The Biggest Bottleneck in AX Is Not AI
The hardest part of AX is not model performance. It is getting people and organizations to agree on how work should actually be done.
Recent Posts
The Product Owner Is the Starting Pitcher
A product owner does not just organize requirements. A good PO controls the early innings of the product game.
AX Starts With Data Visibility
AI agents are not magic. Before they can act on behalf of an organization, the organization has to make its own work visible.
Don’t Build on Rented Land
In the AI gold rush, everyone is building. But most are building on someone else’s land — and the landlord can change the terms overnight.
What Agile Assumes But Never Says: The Contact Point Between Loops
Agile promises that iteration compounds. But iteration only compounds when context survives the loop boundary. Most teams treat context loss as normal. It is...
Where Context Goes to Die: From Domain Expert to Developer
Nobody did anything wrong. Everyone worked hard. And yet the product came out subtly broken. This is the story of how domain knowledge dies in transit — and ...
When AI Replaces the Middleman: Rethinking Relationships in Business
Korea ran on relationship capital — knowing the right person was how things got done. AI is now replacing the middleman. But relationships aren’t disappearin...
Why Korean Online Banking Became a Security Galapagos
How a 1999 law designed to protect online banking created a closed ecosystem that made security worse — and left Korea behind as the world moved to passkeys.
Why We Tear Down the Successful
On envy, moral licensing, and the hidden logic behind public pile-ons
Who Actually Owns the Alpha?
On the sustainability of AI finance tools in an era of commoditized intelligence
The ONE Thing in the Age of AI
When everything is possible, what you choose to focus on becomes everything
Immersion and Anchors
On the art of going deep without losing yourself
AI Literacy Is Table Stakes. Domain Literacy Is the Game.
A few years ago, “AI literacy” was a real debate. How well do you need to use AI? Is prompt engineering a valuable skill? Should you bother learning these to...
Comparative Advantage: What Humans Are For
Stop asking what AI can’t do. Start asking what you’re relatively best at.
How I Built Per-User AI Containers with Docker: The Jiyul Story
A legal & tax AI assistant built with Tauri 2 + Next.js + OpenClaw + Docker
How PMOs Can Finally Read GitHub — Bridging the Jira Gap with AI
Most PMOs are running half-blind.
Own the Domain or Lose to AI
Peter Thiel has a line that most people hear wrong.
What Legacy Code Taught Us — Why We Built ZENT Foundation
For five months, we deleted code.
Why Korea’s Tax Agents Are Fighting the Wrong Enemy
Korea has two professional licenses for tax work: Certified Public Accountants (CPAs) and Licensed Tax Agents (세무사, roughly equivalent to “Enrolled Agents” i...
How I Learned to Stop in the AI Era
Honestly, I was close to burnout last year. But the strange thing was — my mind wasn’t empty. It was the opposite.
Agentic Coding Is About Closed Loops, Not Prompts
Most people who struggle with agentic coding are solving the wrong problem.