Four months after selling Agentnoon for a life changing outcome, I still haven't written about it.
I know what the piece is. The real version of why founders sell. Not the LinkedIn version. 80% financial, 10% exhaustion, 10% strategic at signing. After close it moved 40-30-30. That shift is the whole story. Nobody writes about the shift.
Instead I've been decompressing.
But I recently had my mind blown with Anthropic and Obsidian.
I had five years of notes in Roam Research, Apple Notes, voice memos, email drafts. I moved everything into Obsidian. Vin Verma had shown his setup on Greg Isenberg's podcast. I built on top of it, connecting Claude Code to the vault with slash commands that let me query my notes the way I think. He's been one of the most interesting people I've started following on X in the last year.
It took a while to start. Giving an AI access to five years of private notes is not a small decision. I haven't fully resolved that question. I built it anyway. Some of us have to take the risks early so everyone else doesn't have to.
Over 48 hours in Ramadan, between Sehri and Iftar, I built 21 of them.
/emerge reads everything in the vault and surfaces
ideas I've never explicitly stated. Conclusions the notes imply but
that I haven't said yet.
I ran /emerge on my own writing plans. $50 in API
costs and a lot of reasoning and back-and-forth later, I had answers I
wasn't ready for.
The writing infrastructure is complete. The writing itself doesn't exist.
It also found something I hadn't noticed. I imported all my old blog posts. 40-plus pieces from 2018 to 2020. The longest is maybe 200 words. Most are 50. I was posting five times in two weeks in January 2018. The writing plan I've built for 2026 calls for one essay every two weeks, each drawing from an intellectual library that spans Ibn Khaldun to Nassim Taleb.
Not the same writer.
The old posts were fast and unguarded. The 2026 plan is careful, positioned, aware of its audience. I don't know which mode is right. Probably some of both.
The exit essay is about a decision that's already been made. This one is about something happening right now: building a system for thinking in public, one command at a time, in a month where I'm fasting until sundown and have limited extra time.
The constraint is the advantage. You find out what's actually worth doing when you only have two hours.