One of Claude most valuable traits is that it tells you when it does not know something, is uncertain, or cannot help. That makes it more reliable than tools that always answer confidently — regardless of whether the answer is correct.
By default, Claude cannot search the web or fetch recent news. Its knowledge is based on training data with a specific cutoff date. If you need the latest information, use a search tool or provide Claude with the information yourself in the chat.
Every new conversation starts fresh. Claude does not remember what you talked about yesterday. If you want to build on something, provide context again — or use the Projects feature to save background information.
Claude can produce incorrect information with great confidence. This is called hallucination. It happens especially with specific numbers, dates, quotes from well-known people, niche topics, and legal or medical facts. Always verify important information against primary sources.
Claude cannot send emails on your behalf, order products, make calls, or do anything outside the chat window — unless connected to external tools via an API.
Claude will not help with things that could harm others — regardless of how the question is phrased. This is not a weakness; it is a deliberate design choice by Anthropic.
Gratitude is the shortest path to abundance.— Florence Scovel Shinn