Chapter Zero · What Exactly Is AI

What Exactly Is AI? It's Really Just Playing "Finish the Sentence"

Don't let the words "artificial intelligence" intimidate you. ChatGPT, DeepSeek, Doubao… whatever the name, what they do boils down to one thing: guessing the next word of your sentence. No background needed for this lesson — play two little games and it'll click.

Step 1 · You were born knowing this game

Forget AI for a second. The three sentences below are all unfinished. Tap each one open and see whether the ending that pops into your head matches what most people say.

Notice something? Some sentences have only one way to finish ("Happy birthday to…"), while others have several plausible endings ("The weather today is really…"). In a split second, your brain answered a multiple-choice question — and that's exactly the question AI answers at every step it writes. The only difference: it has read millions of times more than you have, so its guesses are remarkably good.
Step 2 · Watch the AI finish a sentence with your own eyes

Pick an opening line and click "Let the AI continue". The box below is the multiple-choice question in its head at every step, slowed down for you. Two things to watch. First, it does not write one word at a time: "wonton" arrives as won + ton, "check-up" as check + - + up. That unit is called a token — sometimes a whole word, sometimes just a fragment of one; a later lesson covers how the pieces are cut. Second, the length of the bars: at some steps the preceding text all but forces its hand, at others you can watch it waver. Run the same opening a few times — where the bars are close, it doesn't always take the longest one.

Pick an opening:
Worked overtime until 10 tonight,
It's picking the next token…
No lookups, no internet — it continues token by token purely from "the text it has read"
Step 3 · So what about "chatting"?

You might ask: I get the sentence-finishing part, but the AI and I are having a back-and-forth conversation, aren't we? Actually, chat is just sentence-finishing in new packaging —

You
Help me think of somewhere to go this weekend
↓ To the AI, your message is the "first half" — its job is to supply the "second half" ↓
AI
You could go hiking just outside town, or check out a newly opened exhibition. Do you prefer somewhere quiet or somewhere lively?
You
Somewhere quiet
↓ It stitches all the previous messages into one longer "first half" and keeps finishing it ↓
Then I'd suggest the botanical garden or an art museum — fewer crowds, and places to sit when you're tired of walking…
It finishes sentences so well, so naturally, that it looks like it "gets you". If there's one line this lesson would love you to take away, it's this: it has no thoughts of its own — what it has are statistics, gathered from hundreds of millions of books and billions of webpages, about "which words usually follow which words".
This one intuition explains everything that comes later
✍️

Why can it write essays?

It just finishes sentences at length, fluently. Give it an opening (a topic and requirements), and it continues token by token until a whole piece comes out.

🤥

Why does it spout nonsense?

It only cares whether a continuation sounds right, not whether it is right. A smooth-sounding falsehood beats a clumsy truth in its selection. This matters a lot — it gets its own lesson later.

🎲

Why different answers to the same question?

In the demo above it didn't always take the longest bar — where a few candidates are close in likelihood, it rolls a die weighted by those odds and wanders down a different fork of the sentence. At a step where one bar reaches 99%, any roll gives the same result, so the differences only ever show up where it was already wavering.

While we're at it, an old mystery · Why is it bad at math?

You may have heard that "AI gets even elementary-school math wrong". Now that you know about finishing sentences, you can solve this mystery yourself — it was never "calculating"; it's finishing sentences. Send it both of these problems and see how different the candidates in its head look.

See the difference? It has seen "1+1=2" millions of times in its training data — effectively memorized; but it has almost never seen this exact "3847×2963" problem, so it can only pick a number that "looks like an answer" — right number of digits, plausible start and end, but the middle digits are guessed. The good news: many AI products now have the AI call a real calculator or write code for math problems, turning "guessing" into "calculating" — a clever division of labor you'll meet again in later lessons.

✅ What this lesson wants to share with you