Cursor! Google NotebookLM!—When AI Changed My Learning and Creating
2025-05-26
Times really have changed.
Recently, I splurged 20 bucks on a “toy”—Cursor—and it made me deeply realize that the accumulation of knowledge might not be as important as the accumulation of wisdom. Learn by doing—that’s the right way.
With Cursor, I Built a Small Game from Scratch in Half a Day
I used to be someone addicted to gaming, but now, AI has let me experience a bit of what it’s like to be a game developer.
The whole process was like having a conversation with AI. I casually said I wanted to make a “bouncing ball game,” and it generated a rough prototype. Then, I kept giving modification suggestions according to my ideas:
-
“Generate some squares.”
-
“Make them fall down row by row.”
-
“Give the blocks numeric values.”
-
“Set up a value check and clearing mechanism.”
-
“Adjust the ball’s bounce, speed, launch range…” Although new bugs kept appearing throughout the process, and it completely crashed several times, I eventually “molded” this shabby little game into existence, and even spent some time deploying it from local to online.
This experience was so cool. It made me think that maybe I could use Cursor to turn the algorithms from David Silver’s (AlphaGo author’s) ancient courses from years ago into visual mini-games. How awesome would that be!
With NotebookLM, the Logic of Reading Was Completely Changed
Another tool that blew me away is Google’s NotebookLM. I’d only heard about using it to read papers before, never actually tried it myself. Now that I have, I’m amazed!
This thing’s reading ability is really strong:
-
Instant Generation: Upload a document, and it immediately generates mind maps and summary dialogues.
-
Precise Location: All answers can be precisely traced back to the original text.
-
Honest and Reliable: When I asked it to add information to summaries that wasn’t in the text, it kept emphasizing “no reference found”—very rigorous.
-
Multi-dimensional Learning: You can listen to audio while jotting down what you don’t understand to ask questions anytime. To be honest, when I first started using it, I had some worries about new tools, fearing they might bring negative effects. But in the end, I found that how a tool performs ultimately depends on the person using it.
Does AI Have “Individual Experience”?
This amazement suddenly made me feel like the entire logic of reading has changed.
Of course, books like The Courage to Be Disliked that I read this month, along with various classics—these carefully written words still need to be read word by word in the original.
But it made me wonder: Will there come a day when AI can discuss works like Flowers for Algernon and Never Let Me Go with me? At first I thought impossible, but who knows?
On this point, I had an interesting discussion with my friend Will. He believes that while AI is powerful, it lacks “individual experience.”
AI cannot tell you that the girl you met at 18 was the flutter of “if life were only as it was at first meeting, why should the autumn wind grieve the painted fan”; the one you met at 28 was the devotion of “having crossed vast seas, no water compares; having seen Mount Wu’s clouds, all others pale”; and the one you met at 38 was the fated encounter of “searching for her a thousand times in the crowd, suddenly turning back, there she was, in the dim lights.”
I agree with his view. Current AI might only understand correlation, not causation. It can only serve as a spark to inspire thinking.
The various limitations of being human actually add more color to life, creating different sparks from collisions.
However, as other friends said, the speed of learning really can’t keep up with AI’s development. Before, no matter how many courses you watched, without hands-on use, the impact wasn’t that great. Now, I really feel like I need to get on board quickly.