Posts

All the articles I've posted.

• Books I Read in Q1 2026

  14 Apr, 2026

  Notes on the 10 books I read in Q1 2026: Stoner, The Obstacle Is the Way, Meditations, Source Code: My Beginnings, Bird That Drinks Blood, the second Krafton Way book, Lost and Founder, The Score Takes Care of Itself, Formula One, and Principles.

• Reading 'The Score Takes Care of Itself': On Leadership

  13 Apr, 2026

  Notes on Bill Walsh's leadership book 'The Score Takes Care of Itself.' The good-talent-bad-attitude equation, his declaration that teaching is the very definition of leadership, and the upward leadership chapter. The three-time Super Bowl winning coach's principles, read against my own failures.

• The Moment You Spin Up an AX Team, You've Already Lost

  8 Apr, 2026

  The paradox of setting up an AX task force. MIT NANDA found the successful 5% weren't led by central AI labs but by line managers on the ground. What I learned from going from consultant to CTO. The first principle of organizational AX isn't tools — it's people and the organization.

• When Package Install Becomes a Hack: Why Zero Dependency Matters

  4 Apr, 2026

  In March 2026, the litellm and axios supply-chain attacks hit back-to-back. I trace ten years of dependencies turning from convenience to trust to risk, and use code to show why zero-dependency design is survival, not minimalism.

• AI Native Engineer: Taste Built on Principles

  25 Mar, 2026

  Being good with AI tools doesn't make you an AI Native Engineer. Taste without principles is guesswork; principles without taste stay academic. After hitting the wall in iOS, getting blindsided as a Maker, and falling into the sorcerer's apprentice trap, I came to one conclusion: an AI-era engineer's identity isn't the toolkit, it's taste built on principles.

• Installing Claude Code Across Your Org Doesn't Make It AX

  15 Mar, 2026

  Rolling out an AI tool company-wide and going through an AX (AI Transformation) are two completely different things. Real AX starts when you redesign roles, KPIs, pipelines, and governance, not when you bolt AI onto a functional org.

• Between a Working Feature and a Trustworthy Product: Building ToC Recognition

  13 Mar, 2026

  A Do Work → Good → Great engineering journey building a book table-of-contents (ToC) recognition feature with Qwen 3.5 Flash. From a $2,000 pre-collection failure, through a sync API → async job system → SSE real-time streaming evolution. When the model hands you 80, the remaining 20 is engineering.

• Code Review in the AI Era: How Should We Do It?

  8 Mar, 2026

  From the reality of code review during my CTO years to the current AI-era debate. Simon Willison, Kent Beck, Bryan Finster, StrongDM, Salesforce, latent.space — I lay out the spectrum around code review and ask what it is we should actually be reviewing.

• How the World's Largest AI Company Trains the Next Generation of Engineers: A Review of the Anthropic x CodePath Curriculum

  6 Mar, 2026

  I broke down the AI Engineering curriculum Anthropic and CodePath built together, week by week. This piece walks through why fundamentals and critical code reading still sit at the center even when AI writes the code, the AI Native Engineer profile that falls out of the curriculum, and what I see in my own mentoring.

• Wrapping Up January and February 2026 (Not Really a Retro)

  2 Mar, 2026

  After leaving the company, a sabbatical, a trip to Italy, fights and reconciliations with my partner, and finally getting back to work. Too thin on results to call a retro, too long to call a journal entry.