Blog
Thoughts on code, creativity, and the chaos in between.
Dev Diary: Bridges Are Paths, Not Debates
Scrapped the entire bridge system and rebuilt it around entities, lineage, and visual echoes—24,000 bridges that can actually explain why two garments are connected.
The Great Reorganization (and Teaching an AI to Argue Better)
Thirty files moved, twenty-three tests fixed, a temporal classifier rewritten, and seven changes to the narrative engine—all so a Dockerfile can be five lines instead of fifty.
A Dress Doesn’t Just Look Like Another Dress. It Argues With It.
Six-dimensional bridge classification, contrast detection between garments that share structural DNA but argue opposite aesthetics, and a social function explorer that asks how different cultures solve the same problem.
Dev Diary: Cross-Cultural Bridges & a DNS Mystery
Three new bridge fields for cross-cultural garment connections, a Supabase DNS outage, and brute-forcing AWS regions to get the database back.
Teaching an AI What a Dress Means
Replaced MMFashion with Claude Vision, built a 20-term controlled vibe vocabulary for the knowledge graph, and designed a feedback loop that lets the era taxonomy grow from the data.
Migration Complete: Qdrant Is Dead, Long Live pgvector
Twenty migration steps done, one database eliminated, and hundreds of lines of sync code deleted. The full Supabase cutover, from test rewrites to stale bridge cleanup.
Dev Diary: Burning Down the Vector Database & Sketching the Knowledge Graph
Migrated Vintage Vestige from Qdrant to pgvector in a single session — then designed a knowledge graph with AWS Neptune and Getty AAT mappings for what comes next.
Component Library Day: 17 Components From Scratch
17 hand-typed React components in a day — from design tokens to bridge primitives, plus the TypeScript patterns that made composition click.
Building the API Layer: From Schemas to Live Endpoints
13 Pydantic schemas, 13 FastAPI endpoints, and a payload gap between two Qdrant collections that would have broken image search.
Building the Bridge: Connecting Fashion Across Centuries
7,324 style bridges connecting garments across eras and categories — plus the data bugs I caught along the way.
Vintage Vestige: Teaching a Search Engine to Speak Fashion
How I built a semantic search engine for historical fashion collections using Claude, CLIP, and vector embeddings — and what MindCap taught me about input quality.
The Part Nobody Talks About: Keeping an AI on the Same Page
The hardest part of working with AI isn't the code — it's the context. Here's the documentation system that made long-term AI collaboration actually work.
1,212 Lines Deleted, Zero Lines Missed
Full-stack refactor complete: replaced a 15-category system with five behavioral intents, deleted 1,212 lines, and caught a silent bug hiding in the pattern detector.
Building the Fix: 700 Lines That Replaced 1,000
Two phases into replacing MindCap's categorization system with intent detection. Simpler, broader, and half the code.
How I Caught a Blind Spot That Would've Sunk MindCap
I ran keyword extraction against 56,000+ real browsing records and found the system only understood programmers. Here's how I redesigned it.
Dev Diary: Content-First Categorization
Rethinking topic categorization by prioritizing content signals over domain signals, plus URL garbage filtering.
MindCap: Teaching Software to Understand Rabbit Holes
Development diary on rabbit hole tracking, keyword extraction, and the cognitive overhead of building alone.
React Refactor & Finding My Voice
Escaping Intermediate Hell, rebuilding in React, and rediscovering my voice with the help of AI.
The Hard Part: Understanding Attention Through Behavior
The biggest, most ambitious project I've ever built—and an honest look at the difficulties of inferring attention from observable behavior.
Building a Topic Registry: Tracking Digital Attention Over Time
Transforming raw browsing sessions into meaningful insights about what topics capture your attention and how that interest evolves.
MindCap: Building an Attention Tracker with My AI Thought Partner
The story of building a browser extension to understand where your attention actually goes online, and what it's like to develop with Claude as a thought partner.
Firefox History Crawler: Digital Archaeology for the Chronically Curious
Building a Python tool to crawl 56,000+ URLs from my Firefox history and discover what I've actually been doing on the internet.