Building Personal Infrastructure

I read posts, books, and papers by about 350 people in a typical week. That's not a brag—it's a problem. Because what becomes of it? Where does all that go?

Most of it evaporates. I absorb something, feel briefly smarter, and then watch it dissolve into the noise of the next thing. A few weeks later I'll encounter the same idea again and experience a vague sense of recognition—not knowledge, just familiarity. The information passed through me. It didn't stick.

This bothers me more than it should. Not because I want to win trivia contests, but because I suspect the difference between people who seem to compound over time and those who just stay busy is precisely this: whether the information they encounter actually integrates or just flows through.

So I'm building infrastructure for it. I'm calling it EntityBox.

The Vision

EntityBox is a personal knowledge system—a database where information doesn't just live, but merges with me. That phrase sounds mystical, but the mechanics are practical.

Here's a concrete example. I'm on the treadmill, reading a book in a language I'm learning. I hit a word I don't know. Right now, I have two bad options: stop and look it up (lose my reading flow) or keep going (forget the word entirely). What I want is a third option: note the word instantly, in a way that acknowledges it, captures it, and queues it for later review—then immediately return to my book without losing context. The word has effectively entered my system. It's waiting for me in a review session later. It's mine now.

That's the capture side. The other half is retrieval.

Say I've spent a week reading content from 350 different people. I want to be able to sit down—maybe on that same treadmill—and review those people. Who should I read more of? Who's worth writing about? Who should I reach out to? The system should surface this for me, organized, ready for action.

Two interfaces, then: quick capture (for when you're in the middle of something) and active retrieval (for when you're ready to synthesize and act). Both feeding into a database that compounds over time. Both powered, when useful, by language models that can help you search, summarize, and surface connections you'd miss on your own.

The goal isn't to remember everything. It's to stop losing things that matter.

The Public Layer

There's another piece to this: the public interface.

My personal site is meant to be the canonical source for my work. Not the only place it appears—I'll write on Substack, post on X, comment on other people's stuff—but the estuary. The refined output. Everything else is a reflection, a summary, a remix, or a pointer back to the source.

This matters because I have a lot of thoughts I want to memorialize but nowhere good to put them. If I build a following for sharp takes on macroeconomics and then suddenly tweet something about how impressed I am by Jason Calacanis's ability to de-escalate hostile conversations—well, that might cost me a follower. Not because the observation is wrong, but because it's off-topic. And that's a shame.

So I'm thinking about "internal tweets." Thoughts that get captured and stored, but don't broadcast. They're still in the system. They can inform future work. But they don't clutter the public feed with every stray observation.

The architecture of public output should be as intentional as the architecture of private knowledge.

What I'm Actually Building Toward

I've spent a lot of time trying to build earth-changing businesses for the sake of changing the earth. That sounds noble. It also led to some of my worst years financially and emotionally. The strategy now is different: get smart, get ahead, build a platform, and then work on the things that matter most.

Concretely, this means developing a few areas of genuine expertise and talking about them consistently. Macroeconomics. Behavioral psychology—especially group behavior, desire, and the mechanics of how people actually make decisions. Business fundamentals: the stuff you'd learn in an MBA, except better, cheaper, and without the pretense.

I want to make an MBA look stupid. Not because the topics are bad—many of the best leaders have taken MBAs—but because the format is absurd. You don't need two years and a quarter million dollars to learn strategy, accounting, operations, and leadership. You need discipline, good sources, and a system that helps you retain what you learn.

EntityBox is that system. This site is the public face of it. And the plan is to show up consistently: reading, writing, engaging, building.

Alex Finn has a piece where he suggests a simple way to start building an audience: make an X list of everyone who replies to you, add the people whose content you admire, carve out fifteen minutes a day, and put thoughtful replies on everyone's content. It's not glamorous. But it's the work.

What I Won't Do

Focus requires knowing what you're excluding. A few things I've ruled out:

Hard nos: Patent hoarding, domain squatting, landlording purely for rent extraction. You're leeching on society, driving up costs, benefiting no one but yourself. Pass. Also: anything involving animal trading. The incentive structures it creates are perverse.

Uninteresting: Trading time for money without leverage—freelance work, secondary markets for books and furniture, hobby crafting without genuine love for the craft. Not morally wrong, just not the game I want to play.

Maybe later: Seminars and guest speaking. I don't have anything to say from a podium yet. When I do, I'll reconsider.

The point isn't to be judgmental. It's just that you can't build something meaningful if you're constantly distracted by every plausible side hustle. The list of things I won't do is what makes room for the things I will.

Following Along

This is an experiment I'm running in public. The infrastructure is being built—EntityBox is real, even if it's rough. The content is coming—you're reading the first of it now. The platform is forming—this site, this Substack, the slow accumulation of an audience that finds this stuff interesting.

If you're curious about knowledge systems, or personal leverage, or just want to watch someone try to build something unusual in the open, you're welcome to follow along.

Thanks for being here.