Statement of Purpose

This page has a Chinese version.

SOP

About Study

By the time I completed my undergraduate degree, a series of dramatic experiences had convinced me not to rush directly into a PhD. I decided instead to spend some time watching how this era was changing. On the one hand, I wanted to strengthen myself and become better qualified for the next stage of study and research. On the other hand, I wanted to stay flexible in the face of historical change, rather than locking myself too early into a path that might prove too narrow.

At first, I planned to go to UCSD, seek guidance from Professor Yizhuang You, and continue learning machine learning in physics under him. Then the Yau Mathematical Sciences Center at Tsinghua University gave me a rare opportunity to remain in Beijing and continue studying and doing research there. I have always believed that spending more time in Beijing, one of the country’s most important scientific centers, is itself meaningful, so I chose to stay.

Over the past year and more, I have improved very quickly. That progress has not only been an accumulation of knowledge; it has also expanded both my perspective and my mentality. On the one hand, I have continued to work seriously in fundamental physics and have already achieved some initial results. On the other hand, because my co-advisor has long worked on quantum algorithms and quantum artificial intelligence, I have also begun to enter quantum information science in a more systematic way and to build an initial foundation there.

I have also had many deep conversations with Professor Luo Di at Tsinghua University. He is an expert at the intersection of quantum science and AI, and those discussions made me see more clearly that first-rate research today does not have to remain confined within the traditional academic system. Industry can also produce genuinely important work. I have always believed that big tech can do what big tech does best, academia can do what academia does best, and industry has its own distinctive value. In the age of AI especially, the sheer scale of resources available in industry has become impossible to ignore. Some of my friends at Tsinghua have done first-rate AI research while interning at companies and using the resources available there. That reality further changed the way I understand where research can happen.

About AI

My view of AI changed through three decisive moments.

The story begins in 2023. During my summer research experience in California, I was already using large models extensively for translation and text processing. But even by early 2024, my understanding of them was still limited. At that time, I believed they would revolutionize every profession built around the written word, such as drafting, translation, and polishing. I had not yet understood that they might transform almost every industry. Part of that slowness came from the fact that I was still, in many ways, a young student deeply immersed in academic life.

The decisive shift came mainly over the past eight months.

The first moment came in August 2025. I was attending a theoretical physics summer school at the Hangzhou Institute for Advanced Study when GPT-5 was released. That was the first time I felt strongly that large models were no longer just useful tools. They were beginning to look like a new kind of intellectual infrastructure.

The second came in November 2025, at the Turing Forum organized by Tsinghua’s Institute for Interdisciplinary Information Sciences. I spoke with several professors in quantum science from Peking University and Shanghai Jiao Tong University about the latest progress in AI. They told me that GPT Pro could, within tens of minutes, respond impressively to papers they had only just finished or had only just seen accepted, and could even suggest concrete directions for improvement. That was when I truly realized that large models no longer had only textbook-level reasoning ability. They were approaching research-level reasoning. From that point on, I began using GPT broadly throughout my workflow, which accelerated both my research process and my ability to work at a research level. At the same time, I began using Codex to explore the potential of vibe coding, from writing to code generation to website editing.

The third came in February 2026, this time driven by OpenClaw. To me, OpenClaw itself was frankly a very poor product. But what it represented mattered far more than the product itself. It signaled the beginning of an era: the maturation of foundation models, the emergence and systematization of skills, the appearance of harness architecture, and, above all, the genuine opening of the command line to AI. OpenClaw mattered because it truly connected the command-line world with AI agents. In that sense, it helped the AI ecosystem take one of its final steps toward the broader public.

For a long time, I believed that human teaching was irreplaceable. That is one reason I used to care so much about sharing my views on Zhihu. Now I think differently. I increasingly feel that good questions matter more than ready-made answers. I often find myself haunted by another thought: if someone locks himself in a study for five years, what kind of world will he find when he finally comes back out? Every scientific revolution and every technological revolution has a limited window. If that window is missed, the opportunity may never return.

