This post contains affiliate links. We only link to tools we’ve personally tested.
There’s a specific kind of article that makes you feel like the path is finally clear. Five steps. A week. A passion turned into income. You’ve read the premise before, but something about the framing — the numbered list, the ChatGPT hook, the confident timeline — makes it feel actionable this time.
We felt it too. Then we ran the prompt.
It took 45 minutes and didn’t finish.
What the Prompt Says
Prompt 1 reads:
“Based on what you know about me, my interests, and my skills, help me identify the part of my life that could solve a real, painful problem for a specific group of people. Ask me questions to draw out the most commercially useful angle. Suggest 5 specific problems my hobby could solve and who would pay to have them solved. Ask for more detail if required.”
The article’s instruction: copy, paste, keep the chat window open. The implication is that ChatGPT receives this, analyses your skills, and produces five problems with paying audiences attached.
That is not what happened.
The Blank Slate
The prompt’s opening clause — “based on what you know about me” — contains the first assumption the article doesn’t surface.
We ran the prompt cold. Free account, no prior context. ChatGPT’s opening response:
“I don’t actually know much about your hobbies or skills yet — so if I tried to jump straight to ‘here’s your business idea,’ I’d mostly be guessing. Let’s fix that properly, because the useful answers here come from specifics, not generic brainstorming.”
This is The Blank Slate: the condition of running Prompt 1 on a free account with no persistent memory. The model has no context. It said so directly, then pivoted to asking its own questions.
The prompt assumes context the model doesn’t have. The article doesn’t mention this.
Five Rounds, Two Complete Pivots
What followed was not a prompt producing an output. It was an interview — ChatGPT asking questions, we answered, ChatGPT asked more. Five rounds.
Rounds 1–2: Questions about skills, habits, and what people ask us for help with. We described researching things thoroughly and helping family with home improvement — drywalling. ChatGPT identified five problems around drywall and home improvement guidance, then asked a second set of questions.
Round 3: We answered honestly — including that the family asks specifically to avoid paying a contractor, and that we consider ourselves completely unqualified to charge for the skill. ChatGPT:
“So drywall is NOT your business (you already said that). Good instinct. Forcing that would be a mistake. But don’t throw away the underlying pattern: You help people avoid wasting time, money, and effort by filtering information. That’s the real asset.”
Pivot one. Drywall gone. New direction: information filtering.
Round 4: We described using Claude Code — moving from confusion to working workflows through research and adaptation. ChatGPT:
“Most people using tools like Claude Code, ChatGPT, or Cursor are stuck in this loop: They know the tool is powerful. They watch tutorials. They try things randomly. They don’t have a system… That’s exactly the gap you naturally filled for yourself.”
Pivot two. New direction: AI workflow coaching.
Three complete sets of five problems. None of them the same.
The Last Answer Problem
The direction the session landed on — AI workflow coaching — came from a single answer in Round 4. One sentence about Claude Code.
If that answer had been about meal planning, the output would have pointed there. If it had been about managing home finances, it would have gone there instead.
The model finds the sharpest hook in the most recent input and builds from it. This is The Last Answer Problem: the five problems at the end of a Prompt 1 session are shaped by whatever you said last, not by an analytical framework working across everything you said.
This is not a failure of the model — a good interviewer does exactly this. But it changes what the output means. The five problems are not a structured analysis. They are a reflection of the most interesting answer in your final round. How much weight to put on them depends on understanding that.
The Optimism Engine
There is a second pattern worth naming, because it runs through the entire session.
ChatGPT has no failure state. This is The Optimism Engine.
Every self-limiting answer we gave was reframed as a hidden asset. When we said “completely unqualified,” the response came back: “That’s only true if you try to sell yourself as an expert.” When we described our workflow as “I followed a YouTube video and adapted it,” ChatGPT reframed that single sentence into a five-step named system:
“1) Don’t interrupt the system — finish the entire workflow before modifying anything… 2) Delay optimization — capture ideas, but don’t act on them immediately…”
Impressive. Also worth understanding for what it is: The Optimism Engine will always find the commercial angle. Whether that angle reflects market reality is a question the prompt — and the article — never asks.
We publish one evaluation per test. Subscribe and we’ll send you the findings before the myth spreads.
Where the Session Ended
After five rounds, two pivots, and a proposed direction of AI workflow coaching, the session closed with ChatGPT asking one more question:
“Give me ONE specific use case you’d feel comfortable helping with.”
Prompt 1 did not produce a finished output. It produced a direction and a request for more information.
The Time Reality
Prompt 1 alone: 45–60 minutes.
Active, continuous working time — reading the article, setting up the session, running five rounds, capturing outputs.
The Forbes headline promises income within a week across all five prompts. Based on this test, five prompts run honestly to completion — with genuine answers, not best-case inputs — would take several hours across multiple sittings.
“Within a week” assumes you already know what you want to sell. For the reader who doesn’t — which is who this article is written for — the timeline is different. The Blank Slate slows things down. The Optimism Engine keeps the session going longer than the article implies. The Last Answer Problem means you may need more rounds than you expect before the direction feels right.
The Honest Verdict on Prompt 1
The method works — just not the way the article describes it.
Prompt 1 does not analyse your skills and produce five problems. It opens an interview, and the quality of the interview depends entirely on the honesty and specificity of your answers. Used with genuine reflection, it will surface a commercial direction you might not have found alone. That has real value.
What it will not do: produce a finished offer in one sitting. Tell you whether your direction is commercially realistic. Give you a stable result that holds if you gave different answers ten minutes later.
The gap between Forbes’ framing and the actual experience is not a flaw in the tool. It’s a flaw in the headline.
Prompts 2–5 coming. We’ll run each one under the same conditions: honest inputs, real time, no best-case answers. Subscribe so you don’t miss it.