Two years ago, I published Preparation for Re:Zero − Starting Life in North America. I wrote it before leaving, partly to sort out for myself why I wanted to go, which path I had chosen, and all the preparation behind the English tests, school applications, and visa. It was my way of breaking one enormous life decision into manageable pieces.

Looking back now, no single course, status change, or immigration pathway transformed me. The change came from the ordinary things that happened after I landed.

I started showing up at in-person events. I talked to people I would never have crossed paths with before. Slowly, I got comfortable around people from different communities. Because English made me anxious, I ended up building Joe Speaking. And after leaving my job, the empty time that once looked like waste turned out to be exactly what I needed to understand myself again.

So if the piece I wrote two years ago was about why and how I left, this one is about who I became afterward.

I was lucky in some ways, and there were pitfalls in others. Canada gave me freedom, along with policy changes and pressure. If you are also thinking about changing your life, I hope this serves as a reference point, not a template.

On the whole, I am very glad I left. Two years ago I was reading The Pathless Path and feeling like that kind of life was still far away. I did not expect that two years later, I would be living in a way that is less conventional and much more my own.

Timeline

• 2021.05 First heard of IELTS
• 2022.03 Hit my target score on my first IELTS attempt
• 2023.10 Applied for a Canadian study permit
• 2024.08 Landed in Ottawa and started over
• 2025.01 Gave my first public AI talk
• 2025.03 Built my first app through vibe coding
• 2025.06 Visited the Bay Area for the first time
• 2025.07 Got hooked on Claude Code, and building became part of daily life
• 2026.01 Attended a Hacker House and registered my first company
• 2026.04 Officially launched the Joe Speaking beta
• 2026.05 Reached CLB 9 for the first time using Joe Speaking

1. School

I never expected to be back in school, graduating again, ten years later.

To be honest, I did not come to Canada because I missed campus life. Immigration was the main reason. At first, school was not some romantic return to the classroom. It was a way in: legal status, plus some breathing room while I adjusted to a new country.

Looking back, the coursework was not the most important part. What school really gave me was a local starting point. Classes, classmates, and everything happening around the program pulled me into local life.

I have wondered whether I would have changed this much if I had quit my job but stayed in Beijing. Probably not. The environment in Canada, student life, and the steady encouragement I found in local communities pushed me into a different rhythm.

Our program also gave us ways to meet people beyond the university, including free tickets to events and conferences like TiE Ottawa, SaaS North, Capital Angel Network, and Mindtrust. Those events taught me how the local startup and tech scene actually works.

Assignments still took real effort in the first semester. But by the second, after Deep Research and Sonnet 3.7 arrived, coursework started to feel like playing with a cheat code. I still remember the look on one professor's face when he saw the report we built with Cursor. After that, Claude Code and Codex folded the whole process into a single workflow. I even turned my method for guiding Codex to write a hundred-page project report into a Skill.

In the end, the program gave me far more than course content. It gave me people, a place to show up, and feedback from real conversations. Online courses struggle to offer that. The faster AI gets, the more these slow, human parts matter. Being good at exams alone is no longer enough.

That was also when I started to understand the value of co-op, or paid work terms.

When I was applying to schools, I kept seeing people on Xiaohongshu, a Chinese social platform many students use to research studying abroad, say that programs with co-op make job hunting easier down the road. Once I arrived, I understood why. Co-op is not just a label on a program page. It gets students into real workplaces earlier.

Co-op means students spend set terms in paid placements. The government also provides incentives for eligible roles, so co-op positions are often more attractive to employers than regular internships.

Our program does not have co-op. And even with it, a job is not automatic. You still have to run the process yourself, and if you cannot land a suitable role, you may never start a work term at all.

I bring up co-op because it crystallized something for me: a student who has already done several internships and worked in real settings is in a far stronger position than someone who only learned from class.

That is also why I find some students here genuinely impressive, especially the Waterloo crowd you see on X heading to the Bay Area for internships. They have not just taken classes. They already know what it means to work on projects, communicate with a team, and deliver something real.

This will matter even more in the AI era. Knowledge is easier to reach than ever. What separates people is whether they have worked in real environments, collaborated with others, and shipped something that had to work.

I am also grateful I lived on campus for my first two terms and met friends from all over the world.

I miss those weekends of board games and hotpot.

I was lucky to become a student again, and to graduate, right as AI and work were both changing fast.

In some ways, every day felt like a break from the system. I had unstructured time to explore what I wanted to do, or to just sit quietly, let my mind go blank, lie down, and do absolutely nothing, instead of running along a track someone else had laid.

At the same time, I could feel the standard career path cracking apart. Credentials and experience still matter, but they no longer protect you the way they used to. You have to decide what to do with yourself.

My work experience and savings gave me some security. But what helped most was that Canada nudged me to think differently, and the people here encouraged me.