My long conversations with Cai only deepened this realization. Even many people at the top of the education world now seem to agree that the existing model has already failed. We are entering an era in which learning itself carries almost no barrier. In such an era, what matters more is the ability to do things.

This month, and these past six months more broadly, have changed me deeply. That brings me back to my original theme. Since childhood, my dream has been to contribute to humanity. But the reality is that the world of fundamental science has become so difficult that, in many cases, even simple survival within it has become hard. At the same time, the level of attention that the public, investors, and governments now give to AI is unlike anything I had seen before.

I often ask myself whether what I once called formal science really remains the best meeting point between personal vocation and collective need. With the arrival of the AI era, I have for the first time seen a better sweet spot. AI commands broad attention and abundant resources. It can sustain real life, and it may also genuinely change society. At the same time, it still preserves a vast frontier that is worth exploring. In a certain sense, today’s AI is still living in the age of phlogiston theory. We still need to spend enormous time and effort discovering its true physics and waiting for its own Einstein moment.

Since turning toward AI, I have had new progress and new discoveries almost every day. I remain in a state of being overwhelmed, yet excited, by learning. That experience is very different from the way I previously studied, worked, and lived within the community of fundamental theory. For the first time, I have felt directly that an era is opening and that I am present within it.

Of course, if the right opportunity emerges in the future, I would still be willing to return to the broader scientific enterprise. But as the old saying goes: cultivate oneself, regulate the family, govern the state, bring peace to the world. If I truly want to move the world, I must first make myself wiser and stronger.

About Entrepreneurship

Because the current investment environment in China is not very good, my co-founder and I decided to move the project forward in the short term by entering various competitions. This choice had several practical advantages. It could help us win prize money instead of relying immediately on early fundraising; it could increase exposure; it could help us find reliable collaborators; and it could push product iteration forward.

The results came quickly. In only one month, really almost exactly one month, we had already received more than 100,000 RMB in prize money, found reliable hardware engineers and collaborators, gained substantial media exposure, and attracted some investment interest. Even I found the speed of that change surprising. In such a short time, I went from being a fairly typical student from the academic world to becoming someone with a modest name in the young tech entrepreneurship scene.

For a young student stepping out of the academic system, one of the most important steps in entrepreneurship is finding trustworthy collaborators. In that respect, I have been lucky. My collaborator is a full-time entrepreneur close to me in age. He has a strong humanities temperament, while I have a scientific and technical background. That complementarity has allowed us to go very far in a very short time. He can deliver an excellent pitch to investors and judges, and he can use his entrepreneurial experience to help build the team and sustain morale. I often tell him, “Rick, if you need anything on the technical side, I can build it.” Precisely because of that, our two-person team has gone from one success to another over the course of a single month.

Now, through the network we built through competitions, we have already found excellent hardware partners and distribution channels, and we plan to begin pilot hardware production this summer.

Compared with many other entrepreneurs, one of my strengths comes precisely from what I went through in the world of basic mathematics and theoretical science. That is a world that demands enormous effort while offering very little immediate reward. I spent half of my youth on a cold bench, and in doing so I trained myself to work hard and endure long stretches of effort. I am often surprised to find that my diligence is not necessarily greater than it used to be, yet in an entrepreneurial environment it is already seen by many people as unusually intense. That reservoir of energy, together with the habit of valuing time, allows me to keep learning new things and iterating myself in a rapidly changing era.

In this venture, what we want to build is an Apple for the age of AI. We believe that every technological generation has its own interaction gateway, and that AI-native interfaces will therefore be especially important.

Strictly speaking, however, my ambition does not end with consumer products themselves. If the opportunity arises, I still hope to participate in the deeper scientific enterprise of the AI era. Your company offers me precisely such a precious opportunity.

Why am I applying?

Your company offers a very special and genuinely rare possibility to me. The team is made up of well-educated people with physics backgrounds, which means that even if I am working in industry, I can still remain in close and high-quality contact with the academic world. My own scientific training also makes me feel naturally at home in such an environment.

More importantly, this may be the closest opportunity I have in the near term to truly change the world. That is why I hope to join your company.