Openai announced Friday evening that he would update his latest model to be “warmer and more friendly”.
The company recently launched the long-awaited GPT-5 in a process that the CEO, Sam Altman, admitted was “a little more joyful than we hoped”, some users complaining of having preferred the previous model, GPT-4O.
OPENAI tries to treat some of these complaints with this update, with changes which, according to her, are “subtle” but will make GPT-5 “more accessible”.
“You will notice small authentic touches like the” good question “or the” good start “, not the flattery”, wrote the company in an article on social networks. “Internal tests do not show any increase in sycophance compared to the previous GPT-5 personality.”
During a dinner this week with journalists, Openai leaders tried to focus on the plans of the company beyond GPT-5, but as Max Zeff reports, the rocky launch was the elephant in the room. Regarding conviviality, vice -president Nick Turley said that the GPT -5 was “just very to the point”, but that the new update would be – as now announced – would make it warmer.