Do You Remember Why You Started? Three Months Into My First Startup

Today is July 31. I have been building my company for exactly three months.

I do not have a framework, a tactic, or a list of mistakes to share today. I simply want to slow down and record what these three months have felt like, one sentence at a time.

I Think I Had Started to Forget Why I Left

Go back to early January 2026. I had just built my first 60,000-line analytics agent with Claude Code and Opus 4.5, and I was exhilarated.

It was four in the morning, but I was convinced AI had reached a moment that felt like ChatGPT in early 2023. Work that once required engineers suddenly felt accessible through a $100–$200 monthly subscription. I believed I could have an engineering team of my own.

This shift had to create opportunities. I just did not know what they were yet.

The idea of starting a company arrived with equal parts excitement and anxiety. It felt as if I should take a huge swing. There were so many things I could build: AI-generated short dramas, AI cameras, and whatever came next.

What finally made me commit was not one specific idea. It was a directional belief and a very simple understanding of entrepreneurship.

Building something would no longer be scarce. Getting close to the market and developing influence of your own would become far more valuable over the next several years.

A very small group of people can stay close to the frontier of core technology. My path had to be the one closer to customers.

And if entrepreneurship does not have to mean building a unicorn or taking a company public, perhaps a small, beautiful business is simply a way to keep cards in your hand—to earn more chances to experiment.

So I stepped in with confidence, prepared to make no money for six months.

I Turned Down Four Paid Opportunities

I often wanted to prove that I could close the loop. To me, that meant earning the first dollar outside the protection of a platform or employer.

But two ideas kept fighting in my head. Many paid projects were available immediately. They would also consume the time we wanted to invest in the direction we truly believed in: a creative assistant that helps people build a sustainable content practice.

I still do not know whether turning those projects down was the right decision.

To focus on a direction that I believe can compound over the medium and long term, I declined a co-founder opportunity, a full-time Forward Deployed Engineer engagement, and paid AI coaching requested by several followers.

Perhaps growth and focus are simply the repeated act of choosing—giving up something real in exchange for the outcome you want most.

The Certainty Behind Me Started to Look Tempting

Having a job may be the simplest thing in the world.

That is not a humblebrag. The feedback is immediate. A salary arrives every month. Bonuses and equity arrive every year. I had already moved beyond the junior roles dominated by execution.

I was good at being an employee. I complained about the coordination cost and slow pace of large companies, but my body kept getting better at operating inside them.

Compared with entrepreneurship, my old work now looks remarkably straightforward: write documents, run experiments with AI, and keep going.

In three months as a founder, I have already killed two projects: a general-purpose multi-agent system and an AI short-drama product. I became a little more pessimistic and a little less capable of seeing those failures through the lens of growth.

Content and growth are both new disciplines for me. Neither comes with a reliable playbook. They require sustained experiments and the patience to form judgment of your own.

I read countless posts on X and Xiaohongshu. Rationally, I know that many “three steps to achieve X” playbooks are the highly compressed aftermath of much messier work. Owning a copy of The Art of War does not make someone a master strategist.

I do not know how long this exploration will take. I have kept moving, but somewhere along the way I became disoriented.

I see small startup teams working incredibly hard, yet many appear to face a relatively low revenue ceiling. At this point, two paths began to occupy my mind at the same time: continue building a company, or return to an established platform and lead another innovation initiative.

Another discovery intensified the doubt: AI is not as powerful as I wanted it to be.

AI can amplify an existing advantage. But in areas where I lack professional judgment, relying on AI alone still produces work that I find mediocre myself.

My Past Success Became Psychological Baggage

Baidu. Z.ai. Meituan.

Product manager, senior product manager, product lead. Praise from younger graduate hires and interns. Standing on a stage, confidently explaining how AI would change people and organizations.

My rational mind knows I need to reset to zero. In entrepreneurship, I am a rookie.

But the old images still return, especially when the present offers no feedback. I start comparing who I am now with who I appeared to be then.

At work, I did not clearly see my weaknesses. Compared with many colleagues, I was more proactive, more self-directed, and more willing to test new approaches.

Three months into entrepreneurship, the gaps are harder to avoid. I do not always turn understanding into action. When the path is unclear, I struggle to identify the central constraint. I am emotional and can become confused. I have even started imposing limits on myself.

I Was Using the Wrong Incentive Model

I took these doubts to my co-founder.

I am fortunate to have him. He is five years younger than I am and unusually clear about why he chose entrepreneurship. He also has a thoughtful way of stepping outside the day-to-day noise and using a longer time horizon to guide present decisions.

For a moment, I caught myself thinking how nice it would be to be a few years younger.

Then I thought: WTF? Something must be wrong with my recent state. I am young too.

The most important result of that conversation was a pair of ideas: begin with the end in mind, and choose the right incentive model.

I had become too absorbed in solving the immediate problems of content and customer acquisition. When the work became difficult, I failed to step outside the current situation and view it from a higher, longer-term perspective. I was struggling with details that should not have defined the journey.

And then there was the incentive model.

How could I evaluate something as difficult as entrepreneurship with the quarterly KPI system I learned inside a large company?

WTF. The side effects of being employed for too long.

I Overestimated What a Month or Two Could Accomplish

I was still carrying a strong desire for rapid mastery.

Even with AI, the part of the startup loop it accelerates most is production: it makes building faster and cheaper.

In areas where we have not yet developed sound judgment—content-led acquisition or global expansion, for example—AI cannot contribute nearly as much.

We still have to enter the terrain ourselves, make mistakes, and build our own understanding. Only then can AI become serious leverage.

I was simply in too much of a hurry.

The six-month runway I initially gave myself was already short. When I reached month three, I felt an artificial urgency: half the time was gone, yet there was no obvious breakthrough to show for it.

Which of My Original Beliefs Still Hold?

After these conversations—and after slowing down enough to write this reflection word by word—I still believe the directional judgment that made me leave was right.

Build a customer network of my own. Develop influence of my own. Stay close to the market.

None of that has changed. If anything, I now understand those ideas more deeply.

I have also learned something new.

At the beginning, I recognized the importance of intelligence, learning speed, and action. Now I understand the importance of resilience.

I can also feel our team entering a different stage. We are beginning to encounter real disagreements. We need to communicate honestly and speak from the heart, not remain agreeable all the time.

What Kind of Life Do I Want at 35?

Have you seriously imagined what you want your life to look like three to five years from now?

In a little over three years, I will turn 35—a strangely sensitive age in China’s internet industry.

I had a long conversation with Claude about the life I want in the future. The point was to begin with the end in mind: to let an image of the life I hope to have several years from now guide the choices I make today.

We talked about entrepreneurship, family, and growth.

Then I asked it to turn our conclusions into images. One of them became the avatar some of you have seen. I am not quite as handsome as that version of me. I have never even had more than one visible abdominal muscle. But I want to move closer to that picture. I want to fight for the life it represents.

If you also feel lost—unable to find meaning at work, or still uncertain about the road you are taking as a founder—I recommend stopping for a moment and talking with a friend.

A person cannot live on ideals alone. But neither can we live without them.

The Best Time to Start Was Ten Years Ago. The Second Best Time Is Now.

I strongly believe that younger is better when it comes to entrepreneurship.

You have fewer fixed assumptions, greater sensitivity to new information and technology, and a lower cost of failure.

The past is already history. The best available moment to begin is now.

I have set more reasonable milestones for the next several months.

I may return to employment one day. But I am certain that, in the end, I will build again. I will create a business and a body of work that are truly my own.