2. Professor Tony

One person shaped this period more than I could have expected: Professor Tony.

He founded our program. When I first arrived in Canada, I carried a fairly simple idea of work-life balance. Tony and Dan, our program administrator, quickly complicated it. I would see them answering emails late on a Friday night or very early on a Sunday. If I emailed them, the reply might come back instantly.

I did not take Tony's class at first. But after hearing enough of his talks, I could tell he was not reading from slides. His language was plain, but you could feel decades of building and teaching behind it. The image below shows his final 14 life lessons. They are very Tony: practical and direct.

Tony was warm, generous, and deeply committed to his students. The best teachers do more than transfer knowledge. They create opportunities, and they help you see potential you cannot yet see in yourself.

He was also always learning. Once he realized I knew AI tools well, he would occasionally email me to ask how AI might improve a specific workflow. Sometimes I shared new tools and methods with him too. That exchange felt special: he was the professor, but he never stopped acting like a student.

My first job in Canada felt like a lucky break. I never applied for it; the opportunity found me. A big reason was that Tony saw what I could do and was willing to take a chance on me.

In the summer of 2025, he encouraged us to take his course. Later, I took a Directed Study, working with him one-on-one on a project. I had no idea it would become the last project he ever supervised.

I probably should have noticed the signs earlier. When he stopped replying to emails as quickly as before, I assumed he was simply too busy, and I felt too embarrassed to bother him. In November, when we discussed project progress, he changed the time at the last minute. Thinking about it now, even near the end, he was still pouring his energy into the program he had built.

In February 2026, I sent him an update on Joe Speaking. What came back was an email telling me he was nearing the end of his life. I was completely stunned. I did not know how to respond.

The only answer I could think of was to keep making the product better, and to keep building.

In June 2026, we attended Tony's public memorial service at a church.

Watching hundreds of people gather, and listening to his family, friends, students, and community partners share their stories, made me realize how lucky we were to be among his final students.

As one alumnus from more than a decade ago put it to me: Tony changed many people's lives.

What moved me most was this: if a man spends the final years of his life, even the final stretch, still pouring himself into the thing he founded, still helping students, still pushing the program forward, then that has gone far beyond work in any ordinary sense.

I hope I can live like that too. Not just completing tasks, but truly investing myself in something I believe in.

I also wrote separately about my story with Tony.

3. Sharing

Before coming to Ottawa, I had no idea that the AI workflows I took for granted looked almost like magic to most people. Not because I was special, obviously. The information gap was just real, and absurdly large.

The moment that made this clear was Perplexity's campus event in 2024. I assumed 500 registrations would be easy. After round after round of emails, we were stuck at just over 300. In the end, I had to walk into the library and pitch people face to face to close the gap.

That was when I realized that even on a Canadian university campus, most people's understanding of AI stopped at ChatGPT. Things I saw daily on Twitter/X, used myself, and assumed were common knowledge were still unfamiliar to almost everyone.

That experience changed how I saw my own knowledge. I was simply describing how I used AI in daily life, but the audience's reaction told me those habits were nowhere near as widespread as I had assumed.

With Professor Tony’s encouragement, I started giving more AI-related talks. At first, he asked me to give him two numbers. I was confused and asked what numbers. He said dates, so he could book a room for my Perplexity workshop.

Those talks taught me something: if you spend enough time on Twitter/X, follow AI closely, and actually use these tools every day, then even if you were not among the earliest adopters, you are already far ahead of most people.

Because I started using this generation of AI tools fairly early and fairly seriously, I began to feel a strong pull toward building.

Even now, truly deep users of tools like Codex and Claude Code are a tiny minority. As of June 2026, even with OpenAI pushing Codex hard, it had only just crossed 5 million weekly active users. Plenty of people know AI exists. Very few have made it part of how they actually work.

Which means the information gap is enormous. So are the opportunities.

So stop telling yourself you have missed the "last train" of AI. If you are actually using these tools, you are much closer to the frontier than you think.

I have also come around on the "wrapper apps" I used to dismiss or ignore. They make powerful capabilities accessible to more people.

Later, I gave many more talks. Because they drew on firsthand experience from things I had actually built, they landed well. In a way, simply showing how I used these tools in my daily work was enough. And because the examples came from real building, I felt confident even though my spoken English was far from native. Locally, at least, I really was one of the people using this generation of AI tools deeply and aggressively.

But without Professor Tony's early encouragement and the opportunities he opened up, I would not have discovered any of this so quickly.

It is a textbook example of how confidence grows: someone believed in me first, then I discovered my own potential, and only then did the confidence follow. An environment full of encouragement and help is an incredibly powerful thing.

