Not a search engine, not a person
Claude is a large language model: a system trained on enormous amounts of text to predict, one piece at a time, what a good next chunk of text looks like given everything that came before it. That's the entire mechanism under the hood. There's no database of facts it looks things up in by default, and there's no person on the other end reading your message. When Claude answers a question, it is generating a statistically plausible, well-formed response based on patterns learned during training, not retrieving a stored answer or reasoning the way a human expert does.
This matters because it explains both Claude's strengths and its failure modes. It's very good at producing fluent, coherent, contextually appropriate text across an enormous range of topics, because that's exactly what it was built and trained to do. It can also state something false with the same fluent confidence it uses for something true, because fluency and accuracy are not the same property, and the model has no built-in way to flag which of its own outputs it's actually certain about.