ChatGPT has stopped taking your prompts so literally, and that’s a big deal (www.techradar.com)

OpenAI has announced an update to ChatGPT, now powered by GPT-5.5 Instant, which focuses on making interactions feel more conversational and less literal. This update addresses common frustrations users experienced, such as the model’s tendency to adhere strictly to prompts without adapting to the conversation's flow. The aim is to improve user experience by allowing ChatGPT to infer intentions, adjust dynamically, and maintain context through conversations, making it easier for users to communicate effectively. While initial tests reveal a slight improvement in the model's ability to adapt and engage more naturally, particularly in text mode, there are still limitations. Users have noted that the model is more responsive to clarifications and appears to retain feedback more effectively than before. However, in Voice Mode, limitations still persist, with the model struggling to understand sarcasm or nuanced responses. Overall, while the update marks a notable step forward in making AI interactions feel more human-like, the journey toward seamless verbal communication with AI remains ongoing.