How AI actually works
No math, no jargon — just play with the ideas. The core mechanics behind modern AI — text, images, video, tools, and voice — right in your browser.
How a model reads your words
Models don't see letters or whole words — they see tokens. Type something and watch it split.
15 tokens · 58 characters · ␣ marks a leading space (approximate — real models use learned BPE)
How an LLM writes: one token at a time
A model doesn't plan a whole sentence — it predicts a likely next token, adds it, and repeats. Step through it:
The best way to keep up with AI is to
How AI paints from noise
Image models don't “draw.” They start from random noise and remove a little each step — steered by your words — until a picture appears. Step through it:
Prompta neon Charlotte skyline at night, reflections on wet streets, cinematic
How AI does real work: calling tools
A model can't look up live data or take actions on its own. It asks for a tool — your code runs it and hands back the result — then it answers. That loop is what turns a chatbot into an agent.
How many VIP tickets are left?
How AI hears you — and talks back
Talking to AI is really three steps in a row: turn your speech into text, let the model answer, then turn that answer back into speech. Newer models fuse them so it feels like a real conversation.
- You speak
- Speech-to-text
- The model answers
- Text-to-speech
The tools you'll get hands-on with
That's the intuition. Go deeper — hands-on — across two days at the Summit.