What I Wanted
to Ask My
Grandfather

The students had, of course, already been selected from the best. Everyone was clever, and everyone was working hard. The strongest students could continue into Yucai's senior high school on full scholarships. Others would have to pay substantial tuition to remain. Roughly half would not be able to continue at the school. The goal and the rules were clear, but there was little room for error. Knowing the arrangements did not tell a twelve- or thirteen-year-old which outcome she would secure.

Yet I do not remember spending my energy wondering whether effort itself was worth the trouble. This was what I had chosen to do; I wanted to do it as well as I could. If one method failed, I would look for another. If I had not learned something, I would try to understand it. I liked seeing effort take effect, and other people's seriousness moved me too. In December 1996, I wrote in my diary: “There is joy in making progress, joy in striving; I am moved by my own success, and by the spirit of those who are stronger.” That pleasure was real.

I have always been grateful to have grown up in China after reform and opening began. As I grew, the world around me was prospering too. If you were willing to work, to learn, to take a little risk, there seemed always to be another opportunity. I took continued growth for granted, just as I took my extended family's wholeness for granted: wherever I went, however long I stayed away, my relatives would be there when I returned. Confidence in the future became part of me early. Later, things grew less clear than they had been at school, and the world around me no longer moved steadily upward.

I emailed William White to say I did not know what I should study. What I learned now might not even be useful later. His advice was practical. I was only a freshman; the particular course mattered less than the seriousness with which I approached it. Whatever I chose, I should study it honestly, thoroughly, and with a professional's care. Beyond the knowledge itself, I should attend to the people who had discovered or created it: their methods of inquiry, their intellectual habits and character, and the language and logic with which they built a body of knowledge.

I had left McKinsey because I wanted to explore what AI might make possible. Yet in the process of developing my own product, I had gradually trapped myself in another demanding production schedule. Every day I was thinking about what was missing, what needed fixing, and how to keep up with the latest advances. I rarely made time simply to enjoy what the technology could already do.

About two weeks before I discovered my grandfather's photo albums, I deliberately set my work aside. I decided to spend some time being a consumer.

The most advanced models and AI agents were already out there. Why not simply enjoy their capabilities for a while? I didn't need to defend my own product, or insist that every discovery become a commercial opportunity. I could begin with my own life, and think about the things I had long wanted to try but had never pursued because I lacked the time, skills, or means.

For once, I wouldn't worry about whether something could become a business. I wouldn't rush to build another product.

I just wanted to see what I could actually do today that had once been beyond my reach.

Meanwhile, I began catching my father for conversations whenever I could. He disliked talking about himself. Ask him to review his career, and decades contracted into a sequence of positions: dispatcher, chief dispatcher, department head, deputy director, director, retirement. He seemed to think there was little worth saying about particular projects. But mention his childhood, or his mother's father, and he became animated, acting things out for me. I made images of the scenes he described—the Liao River, the courtyard, the games—and sent them to the family chat on his side. Everyone began contributing. The house should look like this; the flower bed was there; the game was played that way. Someone found a photograph, someone drew a plan. An incorrect image could sometimes bring out more memories than an accurate one.

As the material accumulated, practical difficulties appeared. The old albums could only be photographed by phone. Several pictures shared a page, often crookedly photographed, and had to be cropped out one by one. Family recordings needed transcripts, with some way of remembering who had said what and under which circumstances. I began making small tools to simplify the repetitive work. The programming, system building, and use of models I had practiced for more than a year were finding an immediate use.

It was here that my attitude toward AI changed. I did not mind if it could do every task on a computer better than I could. The more capable, useful, and inexpensive the tools became, the happier I was. I had something specific to do: to get to know the people, familiar and unfamiliar, in a real extended family; to collect their stories, understand them, feel something of their lives, and bring their past together. I belonged to this family. I could ask questions, knew whom to approach, and had the desire and ability to undertake the work. Every advance in the technology let me attempt something I could not have done before.

William White's question about competition and cooperation returned in an unexpected form. Once it had concerned classmates and rival classes; now it concerned AI and me. While trying to build a complete product, I had worried about its capabilities surpassing mine. Working on family history, I wanted it to become stronger. I was still absorbed in the work, still intent on doing it better. Now all these capabilities could help me.

My aunt later wrote that in 2008, when he was very ill and could no longer recognize even the relatives beside him, my uncle asked him at the dinner table to define inorganic matter. He answered at once, leaving her astonished. I thought of my calls home from Yale, when his voice gradually disappeared from the telephone. During those years, the family was still discovering, little by little, what he had forgotten and what he continued to remember.

I first put the material into an essay, which brought my older aunt to tears again. Later, I wrote lyrics and used AI to set them to music, then sent the song to the family. Both she and my mother said the words were hard to hear; it would help to see them. Back in America, preparing a presentation for a senior-living community, I made a music video using real photographs, archival material, imagined reconstructions, and subtitles in Chinese and English. I sent it to the family chat. Even my younger aunt, who usually met everything with humor, said it had made her cry. My mother said, “I really do like this.”

Before, I had sat at a computer polishing a large product for imagined users. Now, if people could not hear the lyrics, I added subtitles. If the courtyard was wrong, I looked for photographs or asked someone to draw it. If the interview recordings became unmanageable, I made tools. Each problem came from the people before me and the work in hand. Once it was solved, I could see whether something worked and whom it had moved. Memories once held separately were becoming something the whole family could watch, discuss, and add to. In remembering the older generation, we were acquiring a new shared experience of our own.

I still needed to earn a living and was still testing how this work might become a sustainable undertaking. It had already given me and the people involved a clear sense of value, yet remained difficult to explain briefly to acquaintances who knew none of the background. Once, a company name and a title had been enough. Now there was much to explain before someone could understand what occupied my days. I actually knew more clearly what I was doing and could feel its value more directly, while finding it harder to introduce myself. This required patience. It also tested whether I truly believed in the path I had chosen.

Thinking again of the 1997 letter, I understand more clearly that what I miss is not simply a time when goals were clear and someone told me what to do. At Yucai, I could change my methods and see improvement; what I learned belonged to me. That sense of agency was formed, little by little, in that environment. Clear goals did not mean predetermined results, or the absence of uncertainty for the person pursuing them. I was already learning to put my energy into what I could actually do.

Later experiences kept introducing new variables. At Yale, there was no longer a common standard for what deserved studying or what counted as excellence. After years at work, the relationship between effort, promotion, and reward proved less stable than I had once believed. Now, defining the problem and finding a sense of value had become a problem in its own right. I still wanted to understand it, still wanted to put my efforts into concrete work I believed in and gain something no one could take away, rather than prop up somebody else's story and play a bit part in it.

In the summer of 2012, while at Wharton, I signed up for a leadership trek to climb Mount Kilimanjaro in Tanzania. Two scenes from that climb have stayed with me.