There’s a good chance you’re using AI all wrong. I’ve definitely been frustrated getting generic, unhelpful answers from my prompts. This industry pro just shared a powerful shift in thinking that makes your results incredibly specific and useful!
The core idea is simple but brilliant. Stop asking broad questions like, “How can I do [X]?” Instead, you should first feed the AI your own trusted sources—like PDFs, articles, or even YouTube videos—and then ask your question based on that knowledge. The prompt becomes, “After reading all of these sources, how do I do [X]?”
Here’s what stood out to me from the original post:
- 📌 Source First, Question Second. The biggest mistake is expecting an AI to have perfect context for your specific needs. By providing your own curated library of information, you’re essentially creating a custom, expert brain for the AI to use. The quality of your outputs skyrockets.
- 💡 A Free Google Tool for This. The author points to Google’s NotebookLM as the perfect tool for the job. It’s designed to let you upload your sources and then chat with them, making this whole process super smooth and easy.
- ✅ It Works with ChatGPT, Too. While NotebookLM is the main recommendation, the person who shared it also mentioned that a similar workflow is possible with ChatGPT. You just need to know how to set it up.
⚙️ How to Use NotebookLM
The post’s author laid out a simple 5-step guide to get started, and it’s awesome:
- Go to
notebooklm.google.com. - Create a free account with your Google login.
- Upload your trusted sources (PDFs, YouTube links, etc.).
- Click “Discover” if you want the AI to suggest more related sources.
- Start asking questions in the chat window, just like you would with a normal chatbot.
This completely changes how I think about prompting. It’s about giving the AI the right ingredients before asking it to cook.
For more details and a link to the blog post explaining the ChatGPT method, you have to check out the original post.