Since there's a lot of shade being thrown about imprecise information that LLMs can generate, an ideal doomsday information query database should be constructed

Since there's a lot of shade being thrown about imprecise information that LLMs can generate, an ideal doomsday information query database should be constructed as an LLM + file archive. 1. LLM understands the vague query from human, connects necessary dots, and gives user an overview, and furnishes them with a list of topic names/local file links to actual Wikipedia articles 2. User can then go on to read the precise information from the listed Wikipedia articles directly.