This does explain why OpenAI was trying to moderate people's expectations about GPT 4 prior to announcing it. Bing is clearly an improvement over GPT 3.5, but it's not world shattering and still suffers from a lot of the challenges inherent in LLMs.