Language models & chatbots
A large language model (LLM) is an AI trained on vast amounts of text to predict the next word, which lets it write, answer questions, summarize, and converse.
Your phone's keyboard suggests the next word as you type. An LLM is that idea taken to an extreme. It has read a huge portion of the internet, books, and code, and learned which words tend to follow which. Given a prompt, it predicts one likely next piece of text, adds it, and repeats until the answer is complete.
That simple goal produces surprising abilities. To predict the next word well across all human writing, the model has to absorb grammar, facts, reasoning patterns, and styles. Models with billions of parameters, built on the transformer design, do this well enough to pass exams and write working code.
An LLM does not look things up. Its knowledge is frozen at training time and stored as patterns, not as a database. It can be confidently wrong, and it does not know your private information unless you include it in the prompt. Tools like RAG and web search add the lookup ability.
You ask an assistant to explain photosynthesis to a ten-year-old, and it produces a clear paragraph it has never seen before, built one word at a time.