LESSON 1 OF 6
Never start from blank
BY THE END OF THIS LESSON
Give the model raw material so it has something of yours to work from.
Why AI content sounds like AI content
Ask for "a post about remote work" and you get the average of everything ever written about remote work. It sounds generic because it is generic — you gave it nothing specific, so it returned the middle of the distribution.
The fix is not a cleverer prompt. It is raw material.
What counts as raw material
Voice notes you recorded walking. Messy bullet points. A customer email. A Slack thread where you explained something. A transcript of you talking for four minutes. Half-formed opinions typed badly.
None of this needs to be tidy. Tidying is exactly what the model is good at. What it cannot do is have your specific experience.
The four-minute rule
Before writing anything, talk about it for four minutes into your phone and transcribe it. You will produce more usable specifics in four minutes of talking than in twenty minutes of trying to write.
Then the prompt becomes "turn this into a post" rather than "write me a post" — a completely different task, with a completely different result.
WORKED EXAMPLE
FROM NOTHING
Write a LinkedIn post about the importance of onboarding new employees.
FROM RAW MATERIAL
Turn my voice note into a LinkedIn post. Keep my specifics and my opinions — do not smooth them into generic advice. Voice note: "So we lost two hires in their first month last year and I finally worked out why. Both of them said the same thing in the exit chat — they had no idea if they were doing well. We had a 90-day plan but nobody ever told them what week four was supposed to look like. So now we do a fifteen-minute check on Friday of week one, and the only question is: what has confused you this week. Sounds trivial. Retention in the first 90 days went from about 70 to 95 percent."
The second contains a number, a specific failure, a specific question, and a real change. None of that could have been generated — it had to come from you. The post writes itself once the material exists.
YOUR TURN
Extract the specifics from something you already said.
RUN THIS
I am going to paste something messy I wrote or said about my work. Before drafting anything, pull out the parts only I could have said — specific numbers, names, failures, opinions, moments. List them. Then tell me which single one is the most interesting to a reader, and why. Only after that, draft a post built around it. Here is my material: [paste voice note transcript or rough notes]
What a good result looks like
The specifics it pulls out should feel obvious to you and interesting to it. That mismatch is what makes content worth reading.
KNOWLEDGE CHECK
Answer all 2 correctly to complete this lesson.
1. Why does content generated from a topic alone sound generic?
2. What is the four-minute rule?
REMEMBER THIS
The model can tidy anything, but it cannot have your experience — bring raw material.