About two weeks ago, Openai released GPT-5. The most recent AI model of the GPT line, GPT-5 has been presented as the most intelligent, fastest and most useful model of the company “with” integrated thought “and” expert level intelligence “. But the liberation turned against an important reason.
Part of the changes in the GPT-5 consisted in tackling the sycophetic positivity found in the previous models, where the IA chatbot would constantly praise the user to a degree of cancellation and would categorically accept that the user feels better. Many users have hated this, so GPT-5 was done to be “less effective” and “use fewer unnecessary emojis”.
But GPT -5 faced his own reaction – to be too cold, too far away, too much business. Now, the latest news is that Optai deploys an update that will make the GPT-5 “warmer and more user-friendly”, reports Techcrunch.
The changes include encouraging sentences such as the “good question” or the “good start”, but no flattery and outright shaping. “Internal tests do not show any increase in sycophance compared to the previous GPT-5 personality,” said Openai in an article on social networks.
This article originally appeared on our publication Sister PC För Alla and was translated and located in Swedish.
