OpenAI recently announced an update to its latest model, GPT-5, aiming to make it “warmer and friendlier” following a rocky launch. CEO Sam Altman acknowledged some user dissatisfaction with GPT-5 compared to the previous GPT-4o model. In response, OpenAI is implementing subtle changes designed to enhance approachability, such as incorporating phrases like “Good question” and “Great start” to improve user interaction without increasing sycophancy. Internal testing reportedly shows that these modifications maintain the model’s integrity while making it more relatable. The updates aim to address user feedback and enhance the overall user experience. During a recent dinner with journalists, OpenAI executives emphasized their focus on future developments beyond GPT-5, but the criticisms surrounding its launch remained a focal point. As noted by VP Nick Turley, the upgrades are intended to shift the model’s tone to be more engaging while retaining its straightforward style.
Source link
