What's Behind a Chatbot? (Meet the LLM)
AI That Generates — Not Just Looks Things Up
You just met a chatbot. Now let's open the hood. Every chatbot you've heard of has a "brain" inside called a Large Language Model — or LLM for short. An LLM is a kind of AI that has read billions of texts and learned how words fit together. LLMs are the engine of what we call Generative AI: AI that doesn't just look things up — it generates brand-new stuff every time.
- Large Language Model (LLM) — an AI trained on billions of texts that learned how language works
- LLMs work by predicting the next word, then the next, then the next — over and over at enormous scale
- The "large" in LLM means the model saw billions of examples while it was learning
- Generative AI (GenAI) is the bigger family — AI that can generate text, images, music, code or video
- Unlike a search engine (which finds existing pages), GenAI makes something new each time
Autocomplete's Giant Cousin
You've already met an LLM's little cousin: the autocomplete on a phone keyboard. Type "see you..." and it suggests "later." Cute — but it only looks at the last few words. An LLM is like autocomplete after a thousand years of homework. It pays attention to your whole question, every single word, and it learned from so much text that it picked up grammar, facts, jokes, and different styles of writing along the way. That's why the same LLM can write a pirate story, explain volcanoes, or help plan a lemonade stand. Same trick as autocomplete — just at a mind-boggling scale.
- Phone autocomplete: guesses from the last few words only
- LLM: pays attention to everything you wrote, all at once
- Scale changes everything — a puddle and an ocean are both water, but only one has whales
- One model, many talents: stories, facts, code, jokes
Prefer slides, quizzes, and saved progress? Read this lesson in the library — free, no sign-up.