PART 0 · YOUR FIRST STEPS

0.2 · AI in 2 minutes, no jargon

What you will be able to do

Explain in your own words what an AI is, what it can do, what it cannot, and what you do not hand over to it.

On the left a stripped-down engine, on the right someone happily driving a car. — Nobody takes the engine apart before driving. This chapter hands you the wheel, not the mechanics.

In 2 minutes

An artificial intelligence is a program able to carry out tasks we thought were reserved for humans: writing, summarising, answering questions, translating, creating an image.

The truest image is that of the very fast, very well-read intern. It has read an enormous amount, it works in seconds, it never tires. But it needs clear instructions, it does not know your situation unless you tell it, and it can be wrong with disconcerting confidence.

The important point, the one that changes everything: an AI does not consult a store of truth. It works out what is most likely, word after word, from everything it has read. Most of the time, the most likely happens to be true. Sometimes not.

1 · What you write	2 · What it does	3 · What you get
“Why is the sky blue?”	It works out the most likely sequence of words. By default, it goes and checks nowhere.	A plausible answer. Often right. Not guaranteed.

Note

One exception, which you will see in Part I: if you ask it to search the web, Claude will read real pages and cite its sources. But that is not what it does by default. Until you have asked, assume it is answering from memory.

Let's do it together

Nothing to install for this chapter. Three ideas to anchor, and they will serve you through all the rest of the handbook.

Read this sentence twice: an AI predicts a plausible answer, it does not consult a store of truth.
Remember the consequence. When it does not know, it does not stop: it invents something credible. This is called a hallucination. It is not a rare bug, it is normal behaviour pushed to its limits.
Remember the reflex. For anything that really counts, a figure, a date, a reference, a point of law, a piece of medical information: you check elsewhere. For the rest, you can move fast.

Warning

An AI that invents gives no warning. It changes neither tone nor style, and the false is written as well as the true. That is precisely what makes checking necessary: you cannot spot an error from the way it is written.

The seven words you will come across

Before going further, seven words. Seven, not fifty: these are the ones that turn up everywhere from the first hour, and knowing them will stop you getting stuck on a menu or a message.

The word	What it means, in one sentence
Prompt	Your written request. “Prompting well” simply means putting it well.
Iterate	Ask, look, correct, ask again. The normal method, not a failure.
Hallucination	When the AI states something false with confidence. What this chapter has just explained to you.
Model	The “version” of the AI you are using. There are several, more or less powerful.
Context	Everything the AI has in front of it as it answers: your request, the conversation, the attached files.
Credit	The currency of the image studio in Part II. Every creation uses some up.
Beta	A product still being run in: it works, with a few acknowledged rough edges.

The others will come by themselves

You will come across other words: MCP, agent, canvas, Plan mode. You do not have to learn them now: each one is explained where it is useful, at the moment you need it.

And if a word blocks you along the way, the full glossary is in appendix A, with fifty-one terms. It is also in the troubleshooting file delivered with this handbook, where you can search for it by typing the word.

What you do not hand over to an AI

You will soon be putting your own documents into Claude, and that is the whole point. One simple rule before you go there: what you write is sent to the publisher's servers. There is nothing worrying about that for a set of meeting notes. There is rather more for these five things.

Passwords, codes, access keys. Never, whatever the tool.
Identity documents and official numbers.
Health data, yours or anyone else's.
Personal data belonging to your clients or colleagues that you have no right to pass on. Anonymise before pasting.
Professional secrets covered by a confidentiality agreement.

For everything else, your notes, your drafts, your ordinary working documents: go ahead. And do go and look once, in your Claude account settings, at the section on using your data to improve the models: the setting exists, it changes with one click, and it is better to have seen it now than to discover it in six months.

Your turn

Find a field you genuinely know well: your job, a passion, your town, a sport. Note it next to your sentence from chapter 0.1.

We will use it in the next chapter. It is the best training ground there is: on a subject you have mastered, you spot immediately when the AI is right and when it is padding.

You have got it if…

Your field is precise enough for you to correct a mistake in it (“accounting for small businesses”, not “work”).
You can explain in one sentence, to someone else, why an AI can be confidently wrong.
You can name two things you will never paste into a conversation.

Traps to avoid

Believing everything the AI says because it is well written. How well something is put says nothing about whether the content is accurate.
Thinking it understands the way a human does. It works out probabilities, it does not understand. That takes nothing away from its usefulness, but it explains its mistakes.

What to remember

An AI carries out everyday human tasks: writing, summarising, answering, creating.
It offers the likely, not the true. When it does not know, it invents without warning.
What matters gets checked elsewhere.
Passwords, identity documents, health, third-party data, professional secrets: they do not go in.

← Welcome, and how to use this handbook Open your first account →

The rest of the handbook covers Claude, Higgsfield, Claude Design and Claude Code, with 11 recipes and a 30-day programme.

See the full handbook Open in the reader

Hear about it when the manual changes

The manual gets reworked as the tools change: new versions, screens that move, recipes brought back up to date. If you want to hear about it when that happens, leave your address. It is optional: the free extract stays readable with no account and no email address, and nothing closes if you skip this.

Your address is used only to announce updates to the manual. It is never sold or passed on. To unsubscribe, write to contact@atelieriacomplet.com. Personal data

Written by Sylvère Bamenou, Cotonou

This handbook is not an expert's course: it is the roadmap of my own learning of these tools, written for the least technical among us.

Read the full story