What AI Actually Is (and Isn't)
Separate the hype from reality. Test your assumptions with an interactive quiz, then explore how Large Language Models actually work through an animated diagram.
Myth vs Reality Quiz
For each statement, decide: is it a Myth or Reality? Click to reveal the truth.
How an LLM Actually Works
Click each stage to explore how a Large Language Model processes your prompt and generates a response. It's not magic — it's pattern matching at an extraordinary scale.
Your Prompt
You type a message. The model receives raw text — it has no idea who you are, what you look like, or what you had for breakfast.
Tokenisation
Your text is split into tokens — chunks of words, subwords, or characters. "Understanding" might become ["Under", "stand", "ing"]. Each token maps to a number.
Embedding & Attention
Each token becomes a high-dimensional vector — a position in "meaning space." The attention mechanism then figures out which words relate to which. "Bank" near "river" means something different than "bank" near "money."
Next-Token Prediction
The model calculates the probability of every possible next token. It doesn't "know" the answer — it predicts what text is statistically most likely to follow. This is the core mechanic: autocomplete on steroids.
Token-by-Token Output
One token is selected, appended to the response, and the whole process repeats. Each new token considers everything before it. This is why responses appear to "stream" — they're being built one piece at a time.
Key insight: LLMs don't understand anything. They're incredibly sophisticated pattern-matching machines. This is why they can sound confident while being completely wrong — they're predicting plausible text, not verifying facts.
Try It Yourself
Open any of these AI tools and start a conversation. Notice how the response streams in token by token — now you know why.