I build in Asia-Pacific, then write down what it taught me.

There’s a message in my inbox from three weeks ago. Someone I like, asking for something small I could have done in five minutes. I read it, I meant to reply, I felt the warm little pull of “I know exactly who this person should meet,” and then life happened, the message slid down the screen, and now it sits there as a small monument to a friendship I let cool for no reason at all.

I have a lot of those monuments.

This is the embarrassing truth about me. The thing people know me for is connecting people. The thing I am worst at is keeping up with the people I connect. I’ll introduce two strangers who go on to build something real together, and in the same week forget to text back someone who has been nothing but generous to me. On paper I’m a terrible friend. In practice I care enormously. The gap between those two facts is something I’ve been quietly ashamed of for years.

For a long time I treated it as a character flaw. Try harder. Be more disciplined. Set reminders. None of it stuck, and I want to be honest about why, because the reason turns out to be the whole point.

It isn’t discipline. It’s capacity.

Everything a great connector does happens in one place: their head. You hold who this person is, what they’re building, what they need this month, who they’d love to meet, and which of the other four hundred people you know is the missing piece. That is an extraordinary thing for a brain to do. It’s also a thing brains are bad at past a certain size. The rough limit on the number of stable relationships a person can actively hold is around 150. My network crossed that years ago. Yours probably did too. So the superpower doesn’t fail because I’m lazy. It fails because it was never designed to run on hardware this small.

And the loss isn’t sentimental. It’s real value walking out the door. The most useful introductions in your life probably didn’t come from your closest friends. They came from someone slightly outside your circle who happened to know the right person at the right moment. That’s not a nice idea, it’s one of the most replicated findings in social science, and a massive study of real hiring on LinkedIn confirmed it: the acquaintances, the weak ties, are the ones who change your trajectory. Which means every person I meant to help and forgot wasn’t just a dropped thread. It was an introduction that never happened, a door that stayed shut, value that never got created because my memory ran out before my intentions did.

Here’s the part that took me longest to admit. I never wanted anything back for any of it. I connect people because it’s how I’m wired, not because I’m running some patient long game. But being wired for generosity doesn’t make you immune to being human. When life gets heavy, the follow-ups are the first thing to fall, and the people who fall through the cracks are exactly the ones who deserved better. Generosity that only works when life is easy isn’t much of a virtue. It’s a hobby.

So I stopped trying to fix myself and started trying to fix the system.

The tools that already exist are built for the wrong person. A CRM is a machine for salespeople to track deals, and the moment you turn warmth into data entry, you stop doing it, because it feels like work and warmth isn’t supposed to feel like work. I didn’t need a better spreadsheet for humans. I needed the part of me that never forgets a person to exist somewhere other than my own tired head.

That’s what I’ve been building. It’s called smallworld.chat, and the simplest way to describe it is a second brain for the connector I already am.

It starts with a single link. When someone opens it, they don’t fill out a form, they have an actual conversation with an AI that represents me, curious and warm and genuinely interested in them. In a few minutes it learns the three things a good connector always wants to know: what you’re building, what you can offer the people around you, and what you’re actually looking for. Plus the one person you’d love to meet. Then, instead of vanishing the way every chat does, you become a permanent, understood person in my network.

Quietly, in the background, it does the thing my brain used to do and can’t do at scale anymore. It looks across everyone I know and finds the real fits. This person needs exactly what that person offers. This person is precisely who that one asked to meet. It surfaces those to me, I approve, and it makes the introduction only when both sides say yes. It never invents a connection. It never shares anyone’s details without permission. And with the warmth layer we just shipped, it remembers the human things too, that you just moved to KL, that you’re training for a marathon, that you love good coffee, so a follow-up feels like a friend reaching out instead of a sales touch.

I want to be clear about where the value actually is, because it isn’t the chatbot. Anyone can build a chatbot. The valuable thing, the thing that compounds, is the map: a lasting, growing understanding of who everyone is and what they need, plus an accumulating record of which introductions actually worked. It never forgets a person. It works while I sleep. Every match it makes teaches it who fits with whom, so it gets sharper the longer it runs. A chatbot is a feature. A network that remembers everything and learns from every success is a different kind of thing entirely, and it’s the part nobody can copy by shipping the same feature next quarter.

I’m building it in public, and I’m building it APAC-first, because the instinct at the center of it, that the right introduction at the right time beats almost anything, is one I learned here, as an operator, the hard way. But underneath the thesis and the roadmap, the real reason is smaller and more honest than any of that.

I didn’t build this to network harder. I built it because I was tired of being the generous person who kept quietly letting people down. I didn’t want to care less to protect myself, and I didn’t want to keep failing the people I already care about. Those felt like the only two options I had, until I realized there was a third.

Your network is already bigger and more valuable than you can hold in your head. That was never a personal failing. It was a hardware limit. So I stopped asking myself to remember more, and built something that remembers for me.

Generosity shouldn’t have a memory limit. Now mine doesn’t.