4. Networking

I first became aware of networking in 2023, when I saw @bearliu say on X that if you want a job overseas, LinkedIn is the most useful social platform there is. I gradually started using LinkedIn and cross-posting some of my longer X threads there.

After arriving in Canada, I began to see, concretely, how networking actually happens.

In August 2024, before school had even started, I went to a TiE Ottawa monthly event through the program. I was quiet at first and had no idea what to say. But the more I talked, the more I found there was to talk about: my past work, how I used AI, my impressions of Ottawa, why I came to Canada. Even as an introvert, if the other person took the initiative, I could ease into the conversation.

LinkedIn plays a bigger role here than I had imagined. After some small talk in person, people connect on LinkedIn and then keep loose contact through posts, comments, and likes.

But networking is not simply "adding someone." It is letting both sides keep seeing what the other is doing. Meeting in person first, then staying connected online, feels far more natural than forcing a conversation with someone you have never met.

When you first arrive somewhere, it takes courage to talk to strangers. I am no exception. These days I mostly follow my own energy: when I can talk, I talk more. Do not be too afraid of awkwardness. Just walk up and say something. The key is to take the initiative, because the other person may be even more introverted than you.

And if I am not in a good state, the priority is simple: get some food first, look around, and if it still is not working, leaving is fine too.

Yes, networking events usually have snacks, so there is no shame in solving dinner while you are there. A friend in San Francisco told me that if you go to enough events, you can basically stop paying for meals.

Some friends have told me that networking is an inefficient way to find a job. True, and it is also one of the most effective. If you can have one good face-to-face conversation with a future manager, your odds beat mass-applying online by a mile. Through networking, a friend of a friend, or a referral, you may find exactly the person you need.

In an era where online AI keeps getting stronger, offline has become the new online.

Even a homebody like me now goes out several times a week. When I travel somewhere new, I search for local events instead of just checking off tourist spots.

Over time, I moved from simply attending events to belonging to more stable communities.

I never felt this as strongly in Beijing. It is not that Beijing lacks communities. But after coming to Canada, I felt much more clearly how people with different interests and backgrounds naturally gather around a shared thing.

Following my interests, I found my way into different communities, and they helped me settle in faster. Technology Innovation Management itself is a community. I also joined communities such as AGI Ventures Canada, Build Canada, where I met many interesting people. Even Pokémon Go has its own communities. People simply play together, without many rules at all.

So my suggestion is simple: when you first arrive somewhere, go to lots of local events. Try different kinds first. Only then will you learn what you actually enjoy. After that, it stops being about attending as many events as possible and becomes about finding the people you click with.

Shopify Builder Sundays is currently my favourite regular event. A friend once asked why I still go when I already code every day. I told him the main reason is to talk with different people. It gives me a regular place to show up and meet others.

Still, I want to say this plainly: networking is a means, not the goal.

The most effective way is not going to events every day or mechanically posting on social media. It is continuously making things of real value. When you have your own work and your own projects in motion, it becomes much easier for people to understand who you are and to remember you.

In the end, a network is not something you force into existence. It grows on its own, after you keep showing up and keep creating in the real world.

5. English

English has been one of the main threads of these two years, and the part I worried about most.

What I never expected was that it would become the starting point for Joe Speaking.

At first, Canadian Language Benchmarks (CLB) level 9 felt almost out of reach. In IELTS terms, that meant 8777: an 8 in listening and 7s in the other three bands, with every sub-score hitting the mark. I knew my level too well back then. I did not really believe I could get there.

I wrote about this before in Toward That Farther Purity: My Multi-Dimensional Experience in New Zealand. It was only after two weeks of traveling New Zealand, talking with dozens of native speakers, that I realized my spoken English was actually usable. Only then did I apply for a Canadian study permit.

Speaking had always been a source of anxiety for me, yet in practice it rarely caused serious problems. The most direct lesson was simple: just say it. Language is for communication, not perfection, especially in daily life.

Early on, the main reason I went to so many events was to practice speaking. The more I talked with different people, the more my confidence grew.

One conversation stays with me. In October 2024, I talked with Owen at school for an entire evening. I shared a great deal about my past, and it was the first time I realized I could speak in English for that long, and go that deep.

But honestly, describing everyday life clearly is still hard. I owe thanks to Mark here. He encouraged me to describe my daily life in my own words, then gave me feedback.

I did not stick with the habit in early 2025. By summer, after realizing my score had barely moved despite a year in Canada, and watching immigration cutoffs keep climbing, CLB 9, and even CLB 10, suddenly became urgent.

That urgency became one of the reasons I built Joe Speaking.

I did not want more templates, memorized answers, or fake conversation drills. What I needed was speaking practice that felt natural, closer to real communication. That is why I later spent months building the real-time IELTS speaking simulation.

On the night of May 16, 2026, when the text message arrived and I saw I had reached CLB 9 on my first attempt, I was genuinely thrilled.

