I Tried Claude Design. My Terminal Already Does It.
AI Coding

I Tried Claude Design. My Terminal Already Does It.

No designer, can't pick a layout: Claude made A1-C3 mockups while it was already building, and updated the plan when I chose. Then I tried Claude Design and went back to the terminal.

· 4 min read
Forget the Ladder: Learning Top-Down in the AI Era
AI Takes

Forget the Ladder: Learning Top-Down in the AI Era

Bottom-up learning is a ladder; with LLMs you can start from the top and go down the layers when you need them. How I learned ElevenLabs agents in months, and why my kids are learning Arduino the same way.

· 5 min read
Imperfection Is the New Lean
Product

Imperfection Is the New Lean

From Gantt charts to releasing every day or two: as a solo builder, accepting imperfection and iterating with customers is the only way to survive, and to ship good products.

· 3 min read
I Tested Claude Code’s Ultracode. It Was Amazing, Until It Burned My Whole Week in Half a Day.
AI Coding

I Tested Claude Code’s Ultracode. It Was Amazing, Until It Burned My Whole Week in Half a Day.

I tested Claude Code's Ultracode orchestrator with a long epic and it ate my whole weekly limit in less than a day. Three lessons, and why I'm building a lighter version inside k0.

· 3 min read
The Handover I Needed From Claude Code
AI Coding

The Handover I Needed From Claude Code

A small Claude Code hook that writes the handover I need when I come back to a session hours later.

· 4 min read
Opus Was Down, So I Moved Everything to Sonnet. Then Morning Came.
AI Coding

Opus Was Down, So I Moved Everything to Sonnet. Then Morning Came.

Opus went down, so I moved every session to Sonnet to keep working — until Supabase went down too, and only one of the two models could actually fix it.

· 3 min read
My House Has a Frankenstein Now, and His Name Is Franco
AI Coding

My House Has a Frankenstein Now, and His Name Is Franco

After ten years of writing Alexa skills that either failed, cost money, or wasted my time, I built my own voice assistant out of a decade-old Dell, a USB microphone, and my own Claude Code subscription — and gave him a name that admits exactly what he is.

· 6 min read
Twenty Terminals Open, and No Idea Which One Needed Me
AI Coding

Twenty Terminals Open, and No Idea Which One Needed Me

Running twenty Claude Code sessions at once is easy. Knowing which one is waiting for you is not. So I stopped paying for Cursor, stopped opening VS Code, and built a board of sticky notes where a card is the color of its status.

· 6 min read
How Plan Mode Actually Works, and Why You Don't Always Need It
AI Coding

How Plan Mode Actually Works, and Why You Don't Always Need It

I caught myself hitting Shift+Tab out of habit, so I went back and read the documentation. Plan mode is a permission mode, not a thinking mode — and in bypass it isn't even enforced. Here's what actually changes when you plan, and when to skip it.

· 12 min read
Git Worktree: How to Run Parallel Claude Code Sessions Without Them Fighting Over the Same Files
AI Coding

Git Worktree: How to Run Parallel Claude Code Sessions Without Them Fighting Over the Same Files

Running multiple Claude Code sessions in the same folder eventually means they collide over the same files. Git worktree gives each plan its own isolated folder — work in parallel, then merge back and delete it when you're done. And once a hook triggers it on every plan, you stop having to remember it at all.

· 7 min read
The Skills I Actually Kept: My Claude Code Skills, and What I Use vs Build
AI Coding

The Skills I Actually Kept: My Claude Code Skills, and What I Use vs Build

The Claude Code skills I actually kept — article drafting, a full deploy pipeline, design-system contracts, self-testing — and the one rule I use to decide when to reuse a pre-built skill versus write my own: who's going to maintain it in six months.

· 8 min read
UI Thrashing: The Reset Protocol That Actually Works
AI Coding

UI Thrashing: The Reset Protocol That Actually Works

When AI-driven UI coding turns into an endless loop of micro-tweaks, don't negotiate with the model — kill the session, max the thinking effort, resend everything, and let it plan without more questions.

· 2 min read