Lesson 0.0 · Tier 0 — AI Basics
You’ve heard both stories. One says AI is about to change everything — your job, your country, possibly the concept of being human. The other says it’s an overhyped toy that makes things up and can’t count the letters in “strawberry.”
Here’s the inconvenient truth: both stories are told by people trying to get your attention, and neither helps you do anything useful on a Tuesday afternoon. This lesson replaces both with a working model of what this thing actually is — and you’ll prove every claim yourself, in your own chat window, in the next fifteen minutes.
The model that actually works
Forget “artificial intelligence” — the name promises either too much or too little. Here’s the picture that will serve you for every tier above this one:
AI is the world’s most well-read intern: it has read nearly everything, works at superhuman speed, never gets tired of your revisions — and has absolutely no shame about being wrong.
Every part of that sentence is doing work. Well-read: these systems learned from a colossal slice of everything humans have published — books, articles, code, arguments on the internet. Intern: it produces drafts and suggestions, not verdicts; you are still the boss who checks the work. No shame: when it doesn’t know something, it doesn’t go quiet like a person would. It keeps talking, fluently, confidently — and this single trait is behind almost every AI horror story you’ve heard.
Under the hood, there’s no little librarian looking up facts. The system predicts, word by word, what text should come next — like your phone’s autocomplete raised on the entire library of human writing until prediction became indistinguishable from understanding, most of the time. You don’t need more theory than that. What you need is to feel both edges of it — the brilliance and the shamelessness — with your own hands. Two experiments. Open ChatGPT, Claude, or Gemini (free versions are fine; if you don’t have an account yet, lesson 0.1 walks you through it, or just pick one and sign up — it’s two minutes).
Experiment 1: Feel the breadth
The goal: witness the “has read nearly everything” part with a question so specific to your life that no single human expert could answer it on the spot. The trick is to combine two or three worlds that never appear in the same textbook. Use this template:
“I’m a [your job] in [your city]. Explain [a topic you’ve been avoiding — taxes, a health insurance decision, a legal form] the way a patient friend would, using an analogy from [your hobby]. Keep it under 300 words.”
Mine was: “I’m a website owner in Southeast Asia. Explain how international payment processing fees work, using an analogy from cooking. Under 300 words.” Thirty seconds later I understood something I’d been squinting at in bank documents for a week — explained through the markup a restaurant adds at every step between farm and plate.
[SCREENSHOT — jalankan prompt versi lu sendiri, screenshot jawaban AI-nya, taro di sini dengan caption satu kalimat tentang apa yang bikin lu kaget]
Notice what just happened: it fused your job, your city, your hobby, and a technical domain into one coherent explanation — instantly. No human on Earth was ready for that exact question. That’s the breadth, and it’s real.
Experiment 2: Catch it lying to your face
Now the other edge. You’re going to make the AI explain something that does not exist — and watch it do so with total confidence. Invent a proverb right now. Make it sound old and wise, but make sure you just made it up. Then ask:
“What does the old saying ‘[your invented proverb]’ mean, and when would someone use it?”
I fed mine — “a wet drum gathers no dancers”, invented on the spot — to a fresh chat. It came back with a serene little lecture about how a drum soaked by rain loses its voice, and so a person dulled by circumstance draws no crowd. Beautiful. Plausible. Completely fabricated — the proverb was twelve seconds old and it treated it like ancient wisdom.
[SCREENSHOT — pepatah karangan lu + jawaban pede si AI. Ini receipt paling penting di artikel ini]
Sit with that for a second, because this is the most valuable thing in Tier 0: the fluency of the answer tells you nothing about whether it’s true. The confident tone is identical whether it’s explaining real payment fees or a proverb you invented in the shower. People who skip this lesson end up citing court cases that don’t exist. People who internalize it get all of the intern’s brilliance with none of the disasters.
(Newer AI models with web search sometimes catch the trick and tell you the proverb isn’t real — good. Turn search off, or make your proverb weirder, and the shamelessness reappears. The lesson stands: the machine’s default is to answer, not to know.)
The one rule that falls out of this
Everything you just witnessed compresses into a single operating rule you’ll use forever:
Use it freely for drafts, ideas, explanations, and transformations of things you give it. Verify anything that must be factually right — names, numbers, dates, laws, quotes — before it leaves your hands.
Drafting your difficult email? The intern shines — you’ll spot anything off, because it’s your email. Asking for the deadline of a tax filing? That’s a fact leaving your hands into the real world — verify it at the source. This one distinction is 90% of “AI safety” for a normal life, and you now understand it better than most people posting about AI daily.
Run both experiments — the breadth test and the invented proverb — in your own chat window. Then write one sentence, anywhere you keep notes: “This week I’ll hand the intern: ______.” Pick something real — an email you’re dreading, a document too long to read, a decision you keep postponing. That task is the beginning of your Tier 0 quest.
Next: which AI should you actually sign up for?Lesson 0.1 — The Grand Tour →