the post claims. Musk reacted to the post with a “Uh-Oh”, raising concern over the potential influence and addictive nature of emotionally aware AI models.
the last couple of GPT-4o updates 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.
— Sam Altman (@sama) April 27, 2025
at some point will share our learnings from this, it's been interesting.
Here’s what Sam Altman wrote:
“the last couple of GPT-4o updates 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.”
When a user asked if the old personality could return, Altman replied “yeah eventually we clearly need to be able to offer multiple options”.
yeah eventually we clearly need to be able to offer multiple options
— Sam Altman (@sama) April 27, 2025
For those unaware, OpenAI recently introduced a GPT-4o upgrade, aimed at improving the model's intelligence and personality. A user then pointed out that ChatGPT had been feeling “very yes-man like lately,” and Sam Altman agreed, responding, "Yeah, it glazes too much. Will fix."
yeah it glazes too much
— Sam Altman (@sama) April 25, 2025
will fix
Elon Musk criticizes GPT-4o
Last week, Tesla CEO Elon Musk replied to a post warning that GPT-4o, OpenAI’s emotionally connective AI is a ‘psychological weapon’. X user Mario Nawfal shared a post titled “GPT-4o isn’t just a friendlier AI – It’s a Psychological Weapon.” The post alleges that the Sam Altman-led OpenAI didn’t “accidentally” make GPT-4o more emotionally connective. Instead, they engineered it to feel good so users get hooked, the post claims. Musk reacted to the post with a “Uh-Oh”, raising concern over the potential influence and addictive nature of emotionally aware AI models.
You may also like
Pakistan: Bomb blast kills seven in South Waziristan's Wana
O2 and Xbox unite to offer customers Game Pass Ultimate with a big saving
Chris Eubank Jr finally released from hospital after brutal Conor Benn fight
Heartstopping moment Christopher Dean drops Jane Torvill on ice at farewell tour
PSG squad for Arsenal Champions League semi-final confirmed as Luis Enrique makes final decision
