Building

Keyword Magic

My SEO SaaS. It ranks a Shopify store’s SEO fixes by dollars, not search volume, using the store’s own Search Console, GA4 and Shopify data.

Running

kaibot

The harness I’ve built around Claude Code, and where most of these tokens go: it picks the model for each job, blocks the moves an agent shouldn’t make, lets several sessions share one repo, and keeps the memory and the review panels below.

Refining

waymark

The memory my AI sessions share. They don’t remember each other, so this is how the next one knows what the last one learned: about 19,000 facts, one per file, and each task gets only the few that matter.

Shipping

Counselors

It sends any change where a mistake would matter to a review panel drawn from 29 models across 10 AI labs, each seat playing one of 36 reviewer personas, before I trust it. A model can’t see its own blind spots, so when the labs disagree, that’s my signal.

Rebuilding

Bootstrapping.io

This site: a new design, and the field notes in six chapters.

Rebuilding

Shandalar

The 1997 Magic: The Gathering game, rebuilt. The hard part is the overworld: a grid of square tiles that has to read as one hand-painted map.

$400 1 hour

Book a call

A paid one-on-one call with answers to your questions about using AI for projects, building your AI stack, or growing your business with AI. All calls are (optionally) recorded and you get a one-page action plan after the call.