A former grassroots product manager claiming to have worked in Meituan’s dining department recently posted on Xiaohongshu, listing three “oddities” from the perspective of an insider: First, the role of product manager in the dining business line is highly compressed, with the main focus being on interpreting superiors’ intentions and breaking them down into execution actions, rather than defining problems or identifying opportunities. Second, the massive amount of local life transaction data that Meituan touts to the outside world has not been truly assetized in the specific business chain. Even core questions such as “whether a meal package meets user expectations” lack reliable data-driven representation, with decisions relying more on manual experience. Third, the internal application of AI within the organization is mainly about filling gaps — in the past it was people who filled the gaps, and now large models are filling them — an approach the author calls “wishing well-ization” rather than genuine intelligence.