3rd Anniversary of ChatGPT Release

Has the physical world changed dramatically? Not really. But for me, this year’s progress has been much more significant than the previous two.

One obvious reason is coding. I remember a friend asking me last year whether I had tried Windsurf. My answer was, “Not yet, my class doesn’t require coding.” Looking back, that mindset feels ridiculous. From now on, it is not about classes. It is about personal interest.

As the screenshot of my GitHub commit history shows, I have officially become a developer this year, aided by powerful AI tools. I could not have imagined this when I wrote my annual review last year. As I told a friend this week: “It’s not about productivity. It’s about possibilities.”

This November was my most productive month so far. I am building my first real product with authorization, cloud storage, payment, a credit system, and an admin panel. I hope to ship it by the end of this year. If you are interested in, you can visit english.hubeiqiao.com for an early look. It is an early alpha version, but the basic features already work.

As a former PM, I find it extremely helpful to go through the entire process myself. It gives me a much clearer understanding beyond conversations. The goal is not only to build a product that makes money, but to gain experience.

I told a friend who is preparing for product roles in Toronto that the most effective approach might be to build personal projects instead of spending all their time preparing for behavioral questions. Right now, a single person can create real-world impact.

I also felt that development or engineering is not the bottleneck even before the AI era. The true limitation is myself. If I genuinely want to do something, I can make it happen by collaborating with peers and delegating tasks. Powerful AI is simply makes the process dramatically affordable. As a result, there is no excuse for not shipping something.

Lastly, the most responsible way to use AI is still to articulate your thoughts by documenting carefully. No matter what your tasks are, I spend more time documenting with AI. It is like doing a development document review back and forth. My feeling is that even if you find a small issue, it’s worth allowing AI to do the research. Otherwise, future development is likely to experience the same issue.

That is all I want to share in this piece. Unlike the previous two years, when I wrote long articles and detailed use cases (you can still visit ai.hubeiqiao.com to see them), here I simply present my current thoughts directly and honestly.

Next, I will keep building my first real product and use it to prepare for my English test. I hope I can reach CLB 9.

For next year, I am keeping my options open. I want to build more and eventually support myself and live financially independently. If I need to work to get my PR, I hope to work on something exciting with talented people. I am drafting an unconventional plan.

Anyway, I am optimistic about the future. Let’s see what’s next.

However, after living with my roommate Bagira, I am reminded that we are still far from super intelligence. That gap is the opportunity.