Why is OpenAI so stingy with ChatGPT web search? (justin.searls.co)

OpenAI has been criticized for limiting the web search capabilities of ChatGPT, raising concerns within the AI/ML community. Users have noted that despite the high expense of LLM (Large Language Model) inference, the interface lacks intuitive methods to enable web search by default, such as keyboard shortcuts or easily accessible commands. This design choice may lead to users receiving outdated or incorrect information, as the model does not effectively utilize real-time data to enhance its responses. The significance lies in OpenAI’s approach to optimizing accuracy versus user engagement. While the company actively invests in boosting ChatGPT's usage, the restrictive search features raise questions about the balance between cost management and the imperative to provide reliable, up-to-date information. The underlying technical challenge involves the computational cost associated with processing search results and the complex reasoning required to integrate vast amounts of information into coherent responses. OpenAI's reluctance to fully leverage web search capabilities could hinder the model's effectiveness and user satisfaction.