← Knowledge

Interactive · Guide

What Is AI, Really?

6 topics from movie myths to grounded reality — scroll to read

1

What Is AI, Really?

Separating reality from what movies taught you

AI stands for Artificial Intelligence — sounds scary like in the movies, but it's really just software that learns from examples instead of being programmed with explicit rules.

📚

Analogy: Instead of writing rules like "a cat must have 2 ears, a long tail, soft fur...", you just show it 1 million cat photos and it figures out what cats look like on its own — like someone who's watched a thousand movies and just knows how movies work.

Click each card — true or false?

💡

AI doesn't "think" or "feel" — it predicts patterns from what it's seen before. Brilliant at things it was trained on, surprisingly bad at everything else.

2

The AI Family Tree

AI → ML → Deep Learning → Generative AI

"AI" is the broadest umbrella — everything else nests inside it like Russian dolls. ChatGPT isn't all of AI — it's the innermost box, the most specific layer.

🍕

Analogy: Like "Food" → "Italian Food" → "Pizza" → "Margherita Pizza" — ChatGPT is a specific recipe deep inside, not a stand-in for all of food.

Click each ring to see details

💡

ML is a subset of AI. Deep Learning is a subset of ML. Generative AI is a subset of Deep Learning. People use these terms interchangeably all the time — but they're actually nested layers.

3

How Does AI Learn?

3 main ways AI trains itself

AI doesn't "learn" the way humans do — it adjusts numbers in a formula until predictions get better. But the way you feed it data and feedback has 3 main approaches.

💡

Modern AI typically combines Supervised + Reinforcement learning. ChatGPT trained on trillions of words (supervised), then humans rated responses to fine-tune it (reinforcement).

4

How Smart Is AI, Really?

Narrow AI · AGI · Superintelligence

Every AI that exists today — including ChatGPT, Gemini, Claude — is Narrow AI. Good at specific tasks it was trained on. Not the all-powerful robot from movies.

Exists ✓
Narrow AI
Research
AGI
Theory
ASI
💡

ChatGPT sounds fluent ≠ AGI — it's brilliant at language but gets 7-digit math wrong, doesn't know what happened after its cutoff date, and can't tell you what today's date is. That's Narrow AI by the textbook.

5

AI You Already Use Every Day

Tap each time slot to reveal

Before ChatGPT was a thing, you'd been using AI for years — it was just hidden inside every app on your phone.

Your day

0 / 8 unlocked
💡

TikTok's For You algorithm is arguably the most accurate recommendation AI in the world right now — because its feedback loop is extremely tight. Scroll past = dislike. Watch to end = like. Updates in real-time with every single action you take.

6

Where Do You Go From Here?

3 paths — pick the one that fits your goal

You don't need to be a programmer to benefit from AI — just like you don't need to be a mechanic to drive a car.

💡

90% of people get the most value from AI through the "Use it" path — you don't need to know how neural networks work, just like you don't need to know how an engine works to drive.

← Back to Knowledge