I spent the weekend going through my Projects folder. 30+ repos. Most of them never saw a user. Here's what I found when I stopped calling them failures and started calling them inventory.
The Projects folder on my machine is a graveyard. Or at least that's what I used to call it. Dozens of repos, each one representing a weekend or a month or three months where I was absolutely certain this was the one. The naming alone tells the story: Launch OS. Founder Tools. Solo Builder OS. Career OS. Every one of them was going to change everything.
None of them launched.
But here's what I didn't see until I went through them this weekend: every single one of them fed what I'm building now. Builder's Path didn't come from nowhere. It came from the compost of 30 projects that didn't make it. And when I actually traced the lineage, the pattern was obvious.
Five projects did the heaviest lifting. These aren't the ones I'm most proud of technically. They're the ones where the thinking was sharpest.
Next.js 15, App Router, MDX curriculum. A structured learning platform for Claude's Skills framework -- how to build reusable, context-driven instructions that extend what Claude can do. Fundamentals through deployment, 20 lessons, interactive components, file trees, terminal blocks.
I built this because Claude changed how I work. Not in the "10x productivity" way that everyone posts about. In the "I can actually ship things now" way. The Skills framework is the specific mechanism -- it's how you go from "Claude is a chatbot" to "Claude is a collaborator who knows my codebase, my voice, and my standards." I want to share this one. Other builders should feel what I feel when I sit down to work now.
128 tactical cards across three decks (Strategy, Risk, Execution), 20 curated journeys, a diagnostic flow. This one actually shipped and is live. It's the closest thing I had to a product before Builder's Path. The card format taught me something important: builders don't want essays when they're stuck. They want a sharp question and a next step. That insight shaped how the playbook's skill cards work -- the /system-prompt skill, the /ship-check skill, the /eval-builder skill. All of them follow the PM Cards format: here's your situation, here's the question, here's what to do next.
18 chapters. React SPA. Quizzes, flashcards, an AI tutor powered by Gemini. I built it as a learning platform, and it's still live on Vercel. But nobody was asking for another AI course. What it gave me was the scaffolding. The entire technical depth of the Builder's Path playbook -- Understand the Models, Prompt Engineering, Working with APIs, RAG & Knowledge, Agents & Tools, Security, Fine-Tuning, Local & Edge AI -- all of that content started life in the Masterclass. I didn't copy-paste it. I stripped out the classroom tone and rewrote it through the lens of someone actually shipping a product. The Masterclass taught me. The playbook teaches builders.
Explore the AI Engineering Masterclass →
Build. Distribute. Measure. Concentrate. This is probably my most original thinking, and it lives in a repo with 30+ markdown files that nobody has ever read. The core insight: AI collapsed the cost of building, so the old "build one thing, test it, iterate" model is dead. Build 8-10 offers in a weekend, test them in parallel with $50-100 each in ad spend, let the market pick the winner in 10 days. The Action Contract came from here too -- a four-section commitment document that forces you to stop planning and start talking to humans.
Most of this is still sitting in markdown, waiting. But the DNA is everywhere in Builder's Path. The friction blocks in the playbook, the "ten conversations" framing, the anti-hype voice -- that's all BDMC thinking wearing different clothes.
I intend to test this one for real. The whole point of BDMC is that you don't pick your best idea and pray -- you build a portfolio of offers and let the market tell you which one wins. I want to run that process on my own stuff. Take the frameworks, the playbook content, the tools, package them into 8-10 distinct offers aimed at different audiences, put $50-100 behind each one, and see what people actually click on, sign up for, and pay for. Not in theory. In practice. The framework was always meant to be used, not just documented. It's time I ate my own cooking.
10 strategic frameworks in a Vite/Supabase app. ICE scoring, Pre-Mortem Analysis, First Principles, Devil's Advocate, JTBD, Moat Analysis, FMEA, Bowling Alley Strategy, Hypothesis Testing, Value Chain Analysis. You paste in your strategy and the AI pressure-tests it across all ten lenses.