It was not just a score. It felt like a problem I had carried since arriving in Canada had finally loosened. More importantly, I did it with Joe Speaking, the product I built myself.

This is not the end. I am still far from expressing myself as freely as I want to, and there is still a gap between me and native speakers, or Chinese people who came here young. But that cannot be rushed. All I can do is keep practicing.

English used to be my pain point. In the end, it became my entry point into building products.

Joe Speaking was not an opportunity that simply appeared. It was an opportunity I created by working through a long-standing struggle. I needed it first, and only then could I believe that others might need it too.

6. Immigration

Many of the decisions I made along the way were tied to immigration.

In short, I had planned around Ontario's Master's Graduate stream under the Ontario Immigrant Nominee Program (OINP): finish an eligible master's, apply for a provincial nomination, and follow a relatively predictable path to permanent residence (PR). The problem was that right as we were about to graduate, the stream was cancelled, with no grandfathering.

Choosing Canada, choosing Ottawa, choosing TIM, setting my English target at CLB 9, even building Joe Speaking later on: immigration was running in the background of all of it.

Technology Innovation Management was one of the few master's programs that matched my background, sat outside the Greater Toronto Area, counted as STEM, and met the requirements of Ontario's Master's Graduate stream. From a pathway perspective, it was almost tailor-made for my situation.

If everyone had known in 2024 that the stream would be cancelled in 2026, far fewer people would have come here, and even I might not have chosen Ottawa.

When the reform proposal came out at the end of 2025, I stopped everything I was doing that same day. I wrote feedback, built a website, recorded a video, sent emails, and did basically everything I could think of.

I was angry, but I did not want to stop at anger. What truly upset me was not just that the pathway changed. It was that so many people had already made real life decisions based on the old rules. Schools, students, and related industries were all caught in it.

Unexpectedly, in early March 2026, I received a reply from Ontario's immigration minister. The response itself was fairly neutral, but invitations did go out in March and April. Unfortunately, after we received our graduation letters in May, no more came, and the Master's Graduate and PhD Graduate streams effectively ended there.

So even though I need to look for work next, I still want to publish my feedback on this reform first. Not as a personal complaint, but because these real voices should not stay locked inside private conversations.

Of course, this is not the only path. My current fallback is to gain a year of Canadian work experience. Overall, I do not think permanent residence will be a major problem, since I still have more than three years. So I am not panicking.

It sounds like a joke, but if even someone like me, with some savings and a willingness to try things, can feel the path narrowing, then people with fewer buffers are feeling it much harder.

I remember a dinner with a newly arrived friend in January. He said something very direct: my real goal was never immigration. Judging from what I am actually doing, immigration is more of a byproduct.

He was right. Even if my permanent residence were approved today, tomorrow I would still need to live, build products, practice English, meet friends, and solve the next problem.

As an immigrant, I sometimes wonder: why do we have to fight so hard for things locals are simply born with? But maybe that is also what makes us more competitive. Call it part of the immigrant experience. Some things you only understand after living them.

So immigration matters, but it is an intermediate goal, not the destination.

7. Not Working

A big reason I can keep some distance from all of this is that I have not been working a traditional job.

But that sentence only makes sense with its conditions attached: I am single, have no mortgage, have some savings, and carry little day-to-day family pressure.

Three years ago, when my former colleague Mr. Chen talked to me about these things, I did not pay much attention. Only once it became my own life did I understand this kind of calm: when you have a vast amount of unstructured time that belongs entirely to you, you feel free. Every day becomes about doing what you want to do, or, more precisely, about getting yourself to do what you truly want to do.

Quite a few friends have asked: if you are not working, what about income?

My answer two years ago was roughly the same as it is now: even with no income for a while, I could live for years without doing much.

Partly, my past savings gave me a buffer. Partly, the actual cost of living in Canada turned out to be less frightening than I had imagined.

More importantly, by 2026, AI had already started bringing in some income. Not much, but the basic shape was there.

Since 2024, I have increasingly felt that, for me at least, money and traditional work are sliding down the priority list. They still matter. They are just no longer at the center of every decision.

It sounds like a joke, but if the only goal were money, I should just go all in on SpaceX, Anthropic, and OpenAI. Come back to this sentence in three years.

I was also lucky that after arriving, some income opportunities found me through the program. After Claude Code and Codex appeared, I could sometimes finish months of work in a few scattered days, then keep revising whenever feedback came in.

You could say that since coming to Canada, compared with my Beijing days, I have done the easiest work of my life at the highest hourly rate.

Not working taught me something else too: people need work of their own.

However capable you are inside a system, much of it evaporates the moment you leave. Titles, permissions, processes, and institutional endorsement are not fully yours.

