THE LESSON
Let's understand the essentials.
A model is a learned system that maps inputs to outputs based on training. A chat product wraps a model with instructions, interfaces, memory, tools, and safety controls. This distinction explains why using the same underlying model through two products can feel different.
A model is not a database of guaranteed facts. Language models generate likely continuations and can produce convincing errors. A tool connection may retrieve information, but that retrieved material still needs evaluation.
Try asking two products powered by different models the same factual question. Compare their sources and uncertainty, not just the tone of their answers.
CHECK YOUR UNDERSTANDING
Explain why an AI chat app is not the same thing as the underlying model.
Explain your answer in your own words, then compare it with the main concepts above.