The app itself is a solid MVP -- production-ready, deployed, working auth, database persistence. But the real value was forcing myself to codify how I think about strategy. Every time I write a friction block in the playbook that says "stop and ask yourself this," the question format came from StratIQ.
Here's the part that took me a minute to see. A huge chunk of my Projects folder isn't apps -- it's content. Essays, frameworks, workbooks, all pointed inward. I was creating the content I needed to hear based on what I was learning at the time.
One repo has 39 polished essays on AI governance, enterprise failure patterns, and product management. Written from my compound perspective: FDIC examiner, Dell senior AI PM, Army drill sergeant. Six of those essays are now threaded through the playbook -- "AI Product Management Is Not Software Product Management," "Your AI Pilot Worked. That's the Dangerous Part," "The Vendor Demo Is Not the Product."
The voice thesis workbook is the most honest document I own. It's where I figured out what I actually believe: AI scales faithfully -- make sure it's scaling the right thing. It's where I wrote the voice principles that govern everything on the site. Show the resume, never claim it. Counterposition without dunking. Lead with the observation, not the framework. The layoff is a liberation, not a wound.
I wasn't writing for an audience. I was writing to figure out what I thought. And then Builder's Path became the place to say it out loud.
The graveyard has strata. Some projects were stubs -- a scaffold and nothing else (Launch OS, Founder Tools, Distribution Ninja). Some were duplicates of the same idea in different wrappers (Solo Builder OS, Solo Builder Playbook). Some were genuinely ambitious and just never found their audience:
AI Validation App -- used Claude and the Reddit API to mine problems and validate business ideas. The idea was good. The execution was real. It just never got in front of anyone.
AI Product Mastery -- a resource hub that imported the PM Cards content and added AI-powered search with embeddings. Overengineered for a problem nobody had asked me to solve.
Product Journey -- multi-variant landing pages testing different positioning. The positioning work from this directly informed how I talk about Builder's Path now.
Career OS / Learning OS -- a multi-domain learning tracker. The concept of tracking your own growth across domains lives on in how I think about the playbook's stage-based navigation.
This one deserves its own section because I'm not done with it.
Build Era is a personal development and accountability platform for people in midlife career transitions. Four pillars: Body, Mind, Family, Freedom. Weekly self-assessment via a "Builder Compass." Mini Missions (7-14 day structured projects). Pods (small-group accountability, 3-6 people, weekly check-ins). AI-powered journaling. Habit streaks.
The full-stack MVP is real -- React, Express, PostgreSQL, working auth, API routes for all core features. It targets the "Stuck Professional" -- and honestly, that's who I was when I started building it.
What makes it different from every other habit app: it acknowledges that life is multidimensional. You can't just track your pushups and call it personal development. The Mini Missions format -- 7-14 days, not 30 or 90 -- is the right scope for someone who's been burned by ambitious plans before. And the Pod system avoids the social feed noise of most accountability apps.
I may launch this as a separate app. The thinking is too good to leave in a repo.
It's not a site I built from scratch. It's a site I excavated from 30+ projects, 200,000+ words of writing, and two years of building things nobody used.
The playbook's 19 sections? Masterclass scaffolding + BDMC thinking + consulting essays + friction blocks from the voice thesis.
The free tools? Cost calculator logic from the Masterclass. Readiness assessment questions from AI Validation App thinking. ROI calculator from BDMC's "make the business case" framing.
The skills? PM Cards format. Claude Ninja's mental model. StratIQ's question frameworks.
The voice? 39 essays I wrote to myself, pressure-tested in a workbook, refined until it stopped sounding like everyone else.
Nothing was wasted. Every "failed" project was a draft of the one that's live now. The graveyard was a garden the whole time. I just couldn't see it until I stopped building new things and started looking at what I'd already built.
The Projects folder still has 30+ repos in it. I haven't deleted any of them. Turns out they're not failures. They're inventory.
Builder's Path is a public lab from Sellhausen AI Systems focused on AI-native building, validation, and product judgment.