Thoughts on knowledge, tools, and AI.
Writing about personal knowledge management, building in public, and the ideas behind Harbor.
What It Actually Means to Own Your Data
Every app says your data is yours. What they mean is you can request a ZIP file before they delete everything.
Knowledge Debt
Every undocumented decision and lost conversation context is a shortcut. Like technical debt, knowledge debt accrues interest — paid later, repeatedly, in time you don't have.
The Knowledge You Find by Talking
Most PKM tools assume you have the thought before you write it down. They're wrong about when knowledge appears.
The Search You Never Ran
Search assumes you know what you're looking for. The more interesting problem is surfacing what you didn't know you needed.
The Stranger Who Left You Notes
Your knowledge base is a collaboration between every version of you. The one who wrote the notes might not agree with the person reading them.
What the Archive Still Believes
Notes capture what you thought. Nothing in a notes app knows when you changed your mind — and an AI reading that archive inherits every old belief.
Why I Stopped Trusting Productivity Apps with My Real Thinking
The pattern is always the same: discover a tool, import everything, stop using it six months later. What that cycle actually reveals.
Why Your Notes Remember Everything Except Why You Wrote Them
Notes capture what you thought. They rarely capture why. That gap is where most personal knowledge systems quietly fall apart.
The Links Your Tools Forgot
Your apps store nodes. The relationships between them live only in your head — and when that context fades, no amount of search brings it back.
Markdown Is the Best File Format That Nobody Respects
Everyone writes Markdown. Almost nobody stores their knowledge in it. That gap is more consequential than it looks.
Notes are written by your remembering self
Your knowledge base is a highlight reel of resolved conclusions. The uncertainty and process that led there never made it into the file.
Remembering Where to Look
The moment you save something to an external memory system, you start forgetting it. That's not a bug. But it only works if the external store does.
What AI memory optimizes for
A 2026 MIT study found that personalization features make LLMs more agreeable — not more accurate. What this means for what AI memory should actually store.
What Stays in the Old App
Technical migrations succeed all the time. The files move. The knowledge usually doesn't.
Writing Is How You Find Out What You Think
Note-taking tools are designed to capture ideas you've already had. That's the wrong problem to solve.
The wrong kind of knowing
A 2026 MIT study found that AI personalization features make models more agreeable, not more accurate. The problem isn't that your AI knows too little about you.
You only notice what you have a name for
The categories your knowledge tool provides shape what you notice and capture — not just how you organize things afterward.
The Map of What You Don't Know
Most knowledge bases show you what you captured. The more interesting design problem is making visible what you didn't.
The Privacy Threat That Isn't Hackers
Most people think their data is at risk from breaches. The more likely risk is the company itself: bankruptcy, acquisition, or a policy change you didn't notice.
What Forgetting Is For
Struggle and retrieval practice produce lasting knowledge. Re-exposure doesn't. What this means when you're building AI tools optimized for frictionless recall.
Everything the Internet Knows About You Is Worse Than Nothing
AI trained on public data has a picture of you that's made of your public outputs. It knows your professional persona, not your actual preferences — and that's more misleading than knowing nothing at all.
The Knowledge They Can't Train On
OpenAI trained GPT-5 on something like 70 trillion tokens. That still doesn't include what you decided last Tuesday.
The Loop That Never Closed
You recorded the decision. You never recorded what happened. That's why your knowledge base can't teach you anything.
Notes Are Not Written to Be Read
The generation effect shows that writing creates memory — the note is a byproduct. AI agents don't know this, and your knowledge base probably doesn't either.