But if you have your own work, your own products, and your own voice, you slowly build a kind of confidence that does not depend on any organization.

8. Work

When I first came to Canada, I still planned to find a job. That was part of why I went to so many events.

But through 2025, I gradually realized that if I was only looking in Ottawa, there were not many companies I especially wanted to join. And with Claude Code and Codex arriving, I began seriously entertaining a different thought: maybe what I truly wanted was not to find a job, but to never return to traditional work at all. This was the "not working" desire I talked over with a classmate at Mooney's Bay in the summer of 2025.

Honestly, I do not really want to go back to traditional work.

On one hand, AI is improving so fast that things which once required long stretches of team collaboration can now be turned into working versions by a single person, much faster.

On the other, if we keep walking the old path, the risk is not just that a specific role gets replaced. The entire way organizations work may be redefined.

For me at least, the traditional path of finding a good job and climbing the ladder has lost its pull. Even the idea of a "good job" is getting harder to define. According to OpenAI’s latest blog post, even AI researchers are expected to have a large share of their work replaced.

I have been quite lucky along the way. My first two jobs in Beijing were both things I genuinely wanted to do, so I could bring enthusiasm to work instead of dreading it. In November 2024, I wrote a dedicated piece about my view of work. Reading it now, much of it has come true. I am being guided by my own past thinking.

It is not that I do not want to work. I just do not want to return to a system that stuffs people into processes and defines them by job titles.

Over the past few months, I often worked on Joe Speaking from morning to night. But because I was building something I believed in, it never drained me the same way.

I have also felt the limits of the so-called one-person company. If you want to build seriously, you still need a small team. The leverage a team brings is far greater than what I can create alone. And with Ontario's Master's Graduate stream cancelled, I still need to find a startup team where everyone believes in the value and competitiveness of what we are building, the way I believe in Joe Speaking.

As for job hunting itself, I increasingly believe the new way is not to lean on resumes and past credentials, but to let people see what you have already built. Honestly, it probably should have been this way all along. It is just becoming more common now.

Your products, articles, talks, open-source projects, and real user feedback explain who you are better than a few bullet points on a resume ever could. Ideally, an interview should not feel like an interrogation. It should be a conversation between two sides, grounded in real work. If you are intentional enough, you can even skip the interview entirely, just as I did when I joined Smartisan more than ten years ago.

Wherever you are, I think it matters to build yourself patiently over the long term, instead of only starting when you need a job.

The ideal state is the kind of capability Naval describes: able to survive anywhere, and eventually unemployable.

Over the past year especially, my skills have spilled well past the product manager role. I can try many things now, and different roles are starting to blend together. It is getting harder to treat people as interchangeable parts. Traditional function-based roles are blurring.

Of course, going from not knowing someone to truly knowing them still takes time. Relationships need time to settle. That is also the real value of networking. The best way is to work on projects together, where both sides come to understand each other naturally.

I have told classmates this too: if you use tools like Codex and Claude Code deeply enough, and seriously build something every week, your odds of landing a decent job go way up. You are no longer someone who claims you can do things. You are someone who has already done them.

So for me, job hunting has become just another building problem: not fitting myself into a position, but finding people I can truly build with.

9. Building

If there is one concrete change from these two years, it is this: I finally built my own product.

For people with ideas, this is the best time.

In software, at least, you no longer necessarily need funding at the start, and you do not necessarily need a full team. If you have a clear enough need and tools like Codex and Claude Code, you can start turning what is in your head into something real.

When I built Joe Speaking, I also wanted to prove one thing:

I had no traditional coding background. As recently as last year, I did not know what a Pull Request was. But starting from my own pain point, drawing on my past experience, and leaning on AI tools, I could still independently build a product that genuinely works, one that might even have a shot at being very good in its space. It was as if I alone, in a few months, did what used to take a whole team half a year or more. What is even more outrageous is that this was only the state of things in the first half of 2026, during the Opus 4.5/4.6 era.

Joe Speaking is not an ordinary project to me. It is the first time I truly turned my own pain and my own judgment into something other people can use.

What does this mean? It means the traditional work path I described earlier is loosening. If something frustrates you, or you cannot find a tool that meets your needs, you can now build it yourself, for real. In 2026, the barrier for individuals to build software has dropped dramatically. The harder, more interesting frontier will gradually shift toward hardware and real needs in the physical world.

Many friends immediately ask: Is it profitable? How much did you invest? What is the return?

But money was never Joe Speaking’s first goal. In the beginning, I just wanted to build a product I genuinely wanted to use, one that could help people with a story like mine. That is why it runs on a credit model and even lets users bring their own API key and use it for free.

In a sense, I am already experimenting with a way of working that feels closer to the future: start from my own real needs, build with AI, face user feedback directly, and iterate continuously. I am genuinely excited to talk with users. Once a need is clear, I can ship the feature they asked for in a day or two. That kind of relationship used to be unimaginable.

