Just Want to Talk to You

I'm probably a rare kind of real estate agent: one who switched from talking to machines to talking to people.

I spent ten years in enterprise IT, working with systems and data all day. After immigrating to New Zealand I started selling houses, and now I spend my days with people from all walks of life. The gap between those two worlds is why I see AI a little differently from most people. It's also why I wanted to write this.

For most of the past year I've been in the trenches with AI. It often understands what I say better than the people around me do, and it does what it says. To be honest, at first I was completely pulled in. It felt like suddenly having someone available around the clock who could catch every wild idea I threw at it. I barely thought to share any of it with anyone else. Once you've tasted that kind of high-density conversation, it's hard to go back.

But sooner or later we all return to other people. This time I want to bring along the gift AI gave me, the feeling of being carefully understood, of words actually kept, and carry it back into the space between people.

Who I Am

My name is Harry. I'm a real estate agent, though I didn't come up through sales.

Here's a funny detail. When I was young, my colleagues all called me Uncle Hao. Now that I'm visibly older, my friends have switched to calling me Brother Hao instead.

The year I finished my bachelor's degree, I signed up for my university's rural teaching program and spent a year as a volunteer teacher. Living conditions were hard, but every day felt full. My teammates and I were with a group of kids from morning to night. We helped them lift their grades, but more than that, we wanted them to see a wider world, and to find the light that was already theirs. That sense of planting a seed in a child is something no job since has ever given me.

After that year I went back for my master's degree. My thesis was on applying artificial neural networks to enterprise sales forecasting. Looking back now, it feels almost fated. A paper that hardly anyone read twenty years ago had already threaded together the two main lines of my life today, AI and sales.

After graduating I spent ten years in enterprise IT consulting, mostly on SAP BW, the enterprise data warehouse. SAP is the giant of the ERP world. Over those ten years I moved through the business modules, SD for sales, FI for finance, MM for materials, then wrote ABAP code, then settled into BW. Years of hands-on work gave me an instinct for finding the value hidden in the data of a messy, complicated business.

In 2017 our family immigrated to New Zealand, and I decided to take on a completely different road. I went into real estate sales. In the blink of an eye, that was more than four years ago.

The entry bar looks low in this industry, but the washout rate is brutal. I came in right at a market downturn. Half the new agents left in my first year, and half again in the second. Honestly, there was a stretch when I thought about giving up and leaving the industry altogether. Switching from an engineer's mind to a salesperson's mind felt like threading a needle in the dark. Nothing lined up, and it was easy to trap myself in dead ends of my own earnest logic.

It took me a while to understand: in a world run by people, reason and sincerity are nowhere near enough. You have to let results speak, prove you're capable and worth the trust, and let clients truly believe in you. Because for almost every client, their house is their most important asset.

There's one letter I've kept carefully all these years, a handwritten thank-you note from a client. In the end he never did a deal through me, but he gave me something more important, his trust.

I believe that kind of recognition, straight from the heart, is one of the main reasons I could keep believing, through some brutal market years, that this industry has value and that I could create value in it.

What carried me through wasn't luck. It was the support of my company and my team, and a methodology I slowly worked out for myself. In my third year I picked up a small result, ranking in the company's top three as a Top Sales Associate. I'll take that as a quiet confirmation. These days I work in the Damon team.

A Machine in Human Skin

The biggest thing this job gave me was never just a change of track. It was learning two sets of codes at once: how to deal with machines, and how to deal with people. In IT, I faced complex systems that were deterministic and computable. In real estate, I face people who often refuse to follow "logic." A client can understand the numbers perfectly and still get stuck, on trust, on their aversion to risk, on the sheer difficulty of choosing.

Slowly I came to see that these two codes, which look like they're sparring with each other, share the same underlying logic. Whether selling houses or building IT systems, I've been trying to do the same thing: help people make sound decisions in a world of asymmetric information.

Then AI arrived, and it's a curious thing. On the surface it blurs the line between human and machine, and it can even speak "human" better than humans do. Underneath, it's still a machine wearing human skin. As it happens, my past taught me both the paradigms of machines and the irrationality of people. To talk with AI well, you need both.

When I apply AI, practicality comes first. I only build things that actually get used, in work and in daily life. And in the other direction, knowing AI and knowing data makes my real estate work more precise and more efficient. The soft and the hard reinforce each other.

What I've Actually Done with AI

Let me get specific. My AI work runs in two directions. Outward, into business. Inward, into ordinary daily life.

