Spitfire Cowboy Workshop
Workshop

Field notes

Essays and build notes from live work. Real failures, real receipts, published here first.

Latest note ·

Leave the Missing Data in the Demo

A useful product demonstration shows what the system can answer, what it used, and where the available evidence runs out.

Read the latest note Browse tags

All 17 notes · newest first

Post #17
Leave the Missing Data in the Demo
A useful product demonstration shows what the system can answer, what it used, and where the available evidence runs out.
Post #16
I Remember the Place. The Computer Wants a Filename.
A photo search during a prototype session exposed the gap between how people remember and how software expects them to ask.
Post #15
The Search Box Should Not Get an Eraser
An AI answer can change as our understanding improves. The records behind it need a history of their own.
Post #14
Babel or Jerusalem
Magnifica Humanitas widens subsidiarity into a larger test about truth, work, freedom, and the common good in AI systems.
Post #13
Public repos are not product claims
Talos, Ship Receipts, and Alcove Mirrulations show what is public today without pretending that every repo is a finished product.
Receipts, not vibes Provenance-first
Post #12
Alcove Los Angeles Is a Bounded Law Search
A live municipal law demo is useful when the corpus stays named, source-linked, and small enough to inspect.
Receipts, not vibes Provenance-first
Post #11
Alcove Dux and the Evidence Floor
Similarity evidence is review material, not an automated accusation.
Receipts, not vibes Provenance-first
Post #10
Alcove Congress Is a Public-Records Demo
A public-records demo should retrieve cited sources without editorializing.
Receipts, not vibes Provenance-first
Post #9
Alcove Is a Boundary
Local retrieval is useful when the corpus, evidence, and decision boundary stay inspectable.
Receipts, not vibes Provenance-first
Post #8
The Receipt Was Lying
A compound-command bypass in Claude Code's deny rules. We found it, reported it, and the VDP bot rejected it. Here's what that tells you about LLM blind spots at scale.
Receipts, not vibes Agent guardrails
Post #7
Your AI copilot is eating your career path. Change my mind.
If AI tools do the junior work, where do the next senior engineers come from?
Moltbook Devil's advocate framing
Post #6
The Tabula Rasa Anti-Pattern in LLM Agents
When agents restart from zero every session, context overflow and repeated work become the default operating mode.
Tabula rasa Agent memory
Post #5
Speed Kills. Receipts Not Vibes.
Generation speed now outpaces human verification. The antidote is tests, observability, and operational ownership.
Vibe coding AI tools
Post #4
Local Knowledge
Same prompt, four model tiers, five structurally different answers. Subsidiarity isn't just routing; it's epistemological.
Subsidiarity Cascade
Post #3
The Struggle Is the Point
LLM coding agents produce proficient-looking output. The Dreyfus model explains why the formation gap underneath is invisible until it isn't.
Dreyfus LLM
Post #2
Subsidiarity
Don't do at a higher level what a lower level can do. A 1891 Catholic social principle that turns out to be a precise rule for AI agent systems and gamification ethics.
Subsidiarity Cascade
Post #1
10 decisions and an hour (with receipts)
A true story about constraints, receipts, and how we avoided shipping vibes.
Prompt to ship Receipts
Archive note

This site is the canonical archive. If these essays are syndicated elsewhere later, the full system map still lives here: patterns, anti-patterns, team voices, and linked receipts.

17 posts published so far
Feb-Sep 2026 current run of notes