Stop Getting Generic AI Responses

Ever wondered if you’re getting the most out of your AI conversations?

As part of my AI Product Design course with MIT xPro, I recently discovered something about how we interact with Large Language Models (LLMs). During a prompt engineering exercise, I decided to test ChatGPT’s knowledge in an area I know well — blockchain technology. What followed was an eye-opening lesson in the art of asking questions.

I started with a broad query: “Should enterprises use public permissionless or private permissioned blockchains?” It seemed like a reasonable question at the time, but I quickly learned why this approach rarely yields useful insights.

“It’s like asking ‘What car should I buy?’ without mentioning if you need it for family road trips or Formula 1 racing.”

Adding Context

Here’s where it gets interesting. As soon as I asked specifically about enterprises exploring healthcare solutions. The LLM become much more insightful. Instead of generic pros and cons, I received targeted insights about handling sensitive patient data and meeting regulatory requirements.

However, I noticed something crucial: the AI quickly defaulted to recommending private permissioned blockchains — the “safe” answer — without fully…