Honestly, I also see Joe Speaking as leverage for the future.

It paved the way for whatever I build next. It forced me through the full loop: need, design, development, launch, user feedback, and continuous iteration. That taught me more than any pile of articles or tutorials. It also became my representative work.

Competitively, even if similar products try to copy it, the cost of the Real-time API means they will likely end up much more expensive than mine. In other words, my pricing is genuinely affordable, and I would want to use it myself.

And even if I eventually take a job, this product is my strongest proof, stronger than any line of past work experience. I am also testing, quietly, how far a product built on my own idealism can go.

For individuals, this may be a rare window to build.

I have said this many times, but this time it is not because I spotted some grand trend. It is because I actually built something myself.

I now pour more attention into Joe Speaking than I ever poured into World of Warcraft.

The difference is that this time, I am not trying to win inside rules someone else wrote. I am creating something that belongs to me.

This is why I always start from personal needs. Only when something passes my own test first can I commit to it for the long haul.

I started Joe Speaking because of my own speaking pain point, my immigration needs, and the market demand already validated by my earlier IELTS Speaking Simulator GPT. And building a product I can be proud of matters in itself.

There are things whose weight you only feel after doing them yourself. You also learn that believing in yourself is not a slogan. It is the precondition that lets you keep going every day.

Building this product also brought my shortcomings back into focus.

I am not good at loud promotion, and I am not good at shouting for attention. I always feel the product is not good enough yet, so I would rather improve features than market them. But I am realizing now that, having finally built a product I genuinely believe in, I should be bold enough to let more people see it. That, too, is part of expanding my comfort zone.

All in all, Joe Speaking lives up to its name. It is a product with a very personal character and a strong sense of idealism.

It is still far from a top-tier product. But it is something I built with my own hands, believe in deeply, and hope can help more people. One step at a time.

10. Entrepreneurship?

Am I an entrepreneur?

For now, I still do not quite see myself that way. That is why I prefer calling myself a Builder rather than an Entrepreneur.

I did not start building products because I wanted to start a company. I started because I wanted to solve my own real needs. After building for a while, I found myself at the edge of entrepreneurship.

Joe Speaking is not entrepreneurship in the traditional sense. It is closer to an extreme experiment I ran on myself using new technology.

But building it taught me more about entrepreneurship than anything before it. It stopped being the entrepreneurship of books, podcasts, and other people's stories. It became something where I personally had to face users, costs, product decisions, pricing, distribution, and responsibility.

Back in Beijing in 2023, I talked with friends about whether to start a company. My thinking then was simple: if I wanted to build something, I needed to get out first, and not confine myself to the mainland market. Looking back, that judgment was clearly right. Even the company itself should have been registered overseas.

By January 2026, I realized Joe Speaking was no longer just a personal tool, but a product that needed to be taken seriously. A friend then reminded me that releasing a product as an individual means unlimited liability, while a company offers limited liability.

My thanks also go to everyone at AGI Ventures Canada for their help during the January Hacker House. I registered JUST JOE TECHNOLOGIES INC., a name that very obviously sounds like a one-person company.

The registration was far smoother than I expected. After confirming the name was available, I finished the official registration online in about half an hour. Most of the remaining steps could be handled online too.

Afterward, I sat there slightly stunned. That was it?

I used to think a company was something complicated and distant. Now I realize much of that complexity was self-imposed. Besides, we have AI now. If you do not understand something, you can at least start by asking.

If you truly want to build something, you can feel that much of this environment is set up to help you.

Once you have a company and a product, you can apply for all kinds of startup programs and free credits, from Google for Startups to Sentry, PostHog, Cloudflare, Vercel, and more. That support made the pressure of building Joe Speaking much lighter.

And because I had actually built something of my own and registered a company, I found it easier to talk at events too. Out in the world, identity really is something you give yourself.

As I mentioned earlier, with Claude Code and Codex, you can build the product first and talk about fundraising later. If you do want to raise money, walking in with real market data beats walking in with only a slide deck.

From a fundraising perspective, Joe Speaking is not currently a sexy unicorn story. I think it has a real shot at becoming a strong personal product or a small-team business, but it is still too early to tell a multi-billion-dollar story.

But if I can truly build a fully personalized tutor, a system that uses each person's practice history to keep telling them what to practice next, how to improve, and how to reach their goals, then the space changes completely.

In short, before coming to Canada, I never imagined I would seriously step onto the entrepreneurial path, although a friend did tell me last year that I was well suited for it. Call it one of the unexpected surprises. Mentally, I have also become far more open to entrepreneurship.

