What is important to know about the model's knowledge boundary? Return JSON {"facts": [list of 2+ facts], "why": "briefly"}.
The model has a knowledge boundary (cutoff) - a date after which it knows nothing about the world. Key facts: the model is not aware of recent events, prices, news and updates - but if you ask, she can confidently invent; her “knowledge” is static and does not update itself; For up-to-date data, you need an external source (search, API) connected to the model. Sign of a problem: a question about fresh or changing, and the answer sounds confident, but incorrect. Rule: the model knows the world only up to its learning boundary and does not update itself - for fresh or changing data, connect an external source, otherwise you will get a confident invention.