✍️ Last Updated: December 26, 2024

Attending today @nuva.now Shang Zhe's AI lecture was eye-opening. Even though I already feel confident writing AI prompts, I still gained three valuable insights you probably never thought of using this way.

Three types of AI prompts, using ChatGPT 4o as an example

Many people keep asking how to use ChatGPT. Here I'll share three special ways to use it.

#1 A more comprehensive role background setting method

You already know how to set up character backgrounds, right?

Or like me, for the sake of speed, I might give instructions like this: "You are an all-around administrative assistant who can do presentations, communicate, and express yourself … Please help me with OOXXMost people give AI instructions this way, but if we need AI to give us more complete and accurate answers, you can ask AI to 'write the character background you want more completely' first.

For example, tell AI: "I want to find an all-around administrative professional with a Harvard degree, business management consulting experience, and skilled in OOOO. Please generate a detailed character background descriptionThen, after checking and optimizing the content generated by AI, you can directly feed it to another AI as a character background reference.

#2 Ask AI to explain it 'deeper' or 'more simply'

When you need AI to explain 'what is social media?', you have two approaches to try:

Explain it more simply

You can set the AI instruction as: "You are a second-grade elementary school girl, and you need to explain to your even more clueless little brother what social media is. Please explain it in the simplest way possible」。

Explain it more deeply

Have AI explain this in such a complicated way that even scholars and domain experts can barely understand: "You are a 20-time Nobel Prize winner. Help me explain what social media is in an extremely complicated way that leaves people confused, making it seem incredibly profound so others think you're truly amazing」。

In the same way, when explaining social media, the simple version might use analogy 'social media is like candy OOXX', while the deep version would use a bunch of theories you've never heard of. This approach is really super interesting!

Because we are all 'content creators', the knowledge we create and the perspectives we share must be explained in a way that the general public can understand. After all, everyone's life experience is different. Explaining profound knowledge in the simplest way is the best explanation (just like telling children big truths through storybooks and animations). When you use AI to help you use 'analogy' for explanation, it will be really convenient for our content creation and education work.

#3 Ask AI to optimize your original AI prompt instruction

After you work hard to create the perfect AI instruction, don't you often feel it's still not quite perfect in practice? Surprisingly, you can ask AI to further optimize the instruction you created, upgrading your well-crafted AI bot into something even better with answers that lean more toward what you want.

This technique is absolutely perfect for me as a Virgo! Every time I create a bot, I feel like something's missing somewhere. I never thought I could use this trick to level up my AI!

I also learned great tips for writing AI prompts in Notion today!

Yes, in today's class, Shang Zhe used Notion's /code – Python feature to create AI instructions, with 3 benefits:

#1 Titles automatically change color

Usually, those of us who use AI at an advanced level are accustomed to using Markdown syntax for instructions. With this feature, when you use # heading syntax, the text automatically turns blue (so convenient!), making even the longest AI instructions easy to see at a glance where the key instruction assignments are located, no worries about finding them.

#2 Box symbols automatically change color

When we write AI instructions, we often put information in boxes using symbols like [ ], { }, and "", Notion's code feature automatically color-codes these symbols too, which is very convenient for those of us writing lengthy AI instructions!

Sometimes instructions are too long and it's hard to find where the data selection starts. It usually takes time to locate, and it's easy to miss. But when the brackets auto-change color, we can see it at a glance.

The framework Shang-Zhe most often uses is “”” using these three symbols as opening and closing quotes to select data. The content within the quotes turns green as a whole, making it clearer to see the key points of the task to give to AI.

#3 You can directly see the entire AI instruction globally

When we use ChatGPT, there's always a major drawback: AI instructions are too long but the ChatGPT interface's chat box is too small. After using Notion Code, you don't have to be confined to ChatGPT's tiny chat window, and you don't need to scroll up and down looking for text. This increases work efficiency.

Conclusion

After seeing the event announcement on Shang-Zhe's IG, I immediately signed up for this "Systematic Learning of ChatGPT | Lab. November" event. The four-hour workshop was really short, but with these small gains I think it was worth it. His teaching style is especially lively and engaging (boyish), and I felt energized by him! That's why I wrote this article after such a long time. Thank you Shang-Zhe, and thank myself for writing this piece.

---

🌭 Welcome, Xiaochang 🌭
I haven't written an article in such a long time! I've been stalled for a while. Today's AI course was very rewarding, so I'm sharing this reflection across all channels. Do you like this kind of sharing? Feel free to leave a comment or encourage me! ～

Xiaochang is meticulous about typography; if you spot typos or formatting issues, please reach out to Xiaochang's IG Contact! All articles come from Heptabase Xiaochang's Second Brain🧠 Content is purely human-written, never AI-generated.

If you have any questions about the article content, or if there are any parts you'd like me to elaborate on, feel free to leave a comment below and let me know～
Some links in the article are "Affiliate Marketing" links. If you click through and make a purchase, Xiaochang receives a tiny little bit of commission (it won't affect the price).




by Knowledge Management Mentor Xiaochang🌭