Apr 20 • General discussion
What Is a Large Language Model, in Plain English
Claude, ChatGPT, and Gemini are all built on something called a large language model, or LLM. The name is off-putting. The idea is simple.
An LLM is a system that has read an enormous amount of text and learned the patterns in how humans write. When you type a question or instruction, it predicts what a thoughtful, well-written response would look like, one word at a time.
That is worth rereading. It predicts. It does not retrieve facts from a database. It does not look things up unless you give it that ability. It produces the most likely useful response based on the patterns it has learned.
Three implications follow from this.
First, LLMs are extraordinary at tasks that involve language. Writing, summarising, explaining, translating, structuring ideas, drafting, editing. If your task is mostly words, Claude will help.
Second, LLMs can make mistakes on facts. They can state something incorrect with total confidence. The word for this is hallucination. It does not mean the tool is broken. It means you should verify anything specific, such as dates, figures, names, or citations.
Third, the quality of what you get out depends heavily on what you put in. A vague question produces a vague answer. A clear, specific request produces a clear, specific response. This is the single most important skill we will teach you.
One useful mental model: treat Claude like a brilliant new hire on their first day. They are smart, fast, and willing. They know a lot about the world. But they do not know your company, your clients, your preferences, or your context unless you tell them. The more context you provide, the better they perform.
In the next post, we'll walk through your first real conversation with Claude, step by step.
What questions do you have about LLMs after reading this post? Let us know in the comments and we'll answer all of them!
0
0 comments
What Is a Large Language Model, in Plain English