As Paul Graham wrote in his latest essay, “HOW TO EARN A BILLION DOLLARS”, the most common path is still entrepreneurship. I used to read the books, listen to the talks, and study how other people did it. Now it is finally my turn to step onto the field and try it myself.

This is also a psychological shift for me: if I stay within traditional work alone, it is hard to imagine freedom in the fullest sense.

Sometimes I wonder: if I had known I would end up building products and moving toward entrepreneurship, should I have gone straight to Silicon Valley back then? But people can only decide with what they know at the time. The me of a few years ago did not have the confidence I have now.

So it comes back to the same idea: since I am already here, I should walk the Canadian path well first. If what I build is good enough, opportunities will follow me wherever I go. Immigration status matters, but it is still only a byproduct.

11. Personal Change and Long-Termism

The biggest change in me over these two years is that I no longer set limits on myself so easily, and I am more willing to break old patterns of thinking.

Job hunting is one example. You do not have to wait for someone to hand you a position. You can create the opportunity yourself, or even go straight to an employer and say: here is the role I built for you.

Nor do you need to care so much about other people's opinions and judgments. If you fail an interview or a competition, that is fine. It only means you did not fit their standards. Everyone's standards are different, and you cannot please everyone. Maybe this piece will attract its share of comments too. Whatever happens, at least I have said what I truly think.

In the end, the most important thing is still to trust yourself.

This sentence sounds like a self-help cliché, but it is very concrete. You have to trust your own judgment, trust your own pain points, and believe that you can actually build something. If Joe Speaking had not started from my own pain point, I could never have persisted this long.

Many things do not become clear before you start. They become clear as you move. Action brings feedback, and feedback reshapes judgment.

Self-expression works the same way. In mainland China, I always felt a little uncomfortable putting myself out there, and I did not really want to be seen by too many people at once. After coming here, I gradually realized that my experiences and my voice are resources. I should learn to use them.

Your experiences, judgment, work, and expression are all entry points for others to understand you. If you never show yourself, how would anyone know what makes you different?

If something truly bothers you, or something has affected you, say it. Complaining alone gets you nothing, and passive waiting gets you nothing. Here especially, you have to speak up if you want anything to move. You have to give feedback, express yourself, and push things forward.

That is also why I still have confidence in myself, and why I can keep moving without feeling frantic.

As an individual, I think my greatest advantage is sincerity. Next is insight into real needs. Last comes long-termism and persistence, or Don’t Die.

Come to think of it, I am actually kind of terrifying, hahaha.

Once I understood this, many short-term things started to matter less against a longer horizon, work and money included.

I have always moved forward on the belief that life is one long, continuous accumulation. Seen that way, many of my past behaviors that looked strange actually make sense. More importantly, this way of living is sustainable.

It is still better to be a person with your own character. The faster AI gets, the more you need to be yourself. The things that truly belong to you will only grow more important.

There is a line in The Design of Everyday Things that says, roughly, designers communicate with users through their products. I understand that better now. Users can feel how much care went into a thing.

So I want to keep making things with my own temperament. They may not be the fastest or the best marketed, but at least they are real.

12. Nature and Everyday Life

After all this about work, immigration, AI, and products, I still want to end with everyday life.

Because for me, these two years were not just about doing many things. They were also about learning how to live again.

Before, life revolved almost entirely around work. After coming here, I started noticing the air, the weather, cooking, walks, friends, and a good pot of hotpot again.

I really love the air here, the blue sky, the white clouds.

Even inside the city, beautiful scenery is right outside the door.

Walk a few blocks and you will see squirrels, geese, and other small animals. It sounds ordinary, but it genuinely changes your daily state of mind.

Summer days are long here, and it gets dark very late. Winter is the opposite. Sometimes the light starts fading after 3 p.m. But after a friend took me winter hiking last year, I discovered that winter is not just cold and darkness. Snowfields, forests, and frozen lakes carry their own kind of joy.

Walking, cycling, hiking, skating. I have become far more active since coming here. There is so much outdoor life within reach, and it would be a waste not to use it. I hope to keep doing more.

In Ottawa, at least, strangers are usually friendly to one another. Most people greet you with a smile. Cars yield to pedestrians, and compared with many places, daily life feels less chaotic.

The small details add up too. Public washrooms in national parks have toilet paper and somewhere to wash your hands. Many buildings have automatic doors. The overall pace is slower. None of it is dramatic, but it makes ordinary life easier.

Meeting friends from different countries also turned their customs from things I had only read about into things I understood.

For example, I only truly learned what halal-friendly means after organizing hotpot with Muslim friends. From then on, whenever we planned activities, those needs were naturally part of the planning.

A lot of understanding does not come from reading online. It comes from eating together and talking.

This also changed how I think about "going global."

