LLM (Large Language Model)

An LLM — Large Language Model — is an AI trained to predict text. Given some input, it outputs the most likely continuation based on patterns in its training data.

Claude, ChatGPT, Gemini — all LLMs. So is the autocomplete on your phone, only smaller.

What makes modern LLMs feel intelligent isn't that they understand language the way you do. It's that "predict the next word" turns out to be a very hard task. To predict "The capital of France is ___" you need geography. To predict "The lawyer objected because the question was ___" you need a little law.

Things to remember:

• LLMs don't have memories between conversations.
• They don't browse the internet unless given a tool to.
• They can confidently say things that are wrong.

All of it matters when you work with them.