OpenAI finally fixed the most annoying thing about ChatGPT (www.techradar.com)

OpenAI has launched GPT-5.5 Instant for all ChatGPT users, introducing significant improvements aimed at enhancing user experience. The update boasts a 52.5% reduction in hallucinations—instances where the model generates incorrect or fictitious information—across critical domains such as medicine, law, and finance. Alongside this, responses are crafted to be more concise, avoiding unnecessary embellishments like excessive bullet points or emojis. This streamlined interaction promotes a natural conversational style while enhancing clarity in communication. A notable feature of GPT-5.5 Instant is the introduction of memory sources, which allow users to see and manage the context behind personalized responses. Users can now access information regarding past chats, uploaded files, and linked services to understand how their interactions are tailored. This level of transparency not only empowers users to correct inaccuracies but also optimizes the AI's use of context in future conversations. The update is expected to significantly improve everyday AI interactions, making them more efficient and user-friendly, particularly for frequent users of AI tools.