Why do AI models hallucinate?

Anthropic's 5-minute explainer on why LLMs make things up, what they're doing about it, and how to catch it as a user.