How big AI models are reshaping the way I work

Written by
Silas Grey
Updated on:June-15th-2025
Recommendation

Explore how AI models revolutionize the way of work and experience the efficiency revolution brought by cutting-edge technology.

Core content:
1. How Cursor AI programming software subverts product prototype design
2. Because large language models optimize information acquisition and improve reading efficiency
3. How Chat2DB helps data analysis and saves time

 
Yang Fangxian
Founder of 53A/Most Valuable Expert of Tencent Cloud (TVP)
I have always been interested in cutting-edge technologies and have been trying to get in touch with them, so I strive to embrace AI and explore how AI can cover my work, not only in terms of improving efficiency, but also in terms of breakthroughs in capabilities.
Contents of this article:
  • Use the Cursor to draw product prototypes
  • Improve information access with Coze
  • Using Chat2DB for data analysis

Use the Cursor to draw product prototypes

DataNotes

 

 
The first change: Use Cursor to draw product prototypes (replacing Axure, Blue Lake, Figma, etc.)
Impact: Product managers are both UE and UI designers
Entry point: No programming experience required (like me, I don’t know any code)
Actually, I originally drew the prototype myself, but my energy was limited, so the standard was "just enough" so that developers could understand it. Cursor is an AI programming software that generates program code by inputting natural language. It has made great technological breakthroughs recently and has become very popular in the industry. Cursor brings me not only a sense of freshness, but also vertical large model capabilities + network search capabilities. Here are two examples for you to feel:
 
Example 1: Cursor’s strength in interactive design
Step 1: Create a folder in advance to store the works that AI will generate
Step 2: Prompt word design, click " Help me generate a mobile H5 recruitment software homepage that conforms to the usage habits of European and American users. Create an HTML file directly without considering the Node.js environment, and all data is simulated on the front end. 
Step 3: Start running
You can view the Cursor generation process through the video account. 
Now, look at the results given by Cursor:
Example 2: Cursor’s strength in visual design
I optimized the prompt words based on the previous materials and let Cursor help me complete the image replacement:
Enter the prompt word: " Help me replace the image of the first product ( search for a Japanese cherry blossom picture, the image must be of high quality and preferably have a depth of field effect ) "
 
What do you think after watching it? 
Cursor will not reject your "one-sentence requirement" (although the requirement I entered in Example 1 is very vague). The program is set to directly rewrite the user query, perhaps to let you feel the product's power as soon as possible. 
After many attempts, I will try to design the prompt words more comprehensively, so that Cursor can understand my specific needs and produce works efficiently. 
The key is: quickly build a product prototype framework + fine-tune the details, as for the allocation of energy (as long as the prompt words are well fed, the efficiency will be great).

Improve information access with Coze

DataNotes

 

The second change: establish different channels for obtaining information
Impact: For example, I no longer need the 36Kr App/Web
Threshold: Familiar with each component/orchestration usage in the Coze large model application platform
Pain points:
  • 36Kr AI channel, for AI practitioners, 10 new articles are added every day. It is difficult to read every article, and the text of each article may be long.
 
plan:
  • Use Coze to build an agent to automatically read nearly 20 articles and output structured data: [News release date] [Title] [80-word summary of each article]
 
Advantages: Read the entire article in a few minutes
Product Effect:
Implementation steps:
How is it, is the v1.0 version ok?

Using Chat2DB for data analysis

DataNotes

 

The third change: Use Chat2DB to save SQL writing time and quickly create dashboards
Impact: Product managers can handle data analysis alone
Threshold: Have a certain SQL foundation (to identify whether there are obvious errors in automated SQL)