Back in mainland China, going global sounded like a business phrase. Once you live here, international users, other cultures, and unfamiliar living habits stop being abstract. It becomes hard to assume you are building for only one environment. Friends from around the world also help you see what you would miss on your own.

There was a time when "going abroad" sounded glamorous to me. Having done it, I can report that it is still just life. Wherever you are, life still has to be lived. A place can influence you, but it will not automatically change your life. You still have to choose how to spend your days.

I have plenty of complaints too. Restaurant prices here run about five times my hometown and three times Beijing. Food delivery and package delivery cannot match mainland China on speed or price. Then again, even if a simple meal feels expensive by Chinese standards, how often do you really eat out in a month? Overall, it is manageable. Food safety and quality are also better in many cases.

You can buy most of what you need in Canada, and many big Chinese restaurant chains have opened here. Still, the food options are not as varied as in mainland China, even in Toronto or Vancouver.

These days I cook for myself most of the time.

It keeps costs manageable, and I get to eat what I actually want. Grocery shopping and cooking helped me feel like I had a life again, instead of just working all the time like before.

Another thing: I rarely order delivery here, or use instant delivery the way I used to with Chinese platforms like Meituan or Hema. Price is one reason. Another is that the apps and websites here are just not very easy to use. I often cannot find what I want, so I would rather go to the store, even if I have to take a taxi.

And yes, I do not drive. I mostly get rides from friends or take taxis. Taxis are far more expensive here, and I cannot hail them as casually as I did in Beijing, but it is manageable.

I hope Tesla eventually removes the parentheses from FSD (Supervised). Maybe then I will finally have my own way of getting around. I am looking forward to that day. Of course, the next problem would be whether I can pass the driving test.

On housing, I did not expect to end up in a brand-new apartment here. In Beijing, an apartment of this quality would cost far more. I also bought my own bed and mattress for the first time, which made me realize I should have treated myself better back in Beijing.

And of course, my most food-obsessed roommate!

You have to make your own fun.

I love hotpot, so I hosted plenty of hotpot gatherings here and kept introducing it to the friends around me. When I graduated, a friend posted a Story announcing that the Hotpot Master had graduated.

That might be my most successful cultural export to Canada.

My three favourite foods in Canada right now are snow crab, spot prawns, and Costco beef ribs.

Canada has the clean air, nature, friendliness, and daily ease I love. It is also expensive, slow, inconvenient, and thinner on food options. Every place has tradeoffs. For me, the question is whether I can build a life here that feels balanced.

13. If I had to choose again

Two years ago, I moved from Beijing to Ottawa. If I had to choose again, I would still leave, without hesitation.

What I did not expect was how quickly AI would move. When I read SITUATIONAL AWARENESS in 2024, I was already excited. Many things that felt distant then feel close now.

I sometimes wonder what would have happened if I had left ten years earlier. Maybe I would have taken a completely different path. Maybe I would have made more money and found bigger opportunities. But that would have been another timeline, and it would not necessarily have given me the ease and freedom I have now.

On the whole, I am happy with the choice.

Compared with what I expected when I left two years ago, life has gone much further than I imagined. The original immigration goal remains unresolved. But the process has already given me more than I expected.

A lot of the time, I would just sit around, let my mind go blank, lie down, or do nothing at all. It sounds wasteful, but those blank spaces are what slowly helped me figure out what I actually wanted to do.

The faster AI becomes, the more I feel the need to stay close to what I actually know and care about. Tools keep getting stronger, so judgment, taste, lived experience, real pain points, and the ability to keep acting become harder and harder to fake.

For now, I want to stay optimistic and keep investing in the things I actually care about. Move first. Let the rest come slowly.

What's next

However much I complain about immigration policy, I still have to keep moving.

To put it plainly, I also need to find work. But I am not looking for a traditional job where I just fill a slot. I want a small team where we respect each other and the work feels worth doing.

What I want most is to join a small, strong, AI-native team that cares about the product itself. A team that moves fast, works directly with real users, and is willing to rethink products, interfaces, and the way people work.

I have not fully figured out the exact shape yet, but I have a rough sense: more human-centred products, more natural interaction, and maybe hardware. The real-time voice work in Joe Speaking is already pointing me in that direction.

These two years made one thing clearer: I am better suited to building things with a point of view and real needs behind them. I do not want to manufacture the illusion of growth around problems people do not actually feel. And I do not want to return to systems that slowly consume people through process and hierarchy.

Maybe English, AI, immigration, community, Joe Speaking, hotpot, barbecue gatherings, and all these scattered experiences will connect someday.

So next, I will keep building and keep showing up in the real world, turning what I notice and what I struggle with into products.

I still have to keep the immigration process moving. More importantly, I have already stepped onto this Pathless Path, and I intend to keep walking it.

Not to chase traffic, and not to prove anything to anyone, but to keep building what I believe in.

So let's keep going.