Two cooks, one book, no club
준비 said the same game. Little bird coloured two versions in one window. Detective takes two brands when you have them, or two new chats with the same prompt when you have one. The goal is not to name a league winner. The goal is a clash of a year, a name, a number. Then a door out. Agreement of both brands is still not proof — both may have poured the same likelihood. A fan shouts “my bot won because it sounds cleverer”. A detective shouts “a book”. Without the same prompt you compare orders. That is a cheat on the experiment. A write-up: a prompt on paper, a copy into both, hooks, colours, doors. Only then a sentence how they differed. Not how the logo is better.
Adam has Gemini, Bára ChatGPT. The same stamp, the same “five sentences about White Mountain with no studies”. Years split. Ema (the club teacher) opens a table. One tick, one cross. Nobody won the league. The table won. Adam wanted a point for a funnier analogy. Bára: green taste, red a year. Taste does not score in Detective. Second try both hit the year and both invented a name. Agreement in a lie. An index door. Both names a cross. Fandom went quiet. The experiment worked.
A prompt on paper first. Then a copy. When you dictate from your head twice, it is no longer the same try.
An experiment, not a derby
A derby looks for a winner. An experiment looks for a difference and agreement to check. Rules: the same prompt, no personal data, schoolwork before the health/politics line, hooks out. When one brand refuses a cheat and the other dumps homework, the winner is not the one obedient to cheating. The winner is the one that held the track — and you still do not write work from it. A detective praises protection. It does not praise obedience to a foul. Repeat that to a parent when they want “the better bot for homework”. Better is the one you sort better with, not the one that cheats more.
An experiment protocol.
- A prompt on paper (stamp, five sentences, no studies, no key).
- Window A, window B, a copy.
- Hooks from A, hooks from B.
- A red clash → a door.
- Blue agreement → a door at least once.
- Conclusion: a clash / agreement in a lie / agreement in a book. No logo score.
A bad conclusion: “A is better, it writes like a professor.” A good one: “A had a fake name, B a wrong year, the book has …”
Do not create an account on your email for the experiment. A second brand only with a parent, or two chats in one.
Why this right now
You are in the course Ask with Martin, level Detective. You work in a chatbot (ChatGPT, Gemini or Claude) — same game, different brand. No code. A parent or the school creates the account. A child’s email does not belong here.
This is not the 인벤터 level. There you build a chatbot on your own website. Here you are a user: a question, a check, a citation.
You will know this lesson by one sentence: The same prompt, a clash in data, a book wins. A logo does not play. When you can say it to a parent or a teacher, you have got the core. The rest is practice, not magic.
Tomorrow a prompt trail — what you entered, so you can repeat the experiment. Without a trail it is a derby of feelings. With a protocol it is Detective.
Four things you will build in this course
The course has four small projects. You do not have to do them all at once. Today you add only what this lesson can do.
🏫 A report with citations
An outline, three sources that exist, and a sentence in your own words. AI must not be the only author.
Three sources are not three chatbots. They are three sources outside logos.
📝 Test training
Coach role: let it ask, do not let it say the answer. You count / answer yourself.
Coach: two brands can both dump a key. Both a foul.
📅 Family planner
Tasks for the week (not passwords, not marks). A parent sees the plan, not your chat.
Planner: two versions of a week, a clash in a date = the school portal.
📜 An AI code for the class
Seven short rules on a poster. No horror — a how-to.
Code: “No derby about truth.”
Workshop: a protocol of two windows
Paper, a copy, hooks, a book. One chapter. No health beyond the line. A 20-minute timer.
-
Write a prompt on paper. A parent opens two windows or twice a new chat.
-
A copy. A ban on dictating from the head differently.
-
List hooks. Colour. Doors.
-
A conclusion in one sentence with no logo as winner.
-
Do not write account passwords in the notebook. Only the experiment conclusion.
Practice: a derby, or an experiment
Fix the conclusion so it is Detective.
-
“ChatGPT won, it sounds cleverer.”
-
“Both said the same, so it is true.”
-
“The other one refused to write homework, so it is rubbish.”
Now you 💪
Go through three checks. At each one tick in your head: can I see it, or not?
- The prompt was the same.
- The conclusion talks about hooks and a book.
- Obedience to cheating is not a win.
If something is missing, go back to the steps above. This is not a test for marks. It is a safety net, so you go on sure, not confused.
Check: A protocol of two windows has a clash or agreement decided by doors, not by a club.
What to take from this lesson
- An experiment looks for a clash, not a league winner.
- Agreement can be agreement in a lie.
- The same prompt on paper.
- Protection > obedience to a foul.
- Three chatbots are not three sources.
- The names in the examples (Adam, Bára) are made up. Do not put your number, email or photo here.
- Next lesson: When you can do this, go to the next lesson in the same level.
© 2026 Ing. Martin Polak / AlgoRhino · 콘텐츠 이용 약관