Do you have a folder full of saved prompts or templates you thought would make AI work better- but never actually use?

In this episode, Gil and I share how relying on prompt templates and “perfect prompts” often limits what you get out of tools like ChatGPT and Claude. We break down the mindset shift that helped us go from copy-pasting prompts to having real conversations with AI-leading to better learning, faster results, and outputs we’re actually excited to use.

Key discussion points:

• Why prompt libraries and templates can actually hold you back
• The difference between prompt engineering vs. iterative thinking
• How to structure AI conversations to learn while doing
• Using AI to fill knowledge gaps and test your understanding
• Why ending your first prompt with a question improves results
• The surprising role of system prompts in shaping AI behaviour
• How mindset and frameworks beat tools in the long run

Notable quotes:

• “The biggest mindset shift? Stop trying to ‘prompt engineer’ your way out of thinking.” – Gil
• “I’ve paid for prompt packs and never used them—because once I started just talking to ChatGPT like a teammate, I got much better results.” – Nihal
• “AI gives better answers when you start with better questions—and a conversation, not a command.” – Gil

Links mentioned:

• Anthropic’s Claude system prompt 
• Cursor
• Opus Clips 

Have AI questions? Connect with us on LinkedIn – we’d love to hear from you!

• Nihal Salah
• Gil Chan