Chapter 2: Why We Left the Corporate World to Build Something No One Had Built Before

There's a device that sticks to the back of your iPhone with a magnet. It records meetings, transcribes them, summarizes them. It does exactly what it says it does.

I used one during my consulting period, after Anker, when I was moving between companies and projects and trying to stay on top of more conversations than any notebook could hold. And for a while, I told myself it was working. Then I noticed what happened when I took it out. The room changed.

The Device That Changed the Room

Not dramatically, not in a way anyone named out loud. But you could feel it. The device sat there on the table, visible, a little antenna for everyone's self-consciousness. People chose their words differently. The unguarded moment, the one where someone says what they actually think, became rarer. And the ten-minute debrief on the way out of the building, the conversation after the conversation, that happened in the elevator, without the device, because nobody thought to move it.

The technology worked. The human layer broke. I kept using it anyway, because the alternative was worse. But I filed the feeling away. The right solution wasn't just about capturing conversation. It was about capturing it in a way that didn't change what was being said.

I left Anker without a plan, which was unusual for me. Most of my career had been structured around goals. Ship the product. Build the team. Hit the number. Move to the next thing. At Anker I'd done what I came to do, more than once, and eventually I reached the point where I knew it was time for someone else to take it further. That's not a sad moment, though it can feel like one. It's more like finishing a long book and closing the cover. The story was good. It's done. What's next?

Leaving Anker Without a Plan

I didn't know. And for a while I let myself not know. Consulting filled the gap practically. It kept me close to the problems I understood, product strategy, market positioning, the operational details that only make sense to people who've actually built things. I enjoyed it more than I expected. There's a particular pleasure in walking into a messy situation with fresh eyes, asking the questions no one inside has thought to ask, and watching something unlock.

But I kept bumping into the same friction. Every new client, every new engagement, meant starting from zero. Not zero capability, I knew what I was doing. Zero context. And the more seriously everyone was starting to take AI as a working tool, the more visible that friction became. I was spending enormous energy bridging the gap between what had actually happened, in rooms, on calls, across weeks of work, and what the tools could know about it. The briefing tax. Every single time.

The recording device helped. And then it changed the room. And I kept thinking: there has to be a better way to do this. Elisa had left Anker too. Huilong had been there as well. The three of us had built up the kind of trust that only comes from working through real problems together, shipping things, missing deadlines and recovering from them, having the arguments that needed to happen and coming out the other side still able to work together.

The Team That Kept Talking

When you have that with people, you look for reasons to keep it. So we started talking. Not about a company, at first. About ideas. About where AI was going and what it might mean and where the interesting problems were. We had those conversations over months, in the way that founding teams often do before they know they're founding anything: loosely, honestly, with no particular pressure to arrive anywhere.

A lot of ideas came up. Most of them were good in the way that many ideas are good, interesting enough to discuss, not interesting enough to build. We kept turning them over, stress-testing them, moving on. And then I told them about the gap. Not as a pitch. More as a frustration I'd been carrying around and hadn't fully articulated yet. The consulting device that changed the room. The forty-five minutes of context I'd spent briefing ChatGPT on a problem it should have already known about. The realization that every smart person I knew had built some version of the same workaround, and every workaround was just a more elaborate way of doing by hand what should exist as a product.

Huilong listened. He asked a few questions, the kind of precise, technical questions that tell you someone is already working the problem in their head while you're still talking. And then he said: give me a few days. He came back with an algorithm.

I won't pretend I understood every detail of it. What I understood was what it did. It could take the raw material of a conversation, the actual spoken exchange, with all its noise and repetition and tangents, and find the structure inside it. Not just transcribe. Not just summarize. Identify what mattered, how it connected to other things that had been said, what it meant in context. It was the technical answer to the question I hadn't known how to fully ask.

The Algorithm and First Prototype

We had a solution to half the problem. Now we needed something people could actually use. None of us had built a consumer app before. The obvious move would have been to find someone who had. We talked about it, briefly, and then Elisa said she'd take a crack at it.

She built the first prototype in a week. With Huilong's input, and mine, and Claude, which turned out to be a more capable development partner than any of us had fully appreciated. Elisa had no prior app development experience. The prototype worked.

I remember sitting with it for the first time and thinking two things simultaneously. The first was: this is real, this could actually be the thing. The second was: if the three of us just built a functional app in a week with no background in it, something has genuinely changed about what's possible.

The Moment It Became Real

That was its own kind of signal. Not just that our idea was viable. That the moment we were in was different from any moment that had come before it. The tools had arrived. The capability was there. The gap was real and no one had closed it yet. We were in the right place at the right time with the right problem.

My wife had watched all of this from close range. She'd seen me leave Anker without a plan, move through the consulting period, come home with the growing certainty that something was there. When I told her what the three of us were going to build, she didn't ask me to be careful. She told me to go find it.

When we started showing Memoket to people in the winter of 2025, most of them didn't get it at first. These were smart people, people who understood technology, people who'd felt the exact friction I was describing. They'd listen to the explanation and nod and not quite feel it. I watched it happen more than once. Someone would hear "context capture device" and their brain would file it under voice recorder and move on.

And then they'd see a demo. They'd watch a week of fragmented conversations become a single structured brief. They'd see the thing they'd been losing, organized, connected, ready to use. And something would shift in their face. That face said: I've been waiting for this.

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