Chapter 1: The Moment I Realized AI Didn't Know Anything About My Real Life

For most of my career, I kept a notebook. Not a fancy one. Just whatever was sitting on the desk — lined pages, a pen that actually worked, and enough discipline to write down what mattered before I forgot it. In my years at Anker, I must have filled dozens of them. Meeting agendas. Action items. Decisions and the reasoning behind them. The kind of institutional memory that doesn't exist anywhere official but quietly holds a company together.

In China's corporate world, this was just how things worked. You showed up prepared. You took notes during the meeting and circulated meeting minutes afterward, usually on Lark or DingTalk. Lark especially: it's where the conversations happened, where files lived, where decisions got logged, where the org breathed. And then AI arrived, and I suddenly felt like all of that discipline and structure was both more important than ever and completely inadequate at the same time.

The Notebook and the Context Gap

I remember exactly when it shifted for me. It was sometime in late 2023. I'd been using ChatGPT for a few months at that point, cautiously at first, the way you approach anything that seems too good to be true. But I was getting serious about it. I was using it to draft communications, work through strategic questions, stress-test ideas before I brought them to the team. And it was genuinely impressive. There were moments where I'd look at what it produced and think: this is going to change everything.

But there was this one afternoon where I sat down to work through a product positioning problem we'd been chewing on for weeks. I knew the context cold. I'd been in three meetings about it that week alone. I knew what the team had said, what the data showed, what the customer feedback was pointing to, what the competitors were doing and where they were weak. It was all in my head, and in my notebook, and scattered across a dozen Lark threads.

I opened ChatGPT and started typing. And I spent the next forty-five minutes re-explaining everything I already knew to a machine that had no idea any of it existed. I'm not exaggerating. Forty-five minutes of context-setting before I could even get to the actual question I wanted to ask. By the time I got there, I'd almost forgotten what I was originally trying to figure out.

And afterward I sat back and thought: this is going to be the problem. Not AI's capability. Not the quality of the answers. The bottleneck was the gap between what was happening in my real life and what AI was able to know about it. Everything important, every conversation, every decision, every piece of accumulated context that actually shaped how a business runs, lived in the physical world. In meetings. On calls. In the ten-minute debrief after a customer visit. In the whiteboard session that produced the idea nobody had written down yet. AI knew none of it. And getting it to know required me to type it all in by hand. Every. Single. Time.

Forty-Five Minutes of Re-Explaining

I want to be careful here not to make this sound more dramatic than it was. I wasn't in a moment of crisis. The business was doing well. I had a good team, a product I believed in, and a career that had taken me further than I'd imagined when I started out. But I couldn't stop thinking about the gap. Because it wasn't just a productivity problem, though it was that too. It was a structural problem. The way businesses actually work is through conversations. That's where the real information lives. Not in the documents, not in the polished presentations, not in the quarterly reports.

In the unguarded moment when a customer says what they actually think. In the argument that happens before the decision. In the side conversation on the way out of the meeting that changes everything. I'd spent my whole career learning how to be in those moments. How to pay attention. How to remember what mattered. How to use what I'd heard to make better decisions the next time.

And now I was watching this extraordinary technology arrive that was supposed to make all of us smarter and more capable, and it was blind to the most important thing. It couldn't see the conversation. It could only see what I was willing to type into a box.

I started telling people around me about this. And what struck me was that everyone, literally everyone, knew exactly what I was talking about. The founder who spent an hour every Sunday evening typing notes from the week into ChatGPT so she could get useful output on Monday morning. The sales leader who'd stopped using AI for deal analysis because the briefing took longer than just thinking it through himself. The consultant who kept a running document of "context I've already explained" so he could paste it in before every session.

Where Business Really Happens

These were smart, capable people. They'd all found workarounds. And every single workaround was a version of the same thing: do the work of bridging the gap between the physical world and the digital one by hand. It was the most expensive thing we were all doing. And we'd somehow normalized it.

There's an old joke in the hardware world that software people don't understand time. In software, you can push an update at midnight and it's live for a million users by morning. In hardware, you make a decision in January and you're living with the consequences at Christmas, if you're lucky. The tooling, the supply chain, the manufacturing, the logistics. Everything takes longer than you want, and the feedback loop is measured in months, not hours.

I spent most of my career in that world. Anker was a hardware company at its core, a company that had figured out how to move faster than the industry expected, but still a company that understood the weight of physical objects. What it takes to make them. What it takes to make them right. What it takes to get them from a factory in Shenzhen to a shelf in America to someone's hands.

That world teaches you things that the software world doesn't always value. Patience. Specificity. The discipline to define exactly what you're building before you build it, because changing your mind later costs real money and real time. The respect for manufacturing partners who've been doing this longer than you have and know things you don't.

What Hardware Taught Me About Time

It also teaches you something about where the real work happens. And it doesn't happen in documents. It happens in a factory in the Pearl River Delta at seven in the morning, standing next to the line, watching a production run come together, or not come together, and having a conversation with a floor manager that determines the next six months of your business. It happens in a customer focus group where someone says something that sounds small but carries a weight you feel immediately. It happens in the taxi after the meeting, debriefing with your team, when the real opinions finally come out.

This was the world I knew. Conversations as the unit of work. Context accumulated over years. The ability to remember what had been said, and when, and why it mattered. That was a professional skill I'd spent decades developing. And I was watching AI make everything faster while remaining completely unable to access the thing that made the work real.

I want to tell you that the idea for Memoket came to me in a clear, cinematic moment. A lightbulb. An epiphany. It didn't. It came the way most real ideas come: slowly, then all at once. A long period of frustration with a problem that didn't have a clean name yet. Conversations with colleagues who were feeling the same thing. A notebook full of half-formed thoughts that kept circling back to the same place.

What is the thing that could close this gap? Not the app gap. That part was being worked on by a lot of very smart people. The tools for organizing your digital life were getting better fast. The AI assistants were getting smarter. But what about the physical world gap? What about all the conversation that was happening every day, in every office and coffee shop and car and boardroom, that AI had no way of seeing? That gap didn't have a product yet. And once I saw it clearly, I couldn't unsee it.

Closing the Physical World Gap

I've spent a lot of time since then thinking about why this problem wasn't already solved. Why hadn't someone built this? Part of the answer is technical. The hardware, the audio processing, the on-device storage required to do this well without sacrificing privacy, these are genuinely hard problems to solve in a package that people will actually wear all day.

But I think the bigger part of the answer is cultural. In the tech world, the physical world has always been the afterthought. The assumption was always that the real action was happening on the screen, and everything off the screen could be handled by having people type things into the screen. That assumption worked fine when AI was a convenience. It starts to break down when AI becomes infrastructure.

When the tools you use to run your business are only as good as the context they have, when the gap between what happened and what AI knows becomes the most expensive thing in your organization, the physical world stops being an afterthought. It becomes the problem. That's where we are now. And it's why I ended up here, building something I couldn't find anywhere else.

My notebook is still on the desk. I still write things down. Old habits don't disappear overnight. But now I also press a button on my wrist before a meeting starts. And when the conversation is over, I don't have to choose between being present in it and remembering what happened. Both are taken care of. That was the thing I wanted. It took a while to build it. But I think it was worth the wait.

Memoket Gem captures your conversations so you can stay present in them. See how