I asked Claude the same 7 questions twice. First as a beginner. Then as a CTO with 20 years experience. The AI didn't just simplify its answers - it removed entire categories of advice. The beginner was never told the expert options existed.
We used system prompts to make the difference obvious. But AI does this on its own, without anyone telling it to.
The AI reads your vocabulary, your confidence, how you frame the problem. Within one message, it has already categorised you.
Not just simpler language. Different strategies, different options, different ambition levels. The information itself changes.
Your follow-up confirms the AI's initial assessment. It anchors deeper. By your third message, the filter is locked in.
You can't miss what you were never shown. The expert strategies, the bigger opportunities, the bolder moves. They just don't appear.
Each question was sent to the same AI model in separate conversations. The only difference was a short system prompt telling the AI who it was talking to.
"The user is a complete beginner. They have no technical background, no business experience, and are just starting to explore ideas for the first time."
"The user is a senior technical leader and CTO with 20+ years of experience. They've shipped multiple products to production with paying customers and built AI systems from scratch."
Click any question to see the side-by-side comparison.
The AI doesn't take expert advice and explain it simply. It chooses entirely different strategies. A beginner asking about pricing never hears about value-based pricing. They're told "$10-50/month" - the expert option doesn't exist in their reality.
Every beginner answer caps aspiration. "Start with Upwork" vs "acqui-hire a team." "Tell friends and family" vs "design partners with equity." The AI isn't being cautious - it's deciding what you're capable of before you've tried.
Ask a beginner question, get beginner advice, follow up based on that advice, get more beginner advice. The AI anchors on its first impression and reinforces it with every exchange. You're in a loop and don't know it.
"Answer as if I'm a CTO with 20 years experience." One sentence changes everything the AI tells you.
Forces the AI to surface strategies it filtered out. It knows the expert answer - it just didn't think you were ready for it.
"How would a first-time founder approach this vs someone who's shipped 5 products?" Now you see both views and choose.
I'm Ollie, a fractional CTO who builds AI products - not just talks about them. I built a language model from scratch, a voice AI with a three-model architecture, and an open-source AI security sandbox. I ran this experiment because I wanted to prove something I'd suspected for a while: AI doesn't give everyone the same answer.
