Introducing Yu

Earlier this year I hit burnout in a way I never had before. The kind that doesn't look like slowing down. I was producing more than ever and still felt like I was accomplishing nothing, every decision, every relationship, every open loop pressing down at once.

I started using my AI to reflect on the conversations I was having and how I was feeling. I'd replay conversations and frustrations to it, offloading some of the cognitive load. And in doing that, I started to see my own patterns.

The results were mixed. On one hand, giant leaps, closing and shipping more work than ever. On the other, plenty of nights of prompting and reprompting and still not getting it right. I had files and instructions everywhere, saved in a dozen different places.

Through all of it, I came up with what I hope is the future of AI.

I think the future of AI is hyper-personalization. AI that's so attuned to your thought patterns, history, content, likes, and dislikes that it just fits. This isn't a fringe idea. Grand View Research puts the hyper-personalized technology market at about $36 billion this year, on track for $145 billion by 2033. Companies are racing to build better AI: more configuration, larger context windows, plugins, system instructions, agents. It's endless.

Most of what's being built belongs to the platform, not to you. I want it to be yours.

There are real players building memory and personalization right now, and some of it is very good. But most of it is closer to configuration than personalization. You pick A or B, or you feed it documents and past work and hope it catches on. That's training AI on the output of your thinking. What I haven't seen anyone go after is the thinking itself, the actual flows your brain runs so it can reason the way you do. LLMs are built on all of us at once. They run on the average of everyone, and cognition is the part that gets averaged away.

So I built Yu. It decodes how your brain actually works, the way you reason, not just what you've produced before, and teaches your AI to think with you. One place to build it, one place to keep it, so you stop rewriting yourself into every new chat.

Once I had the idea, I couldn't stop. Take it a step further: what if we could peek into the cognition of our colleagues? Access their AI brain and test work samples for feedback before they ever see them? Adapt feedback to how they actually receive it best? Understand why someone was so frustrated in that meeting? The possibilities are endless.

But it all starts with the same thing: understanding how you actually think. You can't hand your cognition to anyone, or teach your AI to work with it, until you can see it yourself. That's where Yu begins.

The first prototype went out for testing in 2026, and Yu has been growing ever since.

Yu is now a 52 week path you move through at your own pace. Together we map how your mind works, one framework at a time, and you build a Record of how you work plus the Scaffolding your AI runs on. You come out knowing yourself a little better, with an AI that works with you instead of the average of everyone.

I'm building this in the open, and I want the people in the room to help shape it. If any of this sounds like something you've felt, come find out what you're made of.