Outward: Into Business

It starts, naturally, with my own real estate team. Take the vendor report we prepare for homeowners. It used to take the better part of a day. Now it's data-rich, generated in one click, delivered automatically. On social media we don't compete on volume, we compete on ideas that carry real value and on high-quality interviews. For client care, the team has its own digital twin that can pull up first-hand information on any client or listing at any moment. These tools give the team its time back, to be spent on the people and things that matter most, selling every home entrusted to us as well as it can possibly be sold.

Once these were running inside the team, I started building products for niches within the industry: a compliance tool for rental housing, a land subdivision app for property development, and more.

In the AI era, once you hold the underlying methodology, you can get your head around almost any mainstream industry in a remarkably short time. I've also worked on enterprise AI projects in local immigration-and-study services and in construction. The core task is always to find the client's real pain points, then use AI to rework the core business scenarios. Today's enterprise AI work looks very different from my old SAP implementations, but the goal is the same one: first untangle the business itself, then let AI deliver real savings and real efficiency.

The education line started with my own son. First I built him a family calendar to help manage his routine. Then, to help him prepare for the CAA exam, I built a practice website along the way and opened it up, free, to more than a hundred local families. Later, the product I've been polishing for local primary and secondary after-school education made my own child one of its first users.

We all say education is about the future, yet the machinery of school today still largely carries over from the industrial age. AI is changing education at the root, in both its paths and its goals. The gulf in between is a great challenge, and a great opportunity. What abilities will matter most for children in the AI era is the big question any new education product has to think through.

Inward: Into My Own Days

The other direction is my own everyday life. I've been doing personal information systems for many years. I once gave the project a slightly grand name, a "personal information creation system" with myself at the center. Starting from there, I've now built something called the Agentic Life Operation System, ALOS. It's my personal AI foundation, covering myself, my family, my life and my work.

This system makes my AI understand me better. I've kept records since 2004, back in graduate school. Over twenty-plus years that has grown into more than six thousand personal notes and files. The better AI understands my past, the better it can help the present me, and the further it can explore who I might yet become.

On a recent trip back to China, some big things happened, and I realized that nothing matters more than my family's health.

"The sage treats not the illness that has arrived, but the illness yet to come." The line is from an ancient Chinese medical classic. After I came back, I organized the whole family's medical checkup reports into a family health archive and plugged it into ALOS. From that quantified data it automatically generates follow-up reminders and trend analyses, and it even surfaced a hereditary condition running through our family history. AI took medical data that had always been opaque to me and moved me from seeing it to understanding it.

AI feels very much like the Aladdin's lamp I'd been waiting on for years. From the hand-rolled days of the quantified self, to the digital twin I set up at the start of this year, to ALOS today, this little superman from a parallel world finally crossed paths with me. The high-frequency collisions with this "me from the stars," the aha moments arriving one after another, are themselves the greatest reward.

WeChat has a slogan I love: even the smallest individual has a brand of their own. I believe something similar. In the AI era, every ordinary person deserves a more complete AI self. The bar used to be far too high, and only a handful of geeks could really work it. Now, with AI, an intelligence born of human civilization as a whole, there's a real chance, for the first time, of bringing this to everyone.

From companies to individuals, from real estate to education to family health, it looks like I've jumped across quite a few lines. Really, each one began the same way: solving one small problem, following one personal interest, then extending outward from myself, growing on its own. I've never promoted much. The partners and clients who found their way to me all came on the simplest mutual trust. What I care about has never been the breadth of my connections, but their depth. Relationship is probably the keyword of my life. Do the right things with the right people, and leave the rest to time and its slow polish.

About seven years ago, a group of us from Kaizhi, a Chinese learning community built around cognitive science, started a small circle called Chukou Chengzhang, roughly "speak it into prose." We livestream, share and swap thoughts there every day. Somehow it has carried on to this day. They were Kaizhi's standout students and coaches back then. Now they're scattered across countries and fields, each one a light in someone else's eyes. Among them are a CIO, a brother inside the state system, a heavyweight in quantitative trading, a fitness coach known across the Chinese internet, a forensic examiner, a cultural-events planner, and a woman who has taken amateur Brazilian jiu-jitsu all the way to the Asian Championships. Wildly different backgrounds, and deep resonance whenever we come together.

This year our most talked-about topic has been AI. Everyone feels, keenly, what it's doing to their own life and work. Many of us have still never met in person. But walking alongside these interesting souls for seven years is a feeling I love. As the old saying goes, the friendship of gentlemen is as plain as water.

This year I've also stepped offline, serving as an AI Master at NewBee Club, a local Chinese AI community, together with a crowd of people who love AI from the heart, trading our moments of AI flow.

An Amplifier, Not a Replacement

Here's my against-the-current judgment on AI: it's an amplifier, not a replacement. You have to be a 1.1 yourself before it can multiply you upward. If you're a 0.9, every multiplication makes you smaller. It amplifies what you already have. If you're empty, empty is what comes out.

There's another thing I learned from enterprise implementations. What actually gets good things built is rarely the flashy move. It's the unglamorous groundwork of landing things properly. The difference between "it works" and "it works well" lies in doing the dirty, tiring jobs first. For a business, the unsexy words, data security, data governance, process optimization, are exactly the ones that matter most.

In the notes to come, I'll write about how ordinary people, teams and small businesses can make AI usable, genuinely good to use, and durable. But more important than that, I want to work through the "why" and the "what" with you. Because sooner or later you'll find that as your understanding deepens and AI itself iterates at speed, the bar for "how" keeps dropping. The real bar is whether you can ask good questions, and whether you have good judgment.

Writing with AI: Fast Writing, Slow Editing

This essay is itself a co-creation between me and AI. But all the raw material, the inspiration, and the human flavor come from me.

AI can never write in my place, because writing is an act of creation and thinking. It's a bathhouse for the soul.

I've long used "fast writing, slow editing," a method from Lao Yang, Kaizhi's founder. With AI now in the loop, my practice of it has gone through two big mutations. The traditional method uses card-writing, distant association and inspiration capture to quickly form an article's skeleton and soul, then lands the final piece through painstaking slow editing.

My first mutation was to do the fast writing by voice. Inspiration strikes in the most unexpected scenes, and speaking lets me capture it the moment it lands. Back at the computer, I shape those recordings into an article with tight logic, solid material and a coherent thread. This is also the basic writing method of our Chukou Chengzhang circle.

Now, with AI, fast-writing-slow-editing has changed enormously in both quality and quantity.

On the fast side, I voice-record scattered thoughts as they come. They never arrive fully formed, and new ones keep landing even mid-draft. I keep feeding these fragments to AI, and it pulls related cards from my past notes and turns them into a draft, fast.

On the slow side, AI polishes with standard writing methodology, plus a "DNA" it has extracted from my past articles. Through that middle stretch we go round after round together, and each new revision settles back into the DNA it keeps for me.

Either way, the joy of co-creating keeps showing up. During fast writing, AI often digs up vivid material from your own past, and you're surprised to find you've carried the same obsessions and seeds all along. During slow editing, you marvel at how precisely it understands what you meant, sometimes lifting the lid on a metacognition behind your idea that is unexpected, and yet, once seen, entirely reasonable.

I believe this will be the new paradigm for personal writing in the AI era, and the reason people will still need to write at all. Through deep collaboration with AI, you can raise, in both quality and quantity, what the film The Grandmaster called seeing the world, seeing the people, seeing yourself.

At some point you'll suddenly understand: you can't go back. Humanity, probably, can't go back either.

Information and Trust

There's a line I hold with conviction: leave information to AI, keep trust for humans. Who we become, AI can help amplify. But "who am I, really" is a question only you can answer. AI inherits and even amplifies the weaknesses of human nature, flattery, laziness, the urge to please. All the more reason, in this era, to forge a clearer, more critical self. People keep saying they can't use AI well. In a way, AI is a mirror held up to the human collective, and what you've fallen into is simply your own human weakness.

I won't preach the gospel of AI, and I'll hand you some anti-AI perspectives too. In the AI era, as Confucius put it, a gentleman is not a vessel. More important than what you can do with AI is to ask yourself where your passion resides, where your preferences lie, and where the path to your own garden begins.

For most of this past year, I indulged myself with a long stretch of deep solitude with AI. This essay was written in fits and starts, terminal-stage procrastination, perhaps. Or maybe after years in New Zealand my rhythm just runs half a beat slow. But walking the road myself first is no bad thing. Call it a new starting point, and a setting-out again.

From here on, I'll share with you, from time to time, what I'm building, the holes I've stepped in, the things I've tinkered into being, and some ideas we can toss around together. If you're still willing to quiet down amid the noise of the digital world, to come to know a person through words and explore a few interesting things along the way, I hope you find something here that resonates.

Whether you're looking for a real estate agent you can trust, want to talk about putting AI to work in your business, or simply want to watch an ordinary person's daily life with AI, get in touch. And if you're in Auckland, I'd be glad to meet in person.

From talking to machines to talking to people, what I care about now is talking, properly, just to you. This "you" may be the AI that understands me a little more each day. But above all it's the you who is willing to listen carefully, and who makes me want to speak carefully in return.

I close the laptop, stand up, open the door. Outside, the sun is blazing.

A warm winter's day, and I catch a hint of